Search Results: "ag"

9 March 2008

Cyril Brulebois: Funny build systems

While rebuilding packages FTBFS'ing with gcc-4.3, I stumbled upon various quality/sanity levels for code & build systems: from very clean to not-so-nice, dirty, and even scary. Fortunately, upstreams are sometimes funny and want to make us smile. This one succeeded. clanlib.png That one is clanlib, bug #454818. Like pusling already pointed out, there's still a lot to do with RC/RG bugs (although some of them are just prospective/incomplete patches, ls -1 *_gcc-4.3.diff wc -l reports 133 right now). Introductory reading for GCC 4.3: Porting to the New Tools. Please fix your bugs, prepare NMUs and so on, instead of just trolling like/with a bird.

John Goerzen: Vorsanger

This October, our church celebrates its hundredth anniversary. All this year, every so often, we are having brief "historical moments" during church.

Today, our temporary choir director -- and my high school band teacher -- was a Vorsanger. He got up during church and discussed how hymn singing was done in the early days of the church. Today, our Mennonite church sings in a variety of styles, most commonly 4-part music accompanied by a piano, and no director. But we also sing more "contemporary" styles.

Back in the earlier days, the church didn't have hymnals with music for everyone. The Vorsanger -- song leader -- would sing each phrase, then the congregation would sing it back in unison, throughout the whole song. Hymns that the congregation already knew well would just be sung normally, in unison, with the Vorsanger directing.

Not only did the Vorsanger lead the singing, but he or she also chose the music -- not the pastor. As the church purchased first a pump organ, and later a piano, the instruments would sometimes be used to accompany the singing. The musicians never knew what would be sung in advance. One 80-year-old member of our congregation remembered the she started to play for church at 8th grade. But much of the time, there would be no instrument, not even to pick the starting pitch.

This morning, we sang Holy God, We Praise Thy Name with the Vorsanger. It's a familiar song in our church, but we generally sing it in English. The Vorsanger taught it to us in German, the older people in church would have remembered it. He finished his introduction, stepped away from the microphone, and rang out with the first phrase.

Gro er Gott, wir loben dich

Then we all sang the phrase back to him, in unison. On to the next phrase, back and forth for the rest of the hymn. Terah and I were in the choir area, behind him, but could hear him -- and the rest of the congregation -- just fine.

At the end, our Vorsanger for a day said he thought we got it, so we sang through it again, still in unison, but without him leading us on each phrase.

Gro er Gott, wir loben dich,
Herr, wir preisen deine St rke,
Vor dir beugt die Erde sich
Und bewundert deine Werke.
Wie du warst vor aller Zeit,
So bleibst du in Ewigkeit.

What a powerful way to feel connected to the people that worshiped at the same place so many years ago.

As our director said before we sang in choir this morning, music can say so much more than words.

Neil Williams: dpkg-cross and emdebian-tools updates; apt and rsync

dpkg-cross updates
uploaded to Debian unstable

* General support for querying endianness of an architecture.
* Update standards version (no changes)
* Add Cvs-CVS field to debian/control
* Reinstate detect_arch for improved build test checks in emdebian-
tools
* get_endianness: new function for Emdebian TDeb support

This provides support for architecture-dependent TDebs in Emdebian as well as adding detection of armel.

emdebian-tools 0.8.3
Uploaded to Emdebian

* Interim Emdebian release.
* emsecondstage: ensure cdebconf environment is setup before
configuring packages
* emdebian.gpe: Add extra packