Search Results: "nate"

5 January 2026

Vincent Bernat: Using eBPF to load-balance traffic across UDP sockets with Go

Akvorado collects sFlow and IPFIX flows over UDP. Because UDP does not retransmit lost packets, it needs to process them quickly. Akvorado runs several workers listening to the same port. The kernel should load-balance received packets fairly between these workers. However, this does not work as expected. A couple of workers exhibit high packet loss: 
$ curl -s 127.0.0.1:8080/api/v0/inlet/metrics \
>   sed -n s/akvorado_inlet_flow_input_udp_in_dropped//p
packets_total listener="0.0.0.0:2055",worker="0"  0
packets_total listener="0.0.0.0:2055",worker="1"  0
packets_total listener="0.0.0.0:2055",worker="2"  0
packets_total listener="0.0.0.0:2055",worker="3"  1.614933572278264e+15
packets_total listener="0.0.0.0:2055",worker="4"  0
packets_total listener="0.0.0.0:2055",worker="5"  0
packets_total listener="0.0.0.0:2055",worker="6"  9.59964121598348e+14
packets_total listener="0.0.0.0:2055",worker="7"  0
eBPF can help by implementing an alternate balancing algorithm.

Options for load-balancing There are three methods to load-balance UDP packets across workers:
  1. One worker receives the packets and dispatches them to the other workers.
  2. All workers share the same socket.
  3. Each worker has its own socket, listening to the same port, with the SO_REUSEPORT socket option.

SO_REUSEPORT option Tom Hebert added the SO_REUSEPORT socket option in Linux 3.9. The cover letter for his patch series explains why this new option is better than the two existing ones from a performance point of view:
SO_REUSEPORT allows multiple listener sockets to be bound to the same port. [ ] Received packets are distributed to multiple sockets bound to the same port using a 4-tuple hash. The motivating case for SO_RESUSEPORT in TCP would be something like a web server binding to port 80 running with multiple threads, where each thread might have it s own listener socket. This could be done as an alternative to other models:
  1. have one listener thread which dispatches completed connections to workers, or
  2. accept on a single listener socket from multiple threads.
In case #1, the listener thread can easily become the bottleneck with high connection turn-over rate. In case #2, the proportion of connections accepted per thread tends to be uneven under high connection load. [ ] We have seen the disproportion to be as high as 3:1 ratio between thread accepting most connections and the one accepting the fewest. With SO_REUSEPORT the distribution is uniform. The motivating case for SO_REUSEPORT in UDP would be something like a DNS server. An alternative would be to receive on the same socket from multiple threads. As in the case of TCP, the load across these threads tends to be disproportionate and we also see a lot of contection on the socket lock.
Akvorado uses the SO_REUSEPORT option to dispatch the packets across the workers. However, because the distribution uses a 4-tuple hash, a single socket handles all the flows from one exporter.

SO_ATTACH_REUSEPORT_EBPF option In Linux 4.5, Craig Gallek added the SO_ATTACH_REUSEPORT_EBPF option to attach an eBPF program to select the target UDP socket. In Linux 4.6, he extended it to support TCP. The socket(7) manual page documents this mechanism:1
The BPF program must return an index between 0 and N-1 representing the socket which should receive the packet (where N is the number of sockets in the group). If the BPF program returns an invalid index, socket selection will fall back to the plain SO_REUSEPORT mechanism.
In Linux 4.19, Martin KaFai Lau added the BPF_PROG_TYPE_SK_REUSEPORT program type. Such an eBPF program selects the socket from a BPF_MAP_TYPE_REUSEPORT_ARRAY map instead. This new approach is more reliable when switching target sockets from one instance to another for example, when upgrading, a new instance can add its sockets and remove the old ones.

Load-balancing with eBPF and Go Altering the load-balancing algorithm for a group of sockets requires two steps:
  1. write and compile an eBPF program in C,2 and
  2. load it and attach it in Go.

eBPF program in C A simple load-balancing algorithm is to randomly choose the destination socket. The kernel provides the bpf_get_prandom_u32() helper function to get a pseudo-random number. 
volatile const __u32 num_sockets; //  
struct  
    __uint(type, BPF_MAP_TYPE_REUSEPORT_SOCKARRAY);
    __type(key, __u32);
    __type(value, __u64);
    __uint(max_entries, 256);
  socket_map SEC(".maps"); //  
SEC("sk_reuseport")
int reuseport_balance_prog(struct sk_reuseport_md *reuse_md)
 
    __u32 index = bpf_get_prandom_u32() % num_sockets; //  
    bpf_sk_select_reuseport(reuse_md, &socket_map, &index, 0); //  
    return SK_PASS; //  
 
char _license[] SEC("license") = "GPL";
In , we declare a volatile constant for the number of sockets in the group. We will initialize this constant before loading the eBPF program into the kernel. In , we define the socket map. We will populate it with the socket file descriptors. In , we randomly select the index of the target socket.3 In , we invoke the bpf_sk_select_reuseport() helper to record our decision. Finally, in , we accept the packet.

Header files If you compile the C source with clang, you get errors due to missing headers. The recommended way to solve this is to generate a vmlinux.h file with bpftool: 
$ bpftool btf dump file /sys/kernel/btf/vmlinux format c > vmlinux.h
Then, include the following headers:4 
#include "vmlinux.h"
#include <bpf/bpf_helpers.h>
For my 6.17 kernel, the generated vmlinux.h is quite big: 2.7 MiB. Moreover, bpf/bpf_helpers.h is shipped with libbpf. This adds another dependency for users. As the eBPF program is quite small, I prefer to put the strict minimum in vmlinux.h by cherry-picking the definitions I need.

Compilation The eBPF Library for Go ships bpf2go, a tool to compile eBPF programs and to generate some scaffolding code. We create a gen.go file with the following content: 
package main
//go:generate go tool bpf2go -tags linux reuseport reuseport_kern.c
After running go generate ./..., we can inspect the resulting objects with readelf and llvm-objdump: 
$ readelf -S reuseport_bpfeb.o
There are 14 section headers, starting at offset 0x840:
  [Nr] Name              Type             Address           Offset
[ ]
  [ 3] sk_reuseport      PROGBITS         0000000000000000  00000040
  [ 6] .maps             PROGBITS         0000000000000000  000000c8
  [ 7] license           PROGBITS         0000000000000000  000000e8
[ ]
$ llvm-objdump -S reuseport_bpfeb.o
reuseport_bpfeb.o:  file format elf64-bpf
Disassembly of section sk_reuseport:
0000000000000000 <reuseport_balance_prog>:
;  
       0:   bf 61 00 00 00 00 00 00     r6 = r1
;     __u32 index = bpf_get_prandom_u32() % num_sockets;
       1:   85 00 00 00 00 00 00 07     call 0x7
[ ]

Usage from Go Let s set up 10 workers listening to the same port.5 Each socket enables the SO_REUSEPORT option before binding:6 
var (
    err error
    fds []uintptr
    conns []*net.UDPConn
)
workers := 10
listenAddr := "127.0.0.1:0"
listenConfig := net.ListenConfig 
    Control: func(_, _ string, c syscall.RawConn) error  
        c.Control(func(fd uintptr)  
            err = unix.SetsockoptInt(int(fd), unix.SOL_SOCKET, unix.SO_REUSEPORT, 1)
            fds = append(fds, fd)
         )
        return err
     ,
 
for range workers  
    pconn, err := listenConfig.ListenPacket(t.Context(), "udp", listenAddr)
    if err != nil  
        t.Fatalf("ListenPacket() error:\n%+v", err)
     
    udpConn := pconn.(*net.UDPConn)
    listenAddr = udpConn.LocalAddr().String()
    conns = append(conns, udpConn)
The second step is to load the eBPF program, initialize the num_sockets variable, populate the socket map, and attach the program to the first socket.7 
// Load the eBPF collection.
spec, err := loadReuseport()
if err != nil  
    t.Fatalf("loadVariables() error:\n%+v", err)
 
// Set "num_sockets" global variable to the number of file descriptors we will register
if err := spec.Variables["num_sockets"].Set(uint32(len(fds))); err != nil  
    t.Fatalf("NumSockets.Set() error:\n%+v", err)
 
// Load the map and the program into the kernel.
var objs reuseportObjects
if err := spec.LoadAndAssign(&objs, nil); err != nil  
    t.Fatalf("loadReuseportObjects() error:\n%+v", err)
 
t.Cleanup(func()   objs.Close()  )
// Assign the file descriptors to the socket map.
for worker, fd := range fds  
    if err := objs.reuseportMaps.SocketMap.Put(uint32(worker), uint64(fd)); err != nil  
        t.Fatalf("SocketMap.Put() error:\n%+v", err)
     
 
// Attach the eBPF program to the first socket.
socketFD := int(fds[0])
progFD := objs.reuseportPrograms.ReuseportBalanceProg.FD()
if err := unix.SetsockoptInt(socketFD, unix.SOL_SOCKET, unix.SO_ATTACH_REUSEPORT_EBPF, progFD); err != nil  
    t.Fatalf("SetsockoptInt() error:\n%+v", err)
We are now ready to process incoming packets. Each worker is a Go routine incrementing a counter for each received packet:8 
var wg sync.WaitGroup
receivedPackets := make([]int, workers)
for worker := range workers  
    conn := conns[worker]
    packets := &receivedPackets[worker]
    wg.Go(func()  
        payload := make([]byte, 9000)
        for  
            if _, err := conn.Read(payload); err != nil  
                if errors.Is(err, net.ErrClosed)  
                    return
                 
                t.Logf("Read() error:\n%+v", err)
             
            *packets++
         
     )
Let s send 1000 packets: 
sentPackets := 1000
conn, err := net.Dial("udp", conns[0].LocalAddr().String())
if err != nil  
    t.Fatalf("Dial() error:\n%+v", err)
 
defer conn.Close()
for range sentPackets  
    if _, err := conn.Write([]byte("hello world!")); err != nil  
        t.Fatalf("Write() error:\n%+v", err)
If we print the content of the receivedPackets array, we can check the balancing works as expected, with each worker getting about 100 packets: 
=== RUN   TestUDPWorkerBalancing
    balancing_test.go:84: receivedPackets[0] = 107
    balancing_test.go:84: receivedPackets[1] = 92
    balancing_test.go:84: receivedPackets[2] = 99
    balancing_test.go:84: receivedPackets[3] = 105
    balancing_test.go:84: receivedPackets[4] = 107
    balancing_test.go:84: receivedPackets[5] = 96
    balancing_test.go:84: receivedPackets[6] = 102
    balancing_test.go:84: receivedPackets[7] = 105
    balancing_test.go:84: receivedPackets[8] = 99
    balancing_test.go:84: receivedPackets[9] = 88
    balancing_test.go:91: receivedPackets = 1000
    balancing_test.go:92: sentPackets     = 1000

Graceful restart You can also use SO_ATTACH_REUSEPORT_EBPF to gracefully restart an application. A new instance of the application binds to the same address and prepare its own version of the socket map. Once it attaches the eBPF program to the first socket, the kernel steers incoming packets to this new instance. The old instance needs to drain the already received packets before shutting down. To check we are not losing any packet, we spawn a Go routine to send as many packets as possible: 
sentPackets := 0
notSentPackets := 0
done := make(chan bool)
conn, err := net.Dial("udp", conns1[0].LocalAddr().String())
if err != nil  
    t.Fatalf("Dial() error:\n%+v", err)
 
defer conn.Close()
go func()  
    for  
        if _, err := conn.Write([]byte("hello world!")); err != nil  
            notSentPackets++
          else  
            sentPackets++
         
        select  
        case <-done:
            return
        default:
         
     
 ()
Then, while the Go routine runs, we start the second set of workers. Once they are running, they start receiving packets. If we gracefully stop the initial set of workers, not a single packet is lost!9 
=== RUN   TestGracefulRestart
    graceful_test.go:135: receivedPackets1[0] = 165
    graceful_test.go:135: receivedPackets1[1] = 195
    graceful_test.go:135: receivedPackets1[2] = 194
    graceful_test.go:135: receivedPackets1[3] = 190
    graceful_test.go:135: receivedPackets1[4] = 213
    graceful_test.go:135: receivedPackets1[5] = 187
    graceful_test.go:135: receivedPackets1[6] = 170
    graceful_test.go:135: receivedPackets1[7] = 190
    graceful_test.go:135: receivedPackets1[8] = 194
    graceful_test.go:135: receivedPackets1[9] = 155
    graceful_test.go:139: receivedPackets2[0] = 1631
    graceful_test.go:139: receivedPackets2[1] = 1582
    graceful_test.go:139: receivedPackets2[2] = 1594
    graceful_test.go:139: receivedPackets2[3] = 1611
    graceful_test.go:139: receivedPackets2[4] = 1571
    graceful_test.go:139: receivedPackets2[5] = 1660
    graceful_test.go:139: receivedPackets2[6] = 1587
    graceful_test.go:139: receivedPackets2[7] = 1605
    graceful_test.go:139: receivedPackets2[8] = 1631
    graceful_test.go:139: receivedPackets2[9] = 1689
    graceful_test.go:147: receivedPackets = 18014
    graceful_test.go:148: sentPackets     = 18014
Unfortunately, gracefully shutting down a UDP socket is not trivial in Go.10 Previously, we were terminating workers by closing their sockets. However, if we close them too soon, the application loses packets that were assigned to them but not yet processed. Before stopping, a worker needs to call conn.Read() until there are no more packets. A solution is to set a deadline for conn.Read() and check if we should stop the Go routine when the deadline is exceeded: 
payload := make([]byte, 9000)
for  
    conn.SetReadDeadline(time.Now().Add(50 * time.Millisecond))
    if _, err := conn.Read(payload); err != nil  
        if errors.Is(err, os.ErrDeadlineExceeded)  
            select  
            case <-done:
                return
            default:
                continue
             
         
        t.Logf("Read() error:\n%+v", err)
     
    *packets++
With TCP, this aspect is simpler: after enabling the net.ipv4.tcp_migrate_req sysctl, the kernel automatically migrates waiting connections to a random socket in the same group. Alternatively, eBPF can also control this migration. Both features are available since Linux 5.14.

Addendum After implementing this strategy in Akvorado, all workers now drop packets! 
$ curl -s 127.0.0.1:8080/api/v0/inlet/metrics \
>   sed -n s/akvorado_inlet_flow_input_udp_in_dropped//p
packets_total listener="0.0.0.0:2055",worker="0"  838673
packets_total listener="0.0.0.0:2055",worker="1"  843675
packets_total listener="0.0.0.0:2055",worker="2"  837922
packets_total listener="0.0.0.0:2055",worker="3"  841443
packets_total listener="0.0.0.0:2055",worker="4"  840668
packets_total listener="0.0.0.0:2055",worker="5"  850274
packets_total listener="0.0.0.0:2055",worker="6"  835488
packets_total listener="0.0.0.0:2055",worker="7"  834479
The root cause is the default limit of 32 records for Kafka batch sizes. This limit is too low because the brokers have a large overhead when handling each batch: they need to ensure they persist correctly before acknowledging them. Increasing the limit to 4096 records fixes this issue. While load-balancing incoming flows with eBPF remains useful, it did not solve the main issue. At least the even distribution of dropped packets helped identify the real bottleneck.
  1. The current version of the manual page is incomplete and does not cover the evolution introduced in Linux 4.19. There is a pending patch about this.
  2. Rust is another option. However, the program we use is so trivial that it does not make sense to use Rust.
  3. As bpf_get_prandom_u32() returns a pseudo-random 32-bit unsigned value, this method exhibits a very slight bias towards the first indexes. This is unlikely to be worth fixing.
  4. Some examples include <linux/bpf.h> instead of "vmlinux.h". This makes your eBPF program dependent on the installed kernel headers.
  5. listenAddr is initially set to 127.0.0.1:0 to allocate a random port. After the first iteration, it is updated with the allocated port.
  6. This is the setupSockets() function in fixtures_test.go.
  7. This is the setupEBPF() function in fixtures_test.go.
  8. The complete code is in balancing_test.go
  9. The complete code is in graceful_test.go
  10. In C, we would poll() both the socket and a pipe used to signal for shutdown. When the second condition is triggered, we drain the socket by executing a series of non-blocking read() until we get EWOULDBLOCK.

4 January 2026

Sahil Dhiman: Full /8 IPv4 Address Block Holders in 2025

Here, I tried to collect entities that owned full /8 IPv4 address block(s) in 2025. A more comprehensive and historical list can be found at https://en.wikipedia.org/wiki/List_of_assigned_/8_IPv4_address_blocks.

Some notable mentions: Note - HE s list mentions addresses originated (not just owned). Providers also advertise prefixes owned by their customers in Bring Your Own IP (BYOIP) setups. The number of address space owned by entity would mostly be more than their customer owned space (originated by the entity).
  • There s also 126.0.0.0/8 which is mostly Softbank.

Matthew Garrett: What is a PC compatible?

Wikipedia says An IBM PC compatible is any personal computer that is hardware- and software-compatible with the IBM Personal Computer (IBM PC) and its subsequent models . But what does this actually mean? The obvious literal interpretation is for a device to be PC compatible, all software originally written for the IBM 5150 must run on it. Is this a reasonable definition? Is it one that any modern hardware can meet? Before we dig into that, let s go back to the early days of the x86 industry. IBM had launched the PC built almost entirely around off-the-shelf Intel components, and shipped full schematics in the IBM PC Technical Reference Manual. Anyone could buy the same parts from Intel and build a compatible board. They d still need an operating system, but Microsoft was happy to sell MS-DOS to anyone who d turn up with money. The only thing stopping people from cloning the entire board was the BIOS, the component that sat between the raw hardware and much of the software running on it. The concept of a BIOS originated in CP/M, an operating system originally written in the 70s for systems based on the Intel 8080. At that point in time there was no meaningful standardisation - systems might use the same CPU but otherwise have entirely different hardware, and any software that made assumptions about the underlying hardware wouldn t run elsewhere. CP/M s BIOS was effectively an abstraction layer, a set of code that could be modified to suit the specific underlying hardware without needing to modify the rest of the OS. As long as applications only called BIOS functions, they didn t need to care about the underlying hardware and would run on all systems that had a working CP/M port. By 1979, boards based on the 8086, Intel s successor to the 8080, were hitting the market. The 8086 wasn t machine code compatible with the 8080, but 8080 assembly code could be assembled to 8086 instructions to simplify porting old code. Despite this, the 8086 version of CP/M was taking some time to appear, and a company called Seattle Computer Products started producing a new OS closely modelled on CP/M and using the same BIOS abstraction layer concept. When IBM started looking for an OS for their upcoming 8088 (an 8086 with an 8-bit data bus rather than a 16-bit one) based PC, a complicated chain of events resulted in Microsoft paying a one-off fee to Seattle Computer Products, porting their OS to IBM s hardware, and the rest is history. But one key part of this was that despite what was now MS-DOS existing only to support IBM s hardware, the BIOS abstraction remained, and the BIOS was owned by the hardware vendor - in this case, IBM. One key difference, though, was that while CP/M systems typically included the BIOS on boot media, IBM integrated it into ROM. This meant that MS-DOS floppies didn t include all the code needed to run on a PC - you needed IBM s BIOS. To begin with this wasn t obviously a problem in the US market since, in a way that seems extremely odd from where we are now in history, it wasn t clear that machine code was actually copyrightable. In 1982 Williams v. Artic determined that it could be even if fixed in ROM - this ended up having broader industry impact in Apple v. Franklin and it became clear that clone machines making use of the original vendor s ROM code wasn t going to fly. Anyone wanting to make hardware compatible with the PC was going to have to find another way. And here s where things diverge somewhat. Compaq famously performed clean-room reverse engineering of the IBM BIOS to produce a functionally equivalent implementation without violating copyright. Other vendors, well, were less fastidious - they came up with BIOS implementations that either implemented a subset of IBM s functionality, or didn t implement all the same behavioural quirks, and compatibility was restricted. In this era several vendors shipped customised versions of MS-DOS that supported different hardware (which you d think wouldn t be necessary given that s what the BIOS was for, but still), and the set of PC software that would run on their hardware varied wildly. This was the era where vendors even shipped systems based on the Intel 80186, an improved 8086 that was both faster than the 8086 at the same clock speed and was also available at higher clock speeds. Clone vendors saw an opportunity to ship hardware that outperformed the PC, and some of them went for it. You d think that IBM would have immediately jumped on this as well, but no - the 80186 integrated many components that were separate chips on 8086 (and 8088) based platforms, but crucially didn t maintain compatibility. As long as everything went via the BIOS this shouldn t have mattered, but there were many cases where going via the BIOS introduced performance overhead or simply didn t offer the functionality that people wanted, and since this was the era of single-user operating systems with no memory protection, there was nothing stopping developers from just hitting the hardware directly to get what they wanted. Changing the underlying hardware would break them. And that s what happened. IBM was the biggest player, so people targeted IBM s platform. When BIOS interfaces weren t sufficient they hit the hardware directly - and even if they weren t doing that, they d end up depending on behavioural quirks of IBM s BIOS implementation. The market for DOS-compatible but not PC-compatible mostly vanished, although there were notable exceptions - in Japan the PC-98 platform achieved significant success, largely as a result of the Japanese market being pretty distinct from the rest of the world at that point in time, but also because it actually handled Japanese at a point where the PC platform was basically restricted to ASCII or minor variants thereof. So, things remained fairly stable for some time. Underlying hardware changed - the 80286 introduced the ability to access more than a megabyte of address space and would promptly have broken a bunch of things except IBM came up with an utterly terrifying hack that bit me back in 2009, and which ended up sufficiently codified into Intel design that it was one mechanism for breaking the original XBox security. The first 286 PC even introduced a new keyboard controller that supported better keyboards but which remained backwards compatible with the original PC to avoid breaking software. Even when IBM launched the PS/2, the first significant rearchitecture of the PC platform with a brand new expansion bus and associated patents to prevent people cloning it without paying off IBM, they made sure that all the hardware was backwards compatible. For decades, PC compatibility meant not only supporting the officially supported interfaces, it meant supporting the underlying hardware. This is what made it possible to ship install media that was expected to work on any PC, even if you d need some additional media for hardware-specific drivers. It s something that still distinguishes the PC market from the ARM desktop market. But it s not as true as it used to be, and it s interesting to think about whether it ever was as true as people thought. Let s take an extreme case. If I buy a modern laptop, can I run 1981-era DOS on it? The answer is clearly no. First, modern systems largely don t implement the legacy BIOS. The entire abstraction layer that DOS relies on isn t there, having been replaced with UEFI. When UEFI first appeared it generally shipped with a Compatibility Services Module, a layer that would translate BIOS interrupts into UEFI calls, allowing vendors to ship hardware with more modern firmware and drivers without having to duplicate them to support older operating systems1. Is this system PC compatible? By the strictest of definitions, no. Ok. But the hardware is broadly the same, right? There s projects like CSMWrap that allow a CSM to be implemented on top of stock UEFI, so everything that hits BIOS should work just fine. And well yes, assuming they implement the BIOS interfaces fully, anything using the BIOS interfaces will be happy. But what about stuff that doesn t? Old software is going to expect that my Sound Blaster is going to be on a limited set of IRQs and is going to assume that it s going to be able to install its own interrupt handler and ACK those on the interrupt controller itself and that s really not going to work when you have a PCI card that s been mapped onto some APIC vector, and also if your keyboard is attached via USB or SPI then reading it via the CSM will work (because it s calling into UEFI to get the actual data) but trying to read the keyboard controller directly won t2, so you re still actually relying on the firmware to do the right thing but it s not, because the average person who wants to run DOS on a modern computer owns three fursuits and some knee length socks and while you are important and vital and I love you all you re not enough to actually convince a transglobal megacorp to flip the bit in the chipset that makes all this old stuff work. But imagine you are, or imagine you re the sort of person who (like me) thinks writing their own firmware for their weird Chinese Thinkpad knockoff motherboard is a good and sensible use of their time - can you make this work fully? Haha no of course not. Yes, you can probably make sure that the PCI Sound Blaster that s plugged into a Thunderbolt dock has interrupt routing to something that is absolutely no longer an 8259 but is pretending to be so you can just handle IRQ 5 yourself, and you can probably still even write some SMM code that will make your keyboard work, but what about the corner cases? What if you re trying to run something built with IBM Pascal 1.0? There s a risk that it ll assume that trying to access an address just over 1MB will give it the data stored just above 0, and now it ll break. It d work fine on an actual PC, and it won t work here, so are we PC compatible? That s a very interesting abstract question and I m going to entirely ignore it. Let s talk about PC graphics3. The original PC shipped with two different optional graphics cards - the Monochrome Display Adapter and the Color Graphics Adapter. If you wanted to run games you were doing it on CGA, because MDA had no mechanism to address individual pixels so you could only render full characters. So, even on the original PC, there was software that would run on some hardware but not on other hardware. Things got worse from there. CGA was, to put it mildly, shit. Even IBM knew this - in 1984 they launched the PCjr, intended to make the PC platform more attractive to home users. As well as maybe the worst keyboard ever to be associated with the IBM brand, IBM added some new video modes that allowed displaying more than 4 colours on screen at once4, and software that depended on that wouldn t display correctly on an original PC. Of course, because the PCjr was a complete commercial failure, it wouldn t display correctly on any future PCs either. This is going to become a theme. There s never been a properly specified PC graphics platform. BIOS support for advanced graphics modes5 ended up specified by VESA rather than IBM, and even then getting good performance involved hitting hardware directly. It wasn t until Microsoft specced DirectX that anything was broadly usable even if you limited yourself to Microsoft platforms, and this was an OS-level API rather than a hardware one. If you stick to BIOS interfaces then CGA-era code will work fine on graphics hardware produced up until the 20-teens, but if you were trying to hit CGA hardware registers directly then you re going to have a bad time. This isn t even a new thing - even if we restrict ourselves to the authentic IBM PC range (and ignore the PCjr), by the time we get to the Enhanced Graphics Adapter we re not entirely CGA compatible. Is an IBM PC/AT with EGA PC compatible? You d likely say yes , but there s software written for the original PC that won t work there. And, well, let s go even more basic. The original PC had a well defined CPU frequency and a well defined CPU that would take a well defined number of cycles to execute any given instruction. People could write software that depended on that. When CPUs got faster, some software broke. This resulted in systems with a Turbo Button - a button that would drop the clock rate to something approximating the original PC so stuff would stop breaking. It s fine, we d later end up with Windows crashing on fast machines because hardware details will absolutely bleed through. So, what s a PC compatible? No modern PC will run the DOS that the original PC ran. If you try hard enough you can get it into a state where it ll run most old software, as long as it doesn t have assumptions about memory segmentation or your CPU or want to talk to your GPU directly. And even then it ll potentially be unusable or crash because time is hard. The truth is that there s no way we can technically describe a PC Compatible now - or, honestly, ever. If you sent a modern PC back to 1981 the media would be amazed and also point out that it didn t run Flight Simulator. PC Compatible is a socially defined construct, just like Woman . We can get hung up on the details or we can just chill.
  1. Windows 7 is entirely happy to boot on UEFI systems except that it relies on being able to use a BIOS call to set the video mode during boot, which has resulted in things like UEFISeven to make that work on modern systems that don t provide BIOS compatibility
  2. Back in the 90s and early 2000s operating systems didn t necessarily have native drivers for USB input devices, so there was hardware support for trapping OS accesses to the keyboard controller and redirecting that into System Management Mode where some software that was invisible to the OS would speak to the USB controller and then fake a response anyway that s how I made a laptop that could boot unmodified MacOS X
  3. (my name will not be Wolfwings Shadowflight)
  4. Yes yes ok 8088 MPH demonstrates that if you really want to you can do better than that on CGA
  5. and by advanced we re still talking about the 90s, don t get excited

3 January 2026

Benjamin Mako Hill: Effects of Algorithmic Flagging on Fairness: Quasi-experimental Evidence from Wikipedia

Note: I have not published blog posts about my academic papers over the past few years. To ensure that my blog contains a more comprehensive record of my published papers and to surface these for folks who missed them, I will be periodically (re)publishing blog posts about some older published projects. This particular post is closely based on a previously published post by Nate TeBlunthuis from the Community Data Science Blog. Many online platforms are adopting AI and machine learning as a tool to maintain order and high-quality information in the face of massive influxes of user-generated content. Of course, AI algorithms can be inaccurate, biased, or unfair. How do signals from AI predictions shape the fairness of online content moderation? How can we measure an algorithmic flagging system s effects? In our paper published at CSCW, Nate TeBlunthuis, together with myself and Aaron Halfaker, analyzed the RCFilters system: an add-on to Wikipedia that highlights and filters edits that a machine learning algorithm called ORES identifies as likely to be damaging to Wikipedia. This system has been deployed on large Wikipedia language editions and is similar to other algorithmic flagging systems that are becoming increasingly widespread. Our work measures the causal effect of being flagged in the RCFilters user interface.

Screenshot of Wikipedia edit metadata on Special:RecentChanges with RCFilters enabled. Highlighted edits with a colored circle to the left side of other metadata are flagged by ORES. Different circle and highlight colors (white, yellow, orange, and red in the figure) correspond to different levels of confidence that the edit is damaging. RCFilters does not specifically flag edits by new accounts or unregistered editors, but does support filtering changes by editor types.
Our work takes advantage of the fact that RCFilters, like many algorithmic flagging systems, create discontinuities in the relationship between the probability that a moderator should take action and whether a moderator actually does. This happens because the output of machine learning systems like ORES is typically a continuous score (in RCFilters, an estimated probability that a Wikipedia edit is damaging), while the flags (in RCFilters, the yellow, orange, or red highlights) are either on or off and are triggered when the score crosses some arbitrary threshold. As a result, edits slightly above the threshold are both more visible to moderators and appear more likely to be damaging than edits slightly below. Even though edits on either side of the threshold have virtually the same likelihood of truly being damaging, the flagged edits are substantially more likely to be reverted. This fact lets us use a method called regression discontinuity to make causal estimates of the effect of being flagged in RCFilters.
Charts showing the probability that an edit will be reverted as a function of ORES scores in the neighborhood of the discontinuous threshold that triggers the RCfilters flag. The jump in the increase in reversion chances is larger for registered editors compared to unregistered editors at both thresholds.
To understand how this system may affect the fairness of Wikipedia moderation, we estimate the effects of flagging on edits on different groups of editors. Comparing the magnitude of these estimates lets us measure how flagging is associated with several different definitions of fairness. Surprisingly, we found evidence that these flags improved fairness for categories of editors that have been widely perceived as troublesome particularly unregistered (anonymous) editors. This occurred because flagging has a much stronger effect on edits by the registered than on edits by the unregistered. We believe that our results are driven by the fact that algorithmic flags are especially helpful for finding damage that can t be easily detected otherwise. Wikipedia moderators can see the editor s registration status in the recent changes, watchlists, and edit history. Because unregistered editors are often troublesome, Wikipedia moderators attention is often focused on their contributions, with or without algorithmic flags. Algorithmic flags make damage by registered editors (in addition to unregistered editors) much more detectable to moderators and so help moderators focus on damage overall, not just damage by suspicious editors. As a result, the algorithmic flagging system decreases the bias that moderators have against unregistered editors. This finding is particularly surprising because the ORES algorithm we analyzed was itself demonstrably biased against unregistered editors (i.e., the algorithm tended to greatly overestimate the probability that edits by these editors were damaging). Despite the fact that the algorithms were biased, their introduction could still lead to less biased outcomes overall. Our work shows that although it is important to design predictive algorithms to avoid such biases, it is equally important to study fairness at the level of the broader sociotechnical system. Since we first published a preprint of our paper, a follow-up piece by Leijie Wang and Haiyi Zhu replicated much of our work and showed that differences between different Wikipedia communities may be another important factor driving the effect of the system. Overall, this work suggests that social signals and social context can interact with algorithmic signals, and together these can influence behavior in important and unexpected ways.

The full citation for the paper is: TeBlunthuis, Nathan, Benjamin Mako Hill, and Aaron Halfaker. 2021. Effects of Algorithmic Flagging on Fairness: Quasi-Experimental Evidence from Wikipedia. Proceedings of the ACM on Human-Computer Interaction 5 (CSCW): 56:1-56:27. https://doi.org/10.1145/3449130.

We have also released replication materials for the paper, including all the data and code used to conduct the analysis and compile the paper itself.

Louis-Philippe V ronneau: 2025 A Musical Retrospective

2026 already! The winter weather here has really been beautiful and I always enjoy this time of year. Writing this yearly musical retrospective has now become a beloved tradition of mine1 and I enjoy retracing the year's various events through albums I listened to and concerts I went to. Albums In 2025, I added 141 new albums to my collection, around 60% more than last year's haul. I think this might have been too much? I feel like I didn't have time to properly enjoy all of them and as such, I decided to slow down my acquisition spree sometimes in early December, around the time I normally do the complete opposite. This year again, I bought the vast majority of my music on Bandcamp. Most of the other albums I bought as CDs and ripped them.
Try to Win - Stay True
Shotcallers - Tales of Urban Tribes
Lut ce Borgia - Long Live Summer
Lut ce Borgia - Choose
Senza Frontiere - Nonce futuro senza m moria
Self Control - Self Control Session #19 -5-23
Maldita - Un Mundo En Demencia
Bambu - They re Burning the Boats
Paradise Lost - Ascension
Krigsgrav - Stormcaller
Nobro - Live Your Truth Shred Some Gnar
Nobro - Sick Hustle
Balm Squad - Dark Times album
Corrigan Fest - La victoire en chantant
The Planet Smashers - Attack of the Planet Smashers
Extra a humana - Bailando en la oscuridad
Ben Prunty - FTL - Lost Ship EP
Faze - Big Upsetter
Faze - Struggling to Enjoy Ourselves While the World Slowly Implodes
Faze - Songs by Faze
Faze - Content
Blackbraid - Blackbraid III
UZU - qui la libert -
Vulgaires Machins - Contempler l'ab me
The Planet Smashers - On the Dancefloor
Self Control - Still... Can't Bring Us Down
Propagandhi - At Peace
Propagandhi - Today's Empires, Tomorrow's Ashes
Propagandhi - Potemkin City Limits
Propagandhi - Failed States
Rogue Steady Orchestra - Fehlstaat
Balm Squad - Point of No Return
Calva Louise - Edge of the Abyss
Fallujah - Xenotaph
Impureza - Alc zares
Propagandhi - Supporting Caste
Phantom Spell - Immortal's Requiem
In Mourning - The Immortal
Dava ! - Premi re Pression
22 Longs Riffs - Contre Courant
Insania - The Great Apocalypse
Phantom Spell - Heather & Hearth
Tenaz - Existencia Individual
B ton Arm - Renaissance
Messa - The Spin
Jo Corbeau - Yeba A oli
Jo Corbeau - Roi des rois
Lord Kossity - Everlord
Mo'Kalamity - Deeper Revolution
Mo'Kalamity - Freedom of Soul
Mo'Kalamity - Warriors of Light
Mo'Kalamity - Shine
Death Proof - Parallel World
Death Proof - Haunted
Death Proof - Do What You Want
The Buzzers - Cutting It Close
The Buzzers - Bloody Minded EP
Fuerza Bruta - Somos El Mal
Total Nada - Aqu y Ahora
Regress 86 - Primitive Stomp
Petrol Girls - Talk of Violence
Petrol Girls - Baby
Achigan - Les proc s condamn s
Atomic Threat - Nuke You
Atomic Threat - Atomic Threat
Atomic Threat - Nuke Us
Les Ordures Ioniques - Se so lagent
Les Ordures Ioniques - Je m'en viens so l
Les Ordures Ioniques - Les Ordures Ioniques
Nephylim - Circuition
Fighting Chance - Thus Hopes Fades
Capable! - Br le Br le Br le \ Hot-dog chicane
Strength Thru Oi!
Oi! Oi! That's Yer Lot!
Carry On Oi
Oi! The Album
Tulaviok - D che la Ch'touille
Tulaviok - Q-Sec
The Abyssinians - Satta Amasagana
Evil Conduct - Sorry... No
Grima - Nightside
Lambrini Girls - You're Welcome
Lambrini Girls - Who Let the Dogs Out
Dead Pioneers - Dead Pioneers
Dead Pioneers - PO$T AMERICAN
Havukruunu - Tavastland
Mitch Murder - Then Again
Polyrhythmics - Filter System
Ultra Razzia - Ultra Razzia - Armistice Split
The Uncouth - S\T
Ain't Right - Pils Session #21
Kardashev - The Baring of Shadows
Fucked Up - Hidden World
Fucked Up - Epics in Minutes
Liberty & Justice - Empire Down Split
Liberty & Justice - Locked In
Liberty & Justice - Top of the Scrapheap
AUS!Funkt - My Way - Interception
AUS!Funkt - That's a Fact
Dynamite Express - Dynamite Express
Mortier - L'Ennui
Union Thugs - Pils Session #16
Jag Arreola - Unmasked
Le Matos - Turbo Kid Original Game Soundtrack
Union Thugs - Coup de Grisou
BLEMISH - Pils Session #23
The Halo Effect - March of the Unheard
Polyrhythmics - Man from the Future
Crachat - Fille Poup e
Over the Hill - Older, Not Wiser
Taxi Girls - Rainy
Taxi Girls - Coming Up Roses
Brass Lip - vol 2
Brass Lip - vol 1
The Prowlers - On the Run
Ain't Right - Careful What You Wish For
Stranglehold - Trouble
Stranglehold - Stranglehold
B ton Arm - Collection
Flatfoot 56 - Toil
Flatfoot 56 - Blackthorn
Blitz - Voice of a Generation
Bad Bunny - DeB TiRAR M S FOToS
Ultra Razzia - Ultra Razzia
Force Majeure - Force Majeure
Alternate Action - Thin Line
Saturday Night Satan - All Things Black
Alternate Action - Violent Crime EP
The Stress - The War Has Just Begun
The Stress - Never Die
The Stress - More Than You'll Ever Be
Spleen - ll
Spleen - Demo
Ultra Razzia - Collection
Big Boss - Skinhead revenge
Battery March - Boston Eats Its Young \ Futur pour eux EP
Workers Comp - Demo
Khassarat - How Many More Must Perish in the Pursuit of Hope?
Huntsmen - The Dry Land
Julie Christmas - Ridiculous and Full of Blood
Gaerea - Coma
Kanonenfieber - Die Urkatastrophe
Concerts In 2025, I went to the following 25 (!!) concerts: Although I haven't touched metalfinder's code in a good while, my instance still works very well and I get the occasional match when a big-name artist in my collection comes in town. Most the venues that advertise on Bandsintown are tied to Ticketmaster though, which means most underground artists (i.e. most of the music I listen to) end up playing elsewhere. As such, shout out again to the Gancio project and to the folks running the Montreal instance. It continues to be a smash hit and most of the interesting concerts end up being advertised there. See you all in 2026!
  1. see the 2022, 2023 and 2024 entries

2 January 2026

Joachim Breitner: Seemingly impossible programs in Lean

In 2007, Martin Escardo wrote a often-read blog post about Seemingly impossible functional programs . One such seemingly impossible function is find, which takes a predicate on infinite sequences of bits, and returns an infinite sequence for which that predicate hold (unless the predicate is just always false, in which case it returns some arbitrary sequence). Inspired by conversations with and experiments by Massin Guerdi at the dinner of LeaningIn 2025 in Berlin (yes, this blog post has been in my pipeline for far too long), I wanted to play around these concepts in Lean. Let s represent infinite sequences of bits as functions from Nat to Bit, and give them a nice name, and some basic functionality, including a binary operator for consing an element to the front: 
import Mathlib.Data.Nat.Find
abbrev Bit := Bool
def Cantor : Type := Nat   Bit
def Cantor.head (a : Cantor) : Bit := a 0
def Cantor.tail (a : Cantor) : Cantor := fun i => a (i + 1)
@[simp, grind] def Cantor.cons (x : Bit) (a : Cantor) : Cantor
    0 => x
    i+1 => a i
infix:60 " # " => Cantor.cons
With this in place, we can write Escardo s function in Lean. His blog post discusses a few variants; I ll focus on just one of them: 
mutual
  partial def forsome (p : Cantor   Bool) : Bool :=
    p (find p)
  partial def find (p : Cantor   Bool) : Cantor :=
    have b := forsome (fun a => p (true # a))
    (b # find (fun a => p (b # a)))
end
We define find together with forsome, which checks if the predicate p holds for any sequence. Using that find sets the first element of the result to true if there exists a squence starting with true, else to false, and then tries to find the rest of the sequence. It is a bit of a brian twiter that this code works, but it does: 
def fifth_false : Cantor   Bool := fun a => not (a 5)
/-- info: [true, true, true, true, true, false, true, true, true, true] -/
#guard_msgs in
#eval List.ofFn (fun (i : Fin 10) => find fifth_false i)
Of course, in Lean we don t just want to define these functions, but we want to prove that they do what we expect them to do. Above we defined them as partial functions, even though we hope that they are not actually partial: The partial keyword means that we don t have to do a termination proof, but also that we cannot prove anything about these functions. So can we convince Lean that these functions are total after all? We can, but it s a bit of a puzzle, and we have to adjust the definitions. First of all, these seemingly impossible functions are only possible because we assume that the predicate we pass to it, p, is computable and total. This is where the whole magic comes from, and I recommend to read Escardo s blog posts and papers for more on this fascinating topic. In particular, you will learn that a predicate on Cantor that is computable and total necessarily only looks at some initial fragment of the sequence. The length of that prefix is called the modulus . So if we hope to prove termination of find and forsome, we have to restrict their argument p to only such computable predicates. To that end I introduce HasModulus and the subtype of predicates on Cantor that have such a modulus: 
-- Extensional (!) modulus of uniform continuity
def HasModulus (p : Cantor    ) :=   n,   a b : Cantor, (  i < n, a i = b i)   p a = p b
@[ext] structure CantorPred where
  pred : Cantor   Bool
  hasModulus : HasModulus pred
The modulus of such a predicate is now the least prefix lenght that determines the predicate. In particular, if the modulus is zero, the predicate is constant: 
namespace CantorPred
variable (p : CantorPred)
noncomputable def modulus : Nat :=
  open Classical in Nat.find p.hasModulus
theorem eq_of_modulus :  a b : Cantor, (  i < p.modulus, a i = b i)   p a = p b := by
  open Classical in
  unfold modulus
  exact Nat.find_spec p.hasModulus
theorem eq_of_modulus_eq_0 (hm : p.modulus = 0) :   a b, p a = p b := by
  intro a b
  apply p.eq_of_modulus
  simp [hm]
Because we want to work with CantorPred and not Cantor Bool I have to define some operations on that new type; in particular the cons element before predicate operation that we saw above in find: 
def comp_cons (b : Bit) : CantorPred where
  pred := fun a => p (b # a)
  hasModulus := by
    obtain  n, h_n  := p.hasModulus
    cases n with
      zero => exists 0; grind
      succ m =>
      exists m
      intro a b heq
      simp
      apply h_n
      intro i hi
      cases i
        rfl
        grind
@[simp, grind =] theorem comp_cons_pred (x : Bit) (a : Cantor) :
  (p.comp_cons x) a = p (x # a) := rfl
For this operation we know that the modulus decreases (if it wasn t already zero): 
theorem comp_cons_modulus (x : Bit) :
    (p.comp_cons x).modulus   p.modulus - 1 := by
  open Classical in
  apply Nat.find_le
  intro a b hab
  apply p.eq_of_modulus
  cases hh : p.modulus
    simp
    intro i hi
    cases i
      grind
      grind
grind_pattern comp_cons_modulus => (p.comp_cons x).modulus
We can rewrite the find function above to use these operations: 
mutual
  partial def forsome (p : CantorPred) : Bool := p (find p)
  partial def find (p : CantorPred) : Cantor := fun i =>
    have b := forsome (p.comp_cons true)
    (b # find (p.comp_cons b)) i
end
I have also eta-expanded the Cantor function returned by find; there is now a fun i => i around the body. We ll shortly see why that is needed. Now have everything in place to attempt a termination proof. Before we do that proof, we could step back and try to come up with an informal termination argument. We can express all this reasoning as a termination measure in the form of a lexicographic triple. The 0 and 1 in the middle component mean that for zero modulus, we can call forsome from find for free . 
mutual
  def forsome (p : CantorPred) : Bool := p (find p)
  termination_by (p.modulus, if p.modulus = 0 then 0 else 1, 0)
  decreasing_by grind
  def find (p : CantorPred) : Cantor := fun i =>
    have b := forsome (p.comp_cons true)
    (b # find (p.comp_cons b)) i
  termination_by i => (p.modulus, if p.modulus = 0 then 1 else 0, i)
  decreasing_by all_goals grind
end
The termination proof doesn t go through just yet: Lean is not able to see that (_ # p) i will call p with i - 1, and it does not see that p (find p) only uses find p if the modulus of p is non-zero. We can use the wf_preprocess feature to tell it about that: The following theorem replaces a call to p f, where p is a function parameter, with the slightly more complex but provably equivalent expression on the right, where the call to f is no in the else branch of an if-then-else and thus has p.modulus = 0 in scope: 
@[wf_preprocess]
theorem coe_wf (p : CantorPred) :
    (wfParam p) f = p (if _ : p.modulus = 0 then fun _ => false else f) := by
  split
  next h => apply p.eq_of_modulus_eq_0 h
  next => rfl
And similarly we replace (_ # p) i with a variant that extend the context with information on how p is called: 
def cantor_cons' (x : Bit) (i : Nat) (a :   j, j + 1 = i   Bit) : Bit :=
  match i with
    0 => x
    j + 1 => a j (by grind)
@[wf_preprocess] theorem cantor_cons_congr (b : Bit) (a : Cantor) (i : Nat) :
  (b # a) i = cantor_cons' b i (fun j _ => a j) := by cases i <;> rfl
After these declarations, the above definition of forsome and find goes through! It remains to now prove that they do what they should, by a simple induction on the modulus of p: 
@[simp, grind =] theorem tail_cons_eq (a : Cantor) : (x # a).tail = a := by
  funext i; simp [Cantor.tail, Cantor.cons]
@[simp, grind =] theorem head_cons_tail_eq (a : Cantor) : a.head # a.tail = a := by
  funext i; cases i <;> rfl
theorem find_correct (p : CantorPred) (h_exists :   a, p a) : p (find p) := by
  by_cases h0 : p.modulus = 0
    obtain  a, h_a  := h_exists
    rw [  h_a]
    apply p.eq_of_modulus_eq_0 h0
    rw [find.eq_unfold, forsome.eq_unfold]
    dsimp -zeta
    extract_lets b
    change p (_ # _)
    by_cases htrue :   a, p (true # a)
    next =>
      have := find_correct (p.comp_cons true) htrue
      grind
    next =>
      have : b = false := by grind
      clear_value b; subst b
      have hfalse :   a, p (false # a) := by
        obtain  a, h_a  := h_exists
        cases h : a.head
          exists Cantor.tail a
          grind
          exfalso
          apply htrue
          exists Cantor.tail a
          grind
      clear h_exists
      exact find_correct (p.comp_cons false) hfalse
termination_by p.modulus
decreasing_by all_goals grind
theorem forsome_correct (p : CantorPred) :
    forsome p   (  a, p a) where
  mp hfind := by unfold forsome at hfind; exists find p
  mpr hex := by unfold forsome; exact find_correct p hex
This is pretty nice! However there is more to do. For example, Escardo has a massively faster variant of find that we can implement as a partial function in Lean: 
def findBit (p : Bit   Bool) : Bit :=
  if p false then false else true
def branch (x : Bit) (l r : Cantor) : Cantor :=
  fun n =>
    if n = 0      then x
    else if 2   n then r ((n - 2) / 2)
                  else l ((n - 1) / 2)
mutual
  partial def forsome (p : Cantor -> Bool) : Bool :=
    p (find p)
  partial def find (p : Cantor -> Bool) : Cantor :=
    let x := findBit (fun x => forsome (fun l => forsome (fun r => p (branch x l r))))
    let l := find (fun l => forsome (fun r => p (branch x l r)))
    let r := find (fun r => p (branch x l r))
    branch x l r
end
But can we get this past Lean s termination checker? In order to prove that the modulus of p is decreasing, we d have to know that, for example, find (fun r => p (branch x l r)) is behaving nicely. Unforunately, it is rather hard to do termination proof for a function that relies on the behaviour of the function itself. So I ll leave this open as a future exercise. I have dumped the code for this post at https://github.com/nomeata/lean-cantor.

1 January 2026

Dima Kogan: Using libpython3 without linking it in; and old Python, g++ compatibility patches

I just released mrcal 2.5; much more about that in a future post. Here, I'd like to talk about some implementation details.

libpython3 and linking
mrcal is a C library and a Python library. Much of mrcal itself interfaces the C and Python libraries. And it is common for external libraries to want to pass Python mrcal.cameramodel objects to their C code. The obvious way to do this is in a converter function in an O& argument to PyArg_ParseTupleAndKeywords(). I wrote this mrcal_cameramodel_converter() function, which opened a whole can of worms when thinking about the compiling and linking and distribution of this thing. mrcal_cameramodel_converter() is meant to be called by code that implements Python-wrapping of C code. This function will be called by the PyArg_ParseTupleAndKeywords() Python library function, and it uses the Python C API itself. Since it uses the Python C API, it would normally link against libpython. However:
  • The natural place to distribute this is in libmrcal.so, but this library doesn't touch Python, and I'd rather not pull in all of libpython for this utility function, even in the 99% case when that function won't even be called
  • In some cases linking to libpython actually breaks things, so I never do that anymore anyway. This is fine: since this code will only ever be called by libpython itself, we're guaranteed that libpython will already be loaded, and we don't need to ask for it.
OK, let's not link to libpython then. But if we do that, we're going to have unresolved references to our libpython calls, and the loader will complain when loading libmrcal.so, even if we're not actually calling those functions. This has an obvious solution: the references to the libpython calls should be marked weak. That won't generate unresolved-reference errors, and everything will be great. OK, how do we mark things weak? There're two usual methods:
  1. We mark the declaration (or definition?) or the relevant functions with __attribute__((weak))
  2. We weaken the symbols after the compile with objcopy --weaken.
Method 1 is more work: I don't want to keep track of what Python API calls I'm actually making. This is non-trivial, because some of the Py_...() invocations in my code are actually macros that call functions internally that I must weaken. Furthermore, all the functions are declared in Python.h that I don't control. I can re-declare stuff with __attribute__((weak)), but then I have to match the prototypes. And I have to hope that re-declaring these will make __attribute__((weak)) actually work. So clearly I want method 2. I implemented it: 
python-cameramodel-converter.o: %.o:%.c
        $(c_build_rule); mv $@ _$@
        $(OBJCOPY) --wildcard --weaken-symbol='Py*' --weaken-symbol='_Py*' _$@ $@
Works great on my machine! But doesn't work on other people's machines. Because only the most recent objcopy tool actually works to weaken references. Apparently the older tools only weaken definitions, which isn't useful to me, and the tool only started handling references very recently. Well that sucks. I guess I will need to mark the symbols with __attribute__((weak)) after all. I use the nm tool to find the symbols that should be weakened, and I apply the attribute with this macro: 
#define WEAKEN(f) extern __typeof__(f) f __attribute__((weak));
The prototypes are handled by __typeof__. So are we done? With gcc, we are done. With clang we are not done. Apparently this macro does not weaken symbols generated by inline function calls if using clang I have no idea if this is a bug. The Python internal machinery has some of these, so this doesn't weaken all the symbols. I give up on the people that both have a too-old objcopy and are using clang, and declare victory. So the logic ends up being:
  1. Compile
  2. objcopy --weaken
  3. nm to find the non-weak Python references
  4. If there aren't any, our objcopy call worked and we're done!
  5. Otherwise, compile again, but explicitly asking to weaken those symbols
  6. nm again to see if the compiler didn't do it
  7. If any non-weak references still remain, complain and give up.
Whew. This logic appears here and here. There were even more things to deal with here: calling nm and objcopy needed special attention and build-system support in case we were cross-building. I took care of it in mrbuild. This worked for a while. Until the converter code started to fail. Because .

Supporting old Python
. I was using PyTuple_GET_ITEM(). This is a macro to access PyTupleObject data. So the layout of PyTupleObject ended up encoded in libmrcal.so. But apparently this wasn't stable, and changed between Python3.13 and Python3.14. As described above, I'm not linking to libpython, so there's no NEEDED tag to make sure we pull in the right version. The solution was to call the PyTuple_GetItem() function instead. This is unsatisfying, and means that in theory other stuff here might stop working in some Python 3.future, but I'm ready to move on for now. There were other annoying gymnastics that had to be performed to make this work with old-but-not-super old tooling. The Python people deprecated PyModule_AddObject(), and added PyModule_Add() as a replacement. I want to support Pythons before and after this happened, so I needed some if statements. Today the old function still works, but eventually it will stop, and I will have needed to do this typing sooner or later.

Supporting old C++ compilers
mrcal is a C project, but it is common for people to want to #include the headers from C++. I widely use C99 designated initializers (27-years old in C!), which causes issues with not-very-old C++ compilers. I worked around this initialization in one spot, and disabled it a feature for a too-old compiler in another spot. Fortunately, semi-recent tooling supports my usages, so this is becoming a non-issue as time goes on.

31 December 2025

Sergio Cipriano: Zero-Code Instrumentation of an Envoy TCP Proxy using eBPF

Zero-Code Instrumentation of an Envoy TCP Proxy using eBPF I recently had to debug an Envoy Network Load Balancer, and the options Envoy provides just weren't enough. We were seeing a small number of HTTP 499 errors caused by latency somewhere in our cloud, but it wasn't clear what the bottleneck was. As a result, each team had to set up additional instrumentation to catch latency spikes and figure out what was going on. My team is responsible for the LBaaS product (Load Balancer as a Service) and, of course, we are the first suspects when this kind of problem appear. Before going for the current solution, I read a lot of Envoy's documentation. It is possible to enable access logs for Envoy, but they don't provide the information required for this debug. This is an example of the output: 
[2025-12-08T20:44:49.918Z] "- - -" 0 - 78 223 1 - "-" "-" "-" "-" "172.18.0.2:8080"
I won't go into detail about the line above, since it's not possible to trace the request using access logs alone. Envoy also has OpenTelemetry tracing, which is perfect for understanding sources of latency. Unfortunatly, it is only available for Application Load Balancers. Most of the HTTP 499 were happening every 10 minutes, so we managed to get some of the requests with tcpdump, Wireshark and using http headers to filter the requests. This approach helped us reproduce and track down the problem, but it wasn't a great solution. We clearly needed better tools to catch this kind of issue the next time it happened. Therefore, I decided to try out OpenTelemetry eBPF Instrumentation, also referred to as OBI. I saw the announcement of Grafana Beyla before it was renamed to OBI, but I didn't have the time or a strong reason to try it out until now. Even then, I really liked the idea, and the possibility of using eBPF to solve this instrumentation problem had been in the back of my mind. OBI promises zero-code automatic instrumentation for Linux services using eBPF, so I put together a minimal setup to see how well it works.

Reproducible setup I used the following tools: Setting up a TCP Proxy with Envoy was straightforward: 
static_resources:
  listeners:
  - name: go_server_listener
    address:
      socket_address:
        address: 0.0.0.0
        port_value: 8000
    filter_chains:
    - filters:
      - name: envoy.filters.network.tcp_proxy
        typed_config:
          "@type": type.googleapis.com/envoy.extensions.filters.network.tcp_proxy.v3.TcpProxy
          stat_prefix: go_server_tcp
          cluster: go_server_cluster
  clusters:
  - name: go_server_cluster
    connect_timeout: 1s
    type: LOGICAL_DNS
    load_assignment:
      cluster_name: go_server_cluster
      endpoints:
      - lb_endpoints:
        - endpoint:
            address:
              socket_address:
                address: target-backend
                port_value: 8080
This is the simplest Envoy TCP proxy configuration: a listener on port 8000 forwarding traffic to a backend running on port 8080. For the backend, I used a basic Go HTTP server: 
package main

import (
    "fmt"
    "net/http"
)

func main()  
    http.Handle("/", http.FileServer(http.Dir(".")))

    server := http.Server Addr: ":8080" 

    fmt.Println("Starting server on :8080")
    panic(server.ListenAndServe())
Finally, I wrapped everything together with Docker Compose: 
services:
  autoinstrumenter:
    image: otel/ebpf-instrument:main
    pid: "service:envoy"
    privileged: true
    environment:
      OTEL_EBPF_TRACE_PRINTER: text
      OTEL_EBPF_OPEN_PORT: 8000

  envoy:
    image: envoyproxy/envoy:v1.33-latest
    ports:
      - 8000:8000
    volumes:
      - ./envoy.yaml:/etc/envoy/envoy.yaml 
    depends_on:
      - target-backend
  
  target-backend:
    image: golang:1.22-alpine
    command: go run /app/backend.go
    volumes:
      - ./backend.go:/app/backend.go:ro
    expose:
      - 8080
OBI should output traces to the standard output similar to this when a HTTP request is made to Envoy: 
2025-12-08 20:44:49.12884449 (305.572 s[305.572 s]) HTTPClient 200 GET /(/) [172.18.0.3 as envoy:36832]->[172.18.0.2 as localhost:8080] contentLen:78B responseLen:0B svc=[envoy generic] traceparent=[00-529458a2be271956134872668dc5ee47-6dba451ec8935e3e[06c7f817e6a5dae2]-01]
2025-12-08 20:44:49.12884449 (1.260901ms[366.65 s]) HTTP 200 GET /(/) [172.18.0.1 as 172.18.0.1:36282]->[172.18.0.3 as envoy:8000] contentLen:78B responseLen:223B svc=[envoy generic] traceparent=[00-529458a2be271956134872668dc5ee47-06c7f817e6a5dae2[0000000000000000]-01]
This is exactly what we needed, with zero-code. The above trace shows:
  • 2025-12-08 20:44:49.12884449: time of the trace.
  • (1.260901ms[366.65 s]): total response time for the request, with the actual internal execution time of the request (not counting the request enqueuing time).
  • HTTP 200 GET /: protocol, response code, HTTP method, and URL path.
  • [172.18.0.1 as 172.18.0.1:36282]->[172.18.0.3 as envoy:8000]: source and destination host:port. The initial request originates from my machine through the gateway (172.18.0.1), hits the Envoy (172.23.0.3), the proxy then forwards it to the backend application (172.23.0.2).
  • contentLen:78B: HTTP Content-Length. I used curl and the default request size for it is 78B.
  • responseLen:223B: Size of the response body.
  • svc=[envoy generic]: traced service.
  • traceparent: ids to trace the parent request. We can see that the Envoy makes a request to the target and this request has the other one as parent.
Let's add one more Envoy to show that it's also possible to track multiple services. 
  envoy1:
    image: envoyproxy/envoy:v1.33-latest
    ports:
      - 9000:9000
    volumes:
      - ./envoy1.yaml:/etc/envoy/envoy.yaml
    depends_on:
      - envoy
The new Envoy will listen on port 9000 and forward the request to the other Envoy listening on port 8000. Now we just need to change OBI open port variable to look at a range: 
OTEL_EBPF_OPEN_PORT: 8000-9000
And change the pid field of the autoinstrumenter service to use the host's PID namespace inside the container: 
pid: host
This is the output I got after one curl: 
2025-12-09 12:28:05.12912285 (2.202041ms[1.524713ms]) HTTP 200 GET /(/) [172.19.0.1 as 172.19.0.1:59030]->[172.19.0.5 as envoy:9000] contentLen:78B responseLen:223B svc=[envoy generic] traceparent=[00-69977bee0c2964b8fe53cdd16f8a9d19-856c9f700e73bf0d[0000000000000000]-01]
2025-12-09 12:28:05.12912285 (1.389336ms[1.389336ms]) HTTPClient 200 GET /(/) [172.19.0.5 as envoy:59806]->[172.19.0.4 as localhost:8000] contentLen:78B responseLen:0B svc=[envoy generic] traceparent=[00-69977bee0c2964b8fe53cdd16f8a9d19-caa7f1ad1c68fa77[856c9f700e73bf0d]-01]
2025-12-09 12:28:05.12912285 (1.5431ms[848.574 s]) HTTP 200 GET /(/) [172.19.0.5 as 172.19.0.5:59806]->[172.19.0.4 as envoy:8000] contentLen:78B responseLen:223B svc=[envoy generic] traceparent=[00-69977bee0c2964b8fe53cdd16f8a9d19-cbca9d64d3d26b40[caa7f1ad1c68fa77]-01]
2025-12-09 12:28:05.12912285 (690.217 s[690.217 s]) HTTPClient 200 GET /(/) [172.19.0.4 as envoy:34256]->[172.19.0.3 as localhost:8080] contentLen:78B responseLen:0B svc=[envoy generic] traceparent=[00-69977bee0c2964b8fe53cdd16f8a9d19-5502f7760ed77b5b[cbca9d64d3d26b40]-01]
2025-12-09 12:28:05.12912285 (267.9 s[238.737 s]) HTTP 200 GET /(/) [172.19.0.4 as 172.19.0.4:34256]->[172.19.0.3 as backend:8080] contentLen:0B responseLen:0B svc=[backend go] traceparent=[00-69977bee0c2964b8fe53cdd16f8a9d19-ac05c7ebe26f2530[5502f7760ed77b5b]-01]
Each log line represents a span belonging to the same trace (69977bee0c2964b8fe53cdd16f8a9d19). For readability, I ordered the spans by their traceparent relationship, showing the request's path as it moves through the system: from the client-facing Envoy, through the internal Envoy hop, and finally to the Go backend. You can see both server-side (HTTP) and client-side (HTTPClient) spans at each hop, along with per-span latency, source and destination addresses, and response sizes, making it easy to pinpoint where time is spent along the request chain. The log lines are helpful, but we need better ways to visualize the traces and the metrics generated by OBI. I'll share another setup that more closely reflects what we actually use.

Production setup I'll be using the following tools this time: The goal of this setup is to mirror an environment similar to what I used in production. This time, I've omitted the load balancer and shifted the emphasis to observability instead. setup diagram I will run three HTTP servers on port 8080: two inside Incus containers and one on the host machine. The OBI process will export metrics and traces to an OpenTelemetry Collector, which will forward traces to Jaeger and expose a metrics endpoint for Prometheus to scrape. Grafana will also be added to visualize the collected metrics using dashboards. The aim of this approach is to instrument only one of the HTTP servers while ignoring the others. This simulates an environment with hundreds of Incus containers, where the objective is to debug a single container without being overwhelmed by excessive and irrelevant telemetry data from the rest of the system. OBI can filter metrics and traces based on attribute values, but I was not able to filter by process PID. This is where the OBI Collector comes into play, it allows me to use a processor to filter telemetry data by the PID of the process being instrumented. These are the steps to reproduce this setup:
  1. Create the incus containers.
$ incus launch images:debian/trixie server01
Launching server01
$ incus launch images:debian/trixie server02
Launching server02
  1. Start the HTTP server on each container.
$ apt install python3 --update -y
$ tee /etc/systemd/system/server.service > /dev/null <<'EOF'
[Unit]
Description=Python HTTP server
After=network.target
[Service]
User=root
Group=root
Type=simple
ExecStart=/usr/bin/python3 -m http.server 8080
Restart=always
StandardOutput=journal
StandardError=journal
[Install]
WantedBy=multi-user.target
EOF
$ systemctl start server.service
  1. Start the HTTP server on the host.
$ python3 -m http.server 8080
  1. Start the Docker compose.
services:
  autoinstrumenter:
    image: otel/ebpf-instrument:main
    pid: host
    privileged: true
    environment:
      OTEL_EBPF_CONFIG_PATH: /etc/obi/obi.yml 
    volumes:
      - ./obi.yml:/etc/obi/obi.yml

  otel-collector:
    image: otel/opentelemetry-collector-contrib:0.98.0
    command: ["--config=/etc/otel-collector-config.yml"]
    volumes:
      - ./otel-collector-config.yml:/etc/otel-collector-config.yml
    ports:
      - "4318:4318" # Otel Receiver
      - "8889:8889" # Prometheus Scrape
    depends_on:
      - autoinstrumenter
      - jaeger
      - prometheus

  prometheus:
    image: prom/prometheus
    volumes:
      - ./prometheus.yml:/etc/prometheus/prometheus.yml
    ports:
      - "9090:9090" # Prometheus UI

  grafana:
    image: grafana/grafana
    restart: always
    environment:
      - GF_SECURITY_ADMIN_USER=admin
      - GF_SECURITY_ADMIN_PASSWORD=RandomString123!
    volumes:
      - ./grafana-ds.yml:/etc/grafana/provisioning/datasources/datasource.yml
    ports:
      - "3000:3000" # Grafana UI

  jaeger:
    image: jaegertracing/all-in-one
    container_name: jaeger
    ports:
      - "16686:16686" # Jaeger UI
      - "4317:4317"  # Jaeger OTLP/gRPC Collector
Here's what the configuration files look like:
  • obi.yml:
log_level: INFO
trace_printer: text

discovery:
  instrument:
    - open_ports: 8080

otel_metrics_export:
  endpoint: http://otel-collector:4318
otel_traces_export:
  endpoint: http://otel-collector:4318
  • prometheus.yml:
global:
  scrape_interval: 5s

scrape_configs:
  - job_name: 'otel-collector'
    static_configs:
      - targets: ['otel-collector:8889']
  • grafana-ds.yml:
apiVersion: 1

datasources:
  - name: Prometheus
    type: prometheus
    access: proxy
    url: http://prometheus:9090
    isDefault: true
  • otel-collector-config.yml:
receivers:
  otlp:
    protocols:
      http:
        endpoint: otel-collector:4318

exporters:
  otlp/jaeger:
    endpoint: jaeger:4317
    tls:
      insecure: true

  prometheus:
    endpoint: 0.0.0.0:8889
    namespace: default

service:
  pipelines:
    traces:
      receivers: [otlp]
      exporters: [otlp/jaeger]

    metrics:
      receivers: [otlp] 
      exporters: [prometheus]
We're almost there, the OpenTelemetry Collector is just missing a processor. To create the processor filter, we can look at the OBI logs to find the PID of the HTTP server being instrumented: 
autoinstrumenter-1    time=2025-12-30T19:57:17.593Z level=INFO msg="instrumenting process" component=discover.traceAttacher cmd=/usr/bin/python3.13 pid=297514 ino=460310 type=python service=""
autoinstrumenter-1    time=2025-12-30T19:57:18.320Z level=INFO msg="instrumenting process" component=discover.traceAttacher cmd=/usr/bin/python3.13 pid=310288 ino=722998 type=python service=""
autoinstrumenter-1    time=2025-12-30T19:57:18.512Z level=INFO msg="instrumenting process" component=discover.traceAttacher cmd=/usr/bin/python3.13 pid=315183 ino=2888480 type=python service=""
Which can also be obtained using standard GNU/Linux utilities: 
$ cat /sys/fs/cgroup/lxc.payload.server01/system.slice/server.service/cgroup.procs 
297514
$ cat /sys/fs/cgroup/lxc.payload.server02/system.slice/server.service/cgroup.procs 
310288
$ ps -aux   grep http.server
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
1000000   297514  0.0  0.1  32120 14856 ?        Ss   16:03   0:00 /usr/bin/python3 -m http.server 8080
1000000   310288  0.0  0.1  32120 10616 ?        Ss   16:09   0:00 /usr/bin/python3 -m http.server 8080
cipriano  315183  0.0  0.1 103476 11480 pts/3    S+   16:17   0:00 python -m http.server 8080
If we search for the PID in the OpenTelemetry Collector endpoint where Prometheus metrics are exposed, we can find the attribute values to filter on. 
$ curl http://localhost:8889/metrics   rg 297514
default_target_info host_id="148f400ad3ea",host_name="148f400ad3ea",instance="148f400ad3ea:297514",job="python3.13",os_type="linux",service_instance_id="148f400ad3ea:297514",service_name="python3.13",telemetry_sdk_language="python",telemetry_sdk_name="opentelemetry-ebpf-instrumentation",telemetry_sdk_version="main"  1
Now we just need to add the processor to the collector configuration: 
processors: # <--- NEW BLOCK
  filter/host_id:
    traces:
      span:
        - 'resource.attributes["service.instance.id"] == "148f400ad3ea:297514"'

service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: [filter/host_id] # <--- NEW LINE
      exporters: [otlp/jaeger]

    metrics:
      receivers: [otlp] 
      processors:  # <--- NEW BLOCK
        - filter/host_id
      exporters: [prometheus]
That's it! The processor will handle the filtering for us, and we'll only see traces and metrics from the HTTP server running in the server01 container. Below are some screenshots from Jaeger and Grafana: jaeger search will all traces one jaeger trace grafana request duration panel

Closing Notes I am still amazed at how powerful OBI can be. For those curious about the debug, we found out that a service responsible for the network orchestration of the Envoy containers was running netplan apply every 10 minutes because of a bug. Netplan apply causes interfaces to go down temporarily and this made the latency go above 500ms which caused the 499s.

Petter Reinholdtsen: Some of my 2025 free software activities

I guess it is about time I posted a new summary of the free software and open culture activites and projects I have been involved in the last year. The days have been so packed the last year that I have failed with my intention to post at least one blog post per month, so this summary became rather long. I am sorry about this. This year was the year I got tired of the lack of new releases of the multimedia related libraries published via Xiph, and I decided to wrap up the current state and make the releases myself. In a burst of activity early this year, I collected and tested patches, coordinated with other developers and finally made new tarballs and release announcement for theora, and new tarball releases for liboggz, kate and fishsound. This upstreamed several patches accumulated in Debian and other Linux distributions for the last 15 years or so. To change the world and the future, it is important to start with the kids, and one such avenue of change have been created by the current president of FSF Europe, Matthias Kirschner. He wrote a book for children, Ada & Zangemann, and I have been involved in its translation framework for the entire year. The source code has been transformed to Docbook and I have been conducting and coordinating translations into Norwegian Bokm l and Nynorsk, as well as preparing paper editions of the book and an animation movie with Norwegian voices. The Bokm l edition is very close to ready, and will be available early in 2026, and the movie release will follow shortly after this. I intend announce this on my blog and elsewhere when this happen. Please get in touch if you want to help spread the word about this book in Norwegian. I hope we can get the author to Norway when making the Norwegian releases. This year I continued a push for the system I made a few years ago to improve hardware dongle handling on Linux. The Isenkram system use hardware mapping information provided by relevant packages using the AppStream system to propose which Linux distribution packages to install on a given machine to support dongles like cameras, finger print readers, smart card readers, LEGO controllers, ECC memory and other hardware. I have followed up on the list of packages providing such mapping, either to get it into Debian or to upstream the necessary metadata. I am not sure if we are at a point where package maintainers on their own add such information to their packages, but there are Debian lintian reports suggesting it and I have send patches to all packages I am aware of that should include such mappings. Most of the patches are included in Debian now, only 27 was left the last time I checked. As part of my involvement with Debian, I continued my push to get all orphaned packages without a version control repository migrated to git. I am not sure how many packages I went through, but it was in the range of 200-300 packages. In addition to this I updated, sponsored, pushed maintainers for updates upstreamed patches for and fixed RC issues with battery-stats, bs1770gain, isenkram, libonvif, mfiutil, opensnitch, simplescreenrecorder, vlc-plugin-bittorrent and wakeonlan. I've also followed up LEGO related packages, dahdi support for Asterisk, llama.cpp and whisper.cpp in particular for the AMD GPU I was donated by AMD, as well as tried yet again to convince the upstream developers of the photogrammetric workstion e-foto to get their program into a state that could be included in Debian. As I do not buy into the story that it is great to expose oneself to the whims of and priorities of commercial entities to have access to cultural expressions like films and music, I still maintain a huge collection of movies. For this to work well, I have ended up as part of the people maintaining lsdvd upstream and wrapped up a new release fixing several crash bugs caused by DVDs with intentionally broken metadata, and introduced code to list a DVD ID in the lsdvd output. Related to this, I have also worked some add-ons for my main video and music player, and took over upstream maintenance of the Invidious add-on, which sadly stopped working for non-authenticated users when web scrapers made it impossible for Invidious installations to provide a open API, as well as contributed to the NRK and projector control add-ons. As part of my involvement in the Norwegian archiving community and standardisation work, we organised a Noark 5 workshop this spring discussing how to decide what to keep and what to delete in digital archives. We finally managed to apply for Noark 5 certification for the free software archive API Nikita, as well as worked to test and improve the performance of Nikita together with people on my day job at the university. Manufacturing using Free Software is still a focus for me, and I have continued my involved with the LinuxCNC community, organising a developer gathering this summer with the help and sponsoring from the initial start in 2023 from NUUG Foundation and sponsoring from Debian and Redpill-Linpro. We plan to repeat the event also in 2026, but this time NUUG Foundation have told us they do not want a role, so we have found another friendly organisation to handle the money. A popular machine controller with LinuxCNC is the MESA set of electronics, which is centred around a FPGA which now can be programmed using only Free Software. We discussed during this summers gathering how hard it would be to compile the current FPGA source using a Free Software tool chain, and I started looking into this, locating tools to transform the VHDL source into something the Yosys tool chain can handle. Still lot to do there, and I hope to get further next year. An important part of Free Software manufacturing is the ability to design parts and create programs that can be passed to machines making parts, also known as CAD/CAM. The most prominent project for this is FreeCAD, and I have been both pushing to get opencamlib integrated with it in Debian as well as fixing bugs in the handling of Fanuc controlled machines, do make it easier to generate instructions for machines I have access to. I expect to also continue this also next year. This year the UN conference Internet Governance Forum (IGF) was held in Norway, and I tried my best to get a stand for the Norwegian Unix Users Group (NUUG) there. Sadly the effort failed, due to lack of interest with the NUUG Board, but I was happy to see several members at least attend some of the activities related to IGF. Sadly to participate at IGF one need to hand over quite private information, so I decided not to participate in any of the closed forum events myself. Related to NUUG I have been a member of the election board proposing board member candidates to the general assembly, and been part of the program committee of the "Big Tech m vekk" (Big Tech must go away) festival organised by Attac in concert with NUUG and EFN. I've also assisted the Norwegian open TV channel Frikanalen with access to their machines located in a machine room at the university. Related to the University, I have become involved in a small team of students working to build and program robots for the Robocup@Home competition. For 2026 we also plan to use the new features of FreeCAD to make parts for the open hardware robot arm OpenArm. This is also the group that will handle the money for the LinuxCNC gathering in 2026. Also related to the university I was looking into the Linux security auditing system Falco earlier this year, making improvements to the detection rules. This activity is on hold at the moment, and do not expect to continue with this in 2026. I will most likely have to cut down a bit on my free software and open culture activities going forward, as NUUG Foundation, who have funded one day a week for such activities for several years no, sadly have decided they do not want to continue doing this. I am very grateful for their contributions over the years, both with freeing up time for me and supporting several events and projects where I have been involved or taken the initiative on. Now they are reorganizing with more focus on paperwork and applications. As usual, if you use Bitcoin and want to show your support of my activities, please send Bitcoin donations to my address 15oWEoG9dUPovwmUL9KWAnYRtNJEkP1u1b.

kpcyrd: 2025 wrapped

Same as last year, this is a summary of what I ve been up to throughout the year. See also the recap/retrospection published by my friends (antiz, jvoisin, orhun). Thanks to everybody who has been part of my human experience, past or present. Especially those who ve been closest.

30 December 2025

Utkarsh Gupta: FOSS Activites in December 2025

Here s my monthly but brief update about the activities I ve done in the FOSS world.

Debian
Whilst I didn t get a chance to do much, here are still a few things that I worked on:
  • Prepared security update for wordpress for trixie and bookworm.
  • A few discussions with the new DFSG team, et al.
  • Assited a few folks in getting their patches submitted via Salsa.
  • Mentoring for newcomers.
  • Moderation of -project mailing list.

Ubuntu
I joined Canonical to work on Ubuntu full-time back in February 2021. Whilst I can t give a full, detailed list of things I did, here s a quick TL;DR of what I did:
  • Successfully released Resolute Snapshot 2!
    • This one was also done without the ISO tracker and cdimage access.
    • I think this one went rather smooth. Let s see what we re able to do for snapshot 3.
  • Worked on removing GPG keys from the cdimage instance. That took a while, whew!
  • Assisted a bunch of folks with my Archive Admin and Release team hats to:
    • review NEW packages for Ubuntu Studio.
    • remove old binaries that are stalling transition and/or migration.
    • LTS requalification of Ubuntu flavours.
    • bootstrapping dotnet-10 packages for Stable Release Updates.
  • With that, we ve entered the EOY break. :)
    • I was anyway on vacation for majority of this month. ;)

Debian (E)LTS
This month I have worked 72 hours on Debian Long Term Support (LTS) and on its sister Extended LTS project and did the following things:

Released Security Updates
  • ruby-git: Multiple vulnerabilities leading to command line injection and improper path escaping.
  • ruby-sidekiq: Multiple vulnerabilities leading to Cross-site Scripting (XSS) and Denial of Service in Web UI.
  • python-apt: Vulnerability leading to crash via invalid nullptr dereference in TagSection.keys().
    • [LTS]: Fixed CVE-2025-6966 via 2.2.1.1 for bullseye. This has been released as DLA 4408-1.
    • [ELTS]: Fixed CVE-2025-6966 via 1.8.4.4 for buster and 1.4.4 for stretch. This has been released as ELA 1596-1.
    • All of this was coordinated b/w the Security team and Julian Andres Klode. Julian will take care of the stable uploads.
  • node-url-parse: Vulnerability allowing authorization bypass through specially crafted URL with empty userinfo and no host.
  • wordpress: Multiple vulnerabilities in WordPress core, leading to Sent Data & Cross-site Scripting.
  • usbmuxd: Privilege escalation vulnerability via path traversal in SavePairRecord command.
    • [LTS]: Fixed CVE-2025-66004 via 1.1.1-2+deb11u1 for bullseye. This has been released as DLA 4417-1.
    • [ELTS]: Fixed CVE-2025-66004 via 1.1.1~git20181007.f838cf6-1+deb10u1 for buster and 1.1.0-2+deb9u1 for stretch. This has been released as ELA 1599-1.
    • All of this was coordinated b/w the Security team and Yves-Alexis Perez. Yves will take care of the stable uploads.
  • gst-plugins-good1.0: Multiple vulnerabilities in isomp4 plugin leading to potential out-of-bounds reads and information disclosure.
  • postgresql-13: Multiple vulnerabilities including unauthorized schema statistics creation and integer overflow in libpq allocation calculations.
  • gst-plugins-base1.0: Multiple vulnerabilities in SubRip subtitle parsing leading to potential crashes and buffer issues.

Work in Progress
  • ceph: Affected by CVE-2024-47866, using the argument x-amz-copy-source to put an object and specifying an empty string as its content leads to the RGW daemon crashing, resulting in a DoS attack.
  • knot-resolver: Affected by CVE-2023-26249, CVE-2023-46317, and CVE-2022-40188, leading to Denial of Service.
  • adminer: Affected by CVE-2023-45195 and CVE-2023-45196, leading to SSRF and DoS, respectively.
  • u-boot: Affected by CVE-2025-24857, where boot code access control flaw in U-Boot allowing arbitrary code execution via physical access.
    • [ELTS]: As it s only affected the version in stretch, I ve started the work to find the fixing commits and prepare a backport. Not much progress there, I ll roll it over to January.
  • ruby-rack: There were multiple vulnerabilities reported in Rack, leading to DoS (memory exhaustion) and proxy bypass.
    • [ELTS]: After completing the work for LTS myself, Bastien picked it up for ELTS and reached out about an upstream regression and we ve been doing some exchanges. Bastien has done most of the work backporting the patches but needs a review and help backporting CVE-2025-61771.

Other Activities
  • Frontdesk from 01-12-2025 to 07-12-2025.
    • Auto EOL d a bunch of packages.
    • Marked CVE-2025-12084/python2.7 as end-of-life for bullseye, buster, and stretch.
    • Marked CVE-2025-12084/jython as end-of-life for bullseye.
    • Marked CVE-2025-13992/chromium as end-of-life for bullseye.
    • Marked apache2 CVEs as postponed for bullseye, buster, and stretch.
    • Marked CVE-2025-13654/duc as postponed for bullseye and buster.
    • Marked CVE-2025-32900/kdeconnect as ignored for bullseye.
    • Marked CVE-2025-12084/pypy3 as postponed for bullseye.
    • Marked CVE-2025-14104/util-linux as postponed for bullseye, buster, and stretch.
    • Marked several CVEs for fastdds as postponed for bullseye.
    • Marked several CVEs for pytorch as postponed for bullseye.
    • Marked CVE-2025-2486/edk2 as postponed for bullseye.
    • Marked CVE-2025-6172 7,9 /golang-1.15 as postponed for bullseye.
    • Marked CVE-2025-65637/golang-logrus as postponed for bullseye.
    • Marked CVE-2025-12385/qtdeclarative-opensource-src ,gles as postponed for bullseye, buster, and stretch.
    • Marked TEMP-0000000-D08402/rust-maxminddb as postponed for bullseye.
    • Added the following packages to d,e la-needed.txt:
      • liblivemedia, sogo.
    • During my triage, I had to make the bin/elts-eol script robust to determine the lts_admin repository - did a back and forth with Emilio about this on the list.
    • I sent a gentle reminder to the LTS team about the issues fixed in bullseye but not in bookworm via mailing list: https://lists.debian.org/debian-lts/2025/12/msg00013.html.
  • I claimed php-horde-css-parser to work on CVE-2020-13756 for buster and did almost all the work only to realize that the patch already existed in buster and the changelog confirmed that it was intentionally fixed.
    • After speaking with Andreas Henriksson, we figured that the CVE ID was missed when the ELA was generated and so I fixed that via 87afaaf19ce56123bc9508d9c6cd5360b18114ef and 5621431e84818b4e650ffdce4c456daec0ee4d51 in the ELTS security tracker to reflect the situation.
  • Participated in a thread which I started last month around using Salsa CI for E/LTS packages and if we plan to sunset it in favor of using Debusine. The plan for now is to keep it around as it s still beneficial and Debusine is still in its early phase.
  • Did a lot of back and forth with Helmut about debusine uploads on #debian-elts.
    • While debugging a failure in dcut uploads, I ran into an SSH compatibility issue on deb-master.freexian.com that could be fixed on the server-side. I shared all my findings to Freexian s sysadmin team.
    • A minimal fix on the server side would be one of: 
      PubkeyAcceptedAlgorithms -ssh-dss
      or explicitly restricting to modern algorithms, e.g.: 
      PubkeyAcceptedAlgorithms
ssh-ed25519,ecdsa-sha2-nistp256,rsa-sha2-512,rsa-sha2-256
  • Jelly on #debian-lts reported that all my DLA mails had broken GMail s DKIM signature. So I set up sending replies from @debian.org and that seems to have fixed that! \o/
  • [LTS] Attended a rather short monthly LTS meeting on Jitsi. Summary here.
  • [E/LTS] Monitored discussions on mailing lists, IRC, and all the documentation updates.
Until next time.
:wq for today.

28 December 2025

Jonathan Dowland: Our study, 2025

We re currently thinking of renovating our study/home office. I ll likely write more about that project. Embarking on it reminded me that I d taken a photo of the state of it nearly a year ago and forgot to post it, so here it is.
Home workspace, January 2025 Home workspace, January 2025
When I took that pic last January, it had been three years since the last one, and the major difference was a reduction in clutter. I've added a lava lamp (charity shop find) and Rob Sheridan print. We got rid of the PO NG chair (originally bought for breast feeding) so we currently have no alternate seating besides the desk chair. As much as I love my vintage mahogany writing desk, our current thinking is it s likely to go. I m exploring whether we could fit in two smaller desks: one main one for the computer, and another workbench for play: the synthesiser, Amiga, crafting and 3d printing projects, etc.

24 December 2025

Daniel Lange: Getting scanning to work with Gimp on Trixie

Trixie ships Gimp 3.0.4 and the 3.x series has gotten incompatible to XSane, the common frontend for scanners on Linux. Hence the maintainer, J rg Frings-F rst, has disabled the Gimp integration temporarily in response to a Debian bug #1088080. There seems to be no tracking bug for getting the functionality back but people have been commenting on Debian bug #993293 as that is ... loosely related :-). There are two options to get the Scanning functionality back in Trixie until this is properly resolved by an updated XSane in Debian (e.g. via trixie-backports): Lee Yingtong Li (RunasSudo) has created a Python script that calls XSane as a cli application and published it at https://yingtongli.me/git/gimp-xsanecli/. This worked okish for me but needed me to find the scan in /tmp/ a number of times. This is a good stop-gap script if you need to scan from Gimp $now and look for a quick solution. Upstream has completed the necessary steps to get XSane working as a Gimp 3.x plugin at https://gitlab.com/sane-project/frontend/xsane. Unfortunately compiling this is a bit involved but I made a version that can be dropped into /usr/local/bin or $HOME/bin and works alongside Gimp and the system-installed XSane. So:
  1. sudo apt install gimp xsane
  2. Download xsane-1.0.0-fit-003 (752kB, AMD64 executable for Trixie) and place it in /usr/local/bin (as root)
  3. sha256sum /usr/local/bin/xsane-1.0.0-fit-003
    # result needs to be af04c1a83c41cd2e48e82d04b6017ee0b29d555390ca706e4603378b401e91b2
  4. sudo chmod +x /usr/local/bin/xsane-1.0.0-fit-003
  5. # Link the executable into the Gimp plugin directory as the user running Gimp:
    mkdir -p $HOME/.config/GIMP/3.0/plug-ins/xsane/
    ln -s /usr/local/bin/xsane-1.0.0-fit-003 $HOME/.config/GIMP/3.0/plug-ins/xsane/
  6. Restart Gimp
  7. Scan from Gimp via File Create Acquire XSane
The source code for the xsane executable above is available under GPL-2 at https://gitlab.com/sane-project/frontend/xsane/-/tree/c5ac0d921606309169067041931e3b0c73436f00. This points to the last upstream commit from 27. September 2025 at the time of writing this blog article.

22 December 2025

Hellen Chemtai: Overcoming Challenges in OpenQA Images Testing: My Internship Journey

Hello there . Today will be an in depth review on my work with the Debian OpenQA images testing team. I will highlight the struggles that I have had so far during my Outreachy internship. The OpenQA images testing team uses OpenQA to automatically install images e.g. Gnome Images. The images are then tested using tests written in Perl. My current tasks include speech install and capture all audio. I am also installing Live Gnome image to Windows using BalenaEtcher then testing it. A set of similar tasks will also be collaborated on. While working on tasks, I have to go through the guides. I also learn how Perl works so as to edit and create tests. For every change made, I have to re-run the job in developer mode. I have to create needles that have matches and click co-ordinates. I have been stuck on some of these instances:
  1. During installation, my job would not process a second HDD I had added. Roland Clobus , one of my mentors from the team gave me a variable to work with. The solution was adding NUMDISKS=2 as part of the command.
  2. While working on a file, one of the needles would only work after file edits. Afterwards it would fail to assert_and_click . What kept bugging me was why it was passing after the first instance then failing after. The solution was adding a wait_still_screen to the code. This would ensure any screen changes loaded first before clicking happened.
  3. I was stuck on finding the keys that would be needed for a context menu. I added button => right in the assert_and_click code.
  4. Windows 11 installation was constantly failing. Roland pointed out he was working on it so I had to use Windows 10.
  5. Windows 10 Virtual Machine does not connect to the internet because of update restrictions. I had to switch to Linux Virtual Machine for a download job.
When I get stuck, at times I seek guidance from the mentors. I still look for solutions in the documentation. Here are some of the documentation that have helped me get through some of these challenges.
  1. Installation and creating tests guide https://salsa.debian.org/qa/openqa/openqa-tests-debian/-/tree/debian/documentation . These guides help in installation and creating of tests.
  2. OpenQA official documentation https://open.qa/docs/ . This documentation is very comprehensive. I used it recently to read about PUBLISH_HDD_n to save the updated version of a HDD_n I am using.
  3. OpenQA test API documentation https://open.qa/api/testapi/ . This documentation shows me which parameters to use. I have used it recently to find how to right click on a mouse and special characters.
  4. OpenQA variables file in Gitlab https://salsa.debian.org/qa/openqa/openqa-tests-debian/-/blob/debian/VARIABLES.md . This has explanations of the most commonly used variables .
  5. OpenQA repository in Gitlab https://salsa.debian.org/qa/openqa/openqa-tests-debian . I go through the Perl tests. Understand how they work . Then integrate my tests using the a similar manner so that it would look uniform.
  6. OpenQa tests https://openqa.debian.net/tests. I use these tests to find machine settings. I also find test sequences and the assets I would need to create similar tests. I used it recently to look at how graphical login was being implemented then shutdown.
The list above are the documentation that are supposed to be used for these tests and finding solutions. If I don t find anything within these, I then ask Roland for help. I try to go through the autoinst documentation that are from the links provided in the Gitlab README.md file : https://salsa.debian.org/qa/openqa/openqa-tests-debian/-/blob/debian/README.md . They are also comprehensive but are very technical . In general, I get challenges but there is always a means to solve them through documentation provided. The mentors are also very helpful whenever we get challenges. I have gained team contribution skills , upgraded my git skills, learned Perl and how to test using OpenQA. I am still polishing on how to make my needles better. My current progress is thus good. We learn one day at a time.

18 December 2025

Dirk Eddelbuettel: dang 0.0.17: New Features, Plus Maintenance

dang image A new release of my mixed collection of things package dang package arrived at CRAN earlier today. The dang package regroups a few functions of mine that had no other home as for example lsos() from a StackOverflow question from 2009 (!!), the overbought/oversold price band plotter from an older blog post, the market monitor blogged about as well as the checkCRANStatus() function tweeted about by Tim Taylor. And more so take a look. This release retires two functions: the social media site nobody ever visits anymore shut down its API too, so no way to mute posts by a given handle. Similarly, the (never official) ability by Google to supply financial data is no more, so the function to access data this way is gone too. But we also have two new ones: one that helps with CRAN entries for ORCiD ids, and another little helper to re-order microbenchmark results by summary column (defaulting to the median). Other than the usual updates to continuous integrations, as well as a switch to Authors@R which will result in CRAN nagging me less about this, and another argument update. The detailed NEWS entry follows.

Changes in version 0.0.17 (2025-12-18)
  • Added new funtion reorderMicrobenchmarkResults with alias rmr
  • Use tolower on email argument to checkCRANStatus
  • Added new function cranORCIDs bootstrapped from two emails by Kurt Hornik
  • Switched to using Authors@R in DESCRIPTION and added ORCIDs where available
  • Switched to r-ci action with included bootstrap step; updated updated the checkout action (twice); added (commented-out) log accessor
  • Removed googleFinanceData as the (unofficial) API access point no longer works
  • Removed muteTweeters because the API was turned off

Via my CRANberries, there is a comparison to the previous release. For questions or comments use the the issue tracker at the GitHub repo.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. If you like this or other open-source work I do, you can sponsor me at GitHub.

Colin Watson: Preparing a transition in Debusine

We announced a public beta of Debusine repositories recently (Freexian blog, debian-devel-announce). One thing I m very keen on is being able to use these to prepare transitions : changes to multiple packages that need to be prepared together in order to land in testing. As I said in my DebConf25 talk:
We have distribution-wide CI in unstable, but there s only one of it and it s shared between all of us. As a result it s very possible to get into tangles when multiple people are working on related things at the same time, and we only avoid that as much as we do by careful coordination such as transition bugs. Experimental helps, but again, there s only one of it and setting up another one is far from trivial. So, what we want is a system where you can run experiments on possible Debian changes at a large scale without a high setup cost and without fear of breaking things for other people. And then, if it all works, push the whole lot into Debian.
Time to practice what I preach. Setup The setup process is documented on the Debian wiki. You need to decide whether you re working on a short-lived experiment, in which case you ll run the create-experiment workflow and your workspace will expire after 60 days of inactivity, or something that you expect to keep around for longer, in which case you ll run the create-repository workflow. Either one of those will create a new workspace for you. Then, in that workspace, you run debusine archive suite create for whichever suites you want to use. For the case of a transition that you plan to land in unstable, you ll most likely use create-experiment and then create a single suite with the pattern sid-<name>. The situation I was dealing with here was moving to Pylint 4. Tests showed that we needed this as part of adding Python 3.14 as a supported Python version, and I knew that I was going to need newer upstream versions of the astroid and pylint packages. However, I wasn t quite sure what the fallout of a new major version of pylint was going to be. Fortunately, the Debian Python ecosystem has pretty good autopkgtest coverage, so I thought I d see what Debusine said about it. I created an experiment called cjwatson-pylint (resulting in https://debusine.debian.net/debian/developers-cjwatson-pylint/ - I m not making that a proper link since it will expire in a couple of months) and a sid-pylint suite in it. Iteration From this starting point, the basic cycle involved uploading each package like this for each package I d prepared: 
$ dput -O debusine_workspace=developers-cjwatson-pylint \
       -O debusine_workflow=publish-to-sid-pylint \
       debusine.debian.net foo.changes
I could have made a new dput-ng profile to cut down on typing, but it wasn t worth it here. Then I looked at the workflow results, figured out which other packages I needed to fix based on those, and repeated until the whole set looked coherent. Debusine automatically built each upload against whatever else was currently in the repository, as you d expect. I should probably have used version numbers with tilde suffixes (e.g. 4.0.2-1~test1) in case I needed to correct anything, but fortunately that was mostly unnecessary. I did at least run initial test-builds locally of just the individual packages I was directly changing to make sure that they weren t too egregiously broken, just because I usually find it quicker to iterate that way. I didn t take screenshots as I was going along, but here s what the list of top-level workflows in my workspace looked like by the end: Workflows You can see that not all of the workflows are successful. This is because we currently just show everything in every workflow; we don t consider whether a task was retried and succeeded on the second try, or whether there s now a newer version of a reverse-dependency so tests of the older version should be disregarded, and so on. More fundamentally, you have to look through each individual workflow, which is a bit of a pain: we plan to add a dashboard that shows you the current state of a suite as a whole rather than the current workflow-oriented view, but we haven t started on that yet. Drilling down into one of these workflows, it looks something like this: astroid workflow This was the first package I uploaded. The first pass of failures told me about pylint (expected), pylint-flask (an obvious consequence), and python-sphinx-autodoc2 and sphinx-autoapi (surprises). The slightly odd pattern of failures and errors is because I retried a few things, and we sometimes report retries in a slightly strange way, especially when there are workflows involved that might not be able to resolve their input parameters any more. The next level was: pylint workflow Again, there were some retries involved here, and also some cases where packages were already failing in unstable so the failures weren t the fault of my change; for now I had to go through and analyze these by hand, but we ll soon have regression tracking to compare with reference runs and show you where things have got better or worse. After excluding those, that left pytest-pylint (not caused by my changes, but I fixed it anyway in unstable to clear out some noise) and spyder. I d seen people talking about spyder on #debian-python recently, so after a bit of conversation there I sponsored a rope upload by Aeliton Silva, upgraded python-lsp-server, and patched spyder. All those went into my repository too, exposing a couple more tests I d forgotten in spyder. Once I was satisfied with the results, I uploaded everything to unstable. The next day, I looked through the tracker as usual starting from astroid, and while there are some test failures showing up right now it looks as though they should all clear out as pieces migrate to testing. Success! Conclusions We still have some way to go before this is a completely smooth experience that I d be prepared to say that every developer can and should be using; there are all sorts of fit-and-finish issues that I can easily see here. Still, I do think we re at the point where a tolerant developer can use this to deal with the common case of a mid-sized transition, and get more out of it than they put in. Without Debusine, either I d have had to put much more effort into searching for and testing reverse-dependencies myself, or (more likely, let s face it) I d have just dumped things into unstable and sorted them out afterwards, resulting in potentially delaying other people s work. This way, everything was done with as little disruption as possible. This works best when the packages likely to be involved have reasonably good autopkgtest coverage (even if the tests themselves are relatively basic). This is an increasingly good bet in Debian, but we have plans to add installability comparisons (similar to how Debian s testing suite works) as well as optional rebuild testing. If this has got you interested, please try it out for yourself and let us know how it goes!

17 December 2025

Dirk Eddelbuettel: RcppArmadillo 15.2.3-1 on CRAN: Upstream Update

armadillo image Armadillo is a powerful and expressive C++ template library for linear algebra and scientific computing. It aims towards a good balance between speed and ease of use, has a syntax deliberately close to Matlab, and is useful for algorithm development directly in C++, or quick conversion of research code into production environments. RcppArmadillo integrates this library with the R environment and language and is widely used by (currently) 1272 other packages on CRAN, downloaded 43.2 million times (per the partial logs from the cloud mirrors of CRAN), and the CSDA paper (preprint / vignette) by Conrad and myself has been cited 661 times according to Google Scholar. This versions updates to the 15.2.3 upstream Armadillo release from yesterday. It brings minor changes over the RcppArmadillo 15.2.2 release made last month (and described in this post). As noted previously, and due to both the upstream transition to C++14 coupled with the CRAN move away from C++11, the package offers a transition by allowing packages to remain with the older, pre-15.0.0 legacy Armadillo yet offering the current version as the default. If and when CRAN will have nudged (nearly) all maintainers away from C++11 (and now also C++14 !!) we can remove the fallback. Our offer to help with the C++ modernization still stands, so please get in touch if we can be of assistance. As a reminder, the meta-issue #475 regroups all the resources for the C++11 transition. There were no R-side changes in this release. The detailed changes since the last release follow.

Changes in RcppArmadillo version 15.2.3-1 (2025-12-16)
  • Upgraded to Armadillo release 15.2.3 (Medium Roast Deluxe)
    • Faster .resize() for vectors
    • Faster repcube()

Courtesy of my CRANberries, there is a diffstat report relative to previous release. More detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the Rcpp R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. If you like this or other open-source work I do, you can sponsor me at GitHub.

Matthew Garrett: How did IRC ping timeouts end up in a lawsuit?

I recently won a lawsuit against Roy and Rianne Schestowitz, the authors and publishers of the Techrights and Tuxmachines websites. The short version of events is that they were subject to an online harassment campaign, which they incorrectly blamed me for. They responded with a large number of defamatory online posts about me, which the judge described as unsubstantiated character assassination and consequently awarded me significant damages. That's not what this post is about, as such. It's about the sole meaningful claim made that tied me to the abuse.

In the defendants' defence and counterclaim[1], 15.27 asserts in part The facts linking the Claimant to the sock puppet accounts include, on the IRC network: simultaneous dropped connections to the mjg59_ and elusive_woman accounts. This is so unlikely to be coincidental that the natural inference is that the same person posted under both names. "elusive_woman" here is an account linked to the harassment, and "mjg59_" is me. This is actually a surprisingly interesting claim to make, and it's worth going into in some more detail.

The event in question occurred on the 28th of April, 2023. You can see a line reading *elusive_woman has quit (Ping timeout: 2m30s), followed by one reading *mjg59_ has quit (Ping timeout: 2m30s). The timestamp listed for the first is 09:52, and for the second 09:53. Is that actually simultaneous? We can actually gain some more information - if you hover over the timestamp links on the right hand side you can see that the link is actually accurate to the second even if that's not displayed. The first event took place at 09:52:52, and the second at 09:53:03. That's 11 seconds apart, which is clearly not simultaneous, but maybe it's close enough. Figuring out more requires knowing what a "ping timeout" actually means here.

The IRC server in question is running Ergo (link to source code), and the relevant function is handleIdleTimeout(). The logic here is fairly simple - track the time since activity was last seen from the client. If that time is longer than DefaultIdleTimeout (which defaults to 90 seconds) and a ping hasn't been sent yet, send a ping to the client. If a ping has been sent and the timeout is greater than DefaultTotalTimeout (which defaults to 150 seconds), disconnect the client with a "Ping timeout" message. There's no special logic for handling the ping reply - a pong simply counts as any other client activity and resets the "last activity" value and timeout.

What does this mean? Well, for a start, two clients running on the same system will only have simultaneous ping timeouts if their last activity was simultaneous. Let's imagine a machine with two clients, A and B. A sends a message at 02:22:59. B sends a message 2 seconds later, at 02:23:01. The idle timeout for A will fire at 02:24:29, and for B at 02:24:31. A ping is sent for A at 02:24:29 and is responded to immediately - the idle timeout for A is now reset to 02:25:59, 90 seconds later. The machine hosting A and B has its network cable pulled out at 02:24:30. The ping to B is sent at 02:24:31, but receives no reply. A minute later, at 02:25:31, B quits with a "Ping timeout" message. A ping is sent to A at 02:25:59, but receives no reply. A minute later, at 02:26:59, A quits with a "Ping timeout" message. Despite both clients having their network interrupted simultaneously, the ping timeouts occur 88 seconds apart.

So, two clients disconnecting with ping timeouts 11 seconds apart is not incompatible with the network connection being interrupted simultaneously - depending on activity, simultaneous network interruption may result in disconnections up to 90 seconds apart. But another way of looking at this is that network interruptions may occur up to 90 seconds apart and generate simultaneous disconnections[2]. Without additional information it's impossible to determine which is the case.

This already casts doubt over the assertion that the disconnection was simultaneous, but if this is unusual enough it's still potentially significant. Unfortunately for the Schestowitzes, even looking just at the elusive_woman account, there were several cases where elusive_woman and another user had a ping timeout within 90 seconds of each other - including one case where elusive_woman and schestowitz[TR] disconnect 40 seconds apart. By the Schestowitzes argument, it's also a natural inference that elusive_woman and schestowitz[TR] (one of Roy Schestowitz's accounts) are the same person.

We didn't actually need to make this argument, though. In England it's necessary to file a witness statement describing the evidence that you're going to present in advance of the actual court hearing. Despite being warned of the consequences on multiple occasions the Schestowitzes never provided any witness statements, and as a result weren't allowed to provide any evidence in court, which made for a fairly foregone conclusion.

[1] As well as defending themselves against my claim, the Schestowitzes made a counterclaim on the basis that I had engaged in a campaign of harassment against them. This counterclaim failed.

[2] Client A and client B both send messages at 02:22:59. A falls off the network at 02:23:00, has a ping sent at 02:24:29, and has a ping timeout at 02:25:29. B falls off the network at 02:24:28, has a ping sent at 02:24:29, and has a ping timeout at 02:25:29. Simultaneous disconnects despite over a minute of difference in the network interruption.

comment count unavailable comments

14 December 2025

Dirk Eddelbuettel: BH 1.90.0-1 on CRAN: New Upstream

Boost Boost is a very large and comprehensive set of (peer-reviewed) libraries for the C++ programming language, containing well over one hundred individual libraries. The BH package provides a sizeable subset of header-only libraries for (easier, no linking required) use by R. It is fairly widely used: the (partial) CRAN mirror logs (aggregated from the cloud mirrors) show over 41.5 million package downloads. Version 1.90.0 of Boost was released a few days ago following the regular Boost release schedule of April, August and December releases. As before, we packaged it almost immediately and started testing following our annual update cycle which strives to balance being close enough to upstream and not stressing CRAN and the user base too much. The reverse depends check revealed only one really minor issue among the over three hundred direct reverse dependencies. And that issue was addressed yesterday within hours by a truly responsove maintainer (and it helped that a related issue had been addressed months earlier with version 1.89.). So big thanks to Jean-Romain Roussel for the prompt fix, and to Andrew Johnson for the earlier test with 1.89.0. As last year with 1.87.0, no new Boost libraries were added to BH so the (considerable) size is more or less unchanged. It lead to CRAN doing a manual inspection but as there were no other issues it sailed through as is now in the CRAN repository. The short NEWS entry follows.

Changes in version 1.90.0-1 (2025-12-13)
  • Upgrade to Boost 1.90.0, patched as usual to comment-out diagnostic suppression messages per the request of CRAN
  • Minor upgrades to continuous integration

Via my CRANberries, there is a diffstat report relative to the previous release. Comments and suggestions about BH are welcome via the issue tracker at the GitHub repo.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. If you like this or other open-source work I do, you can now sponsor me at GitHub.

12 December 2025

Freexian Collaborators: Debian Contributions: Updates about DebConf Video Team Sprint, rebootstrap, SBOM tooling in Debian and more! (by Anupa Ann Joseph)

Debian Contributions: 2025-11 Contributing to Debian is part of Freexian s mission. This article covers the latest achievements of Freexian and their collaborators. All of this is made possible by organizations subscribing to our Long Term Support contracts and consulting services.

DebConf Video Team Sprint The DebConf Video Team records, streams, and publishes talks from DebConf and many miniDebConfs. A lot of the infrastructure development happens during setup for these events, but we also try to organize a sprint once a year to work on infrastructure, when there isn t a DebConf about to happen. Stefano attended the sprint in Herefordshire this year and wrote up a report.

rebootstrap, by Helmut Grohne A number of jobs were stuck in architecture-specific failures. gcc-15 and dpkg still disagree about whether PIE is enabled occasionally and big endian mipsen needed fixes in systemd. Beyond this regular uploads of libxml2 and gcc-15 required fixes and rebasing of pending patches. Earlier, Loongson used rebootstrap to create the initial package set for loong64 and Miao Wang now submitted their changes. Therefore, there is now initial support for suites other than unstable and use with derivatives.

Building the support for Software Bill Of Materials tooling in Debian, by Santiago Ruano Rinc n Vendors of Debian-based products may/should be paying attention to the evolution of different jurisdictions (such as the CRA or updates on CISA s Minimum Elements for a Software Bill of Materials) that require to make available Software Bill of Materials (SBOM) of their products. It is important then to have tools in Debian to make it easier to produce such SBOMs. In this context, Santiago continued the work on packaging libraries related to SBOMs. This includes the packaging of the SPDX python library (python-spdx-tools), and its dependencies rdflib and mkdocs-include-markdown-plugin. System Package Data Exchange (SPDX), defined by ISO/IEC 5962:2021, is an open standard capable of representing systems with software components as SBOMs and other data and security references. SPDX and CycloneDX (whose python library python3-cyclonedx-lib was packaged by prior efforts this year), encompass the two main SBOM standards available today.

Miscellaneous contributions
  • Carles improved po-debconf-manager: added checking status of bug reports automatically via python-debianbts; changed some command line options naming or output based on user feedback; finished refactoring user interaction to rich; codebase is now flake8-compliant; added type safety with mypy.
  • Carles, using po-debconf-manager, created 19 bug reports for translations where the merge requests were pending; reviewed and created merge requests for 4 packages.
  • Carles planned a second version of the tool that detects packages that Recommends or Suggests packages which are not in Debian. He is taking ideas from dumat.
  • Carles submitted a pull request to python-unidiff2 (adapted from the original pull request to python-unidiff). He also started preparing a qnetload update.
  • Stefano did miscellaneous python package updates: mkdocs-macros-plugin, python-confuse, python-pip, python-mitogen.
  • Stefano reviewed a beets upload for a new maintainer who is taking it over.
  • Stefano handled some debian.net infrastructure requests.
  • Stefano updated debian.social infrastructure for the trixie point release.
  • The update broke jitsi.debian.social, Stefano put some time into debugging it and eventually enlisted upstream assistance, who solved the problem!
  • Stefano worked on some patches for Python that help Debian:
    • GH-139914: The main HP PA-RISC support patch for 3.14.
    • GH-141930: We observed an unhelpful error when failing to write a .pyc file during package installation. We may have fixed the problem, and at least made the error better.
    • GH-141011: Ignore missing ifunc support on HP PA-RISC.
  • Stefano spun up a website for hamburg2026.mini.debconf.org.
  • Rapha l reviewed a merge request updating tracker.debian.org to rely on bootstrap
    version 5.
  • Emilio coordinated various transitions.
  • Helmut sent patches for 26 cross build failures.
  • Helmut officially handed over the cleanup of the /usr-move transition.
  • Helmut monitored the transition moving libcrypt-dev out of build-essential and bumped the remaining bugs to rc-severity in coordination with the release team.
  • Helmut updated the Build-Profiles patch for debian-policy incorporating feedback from Sean Whitton with a lot of help from Nattie Mayer-Hutchings and Freexian colleagues.
  • Helmut discovered that the way mmdebstrap deals with start-stop-daemon may result in broken output and sent a patch.
  • As a result of armel being removed from sid , but not from forky , the multiarch hinter broke. Helmut fixed it.
  • Helmut uploaded debvm accepting a patch from Luca Boccassi to fix it for newer
    systemd.
  • Colin began preparing for the second stage of the OpenSSH GSS-API key exchange package split.
  • Colin caught and fixed a devscripts regression due to it breaking part of Debusine.
  • Colin packaged django-pgtransaction and backported it to trixie , since it looks useful for Debusine.
  • Thorsten uploaded the packages lprng, cpdb-backend-cups, cpdb-libs and ippsample to fix some RC bugs as well as other bugs that accumulated over time. He also uploaded cups-filters to all Debian releases to fix three CVEs.

Next.

Previous.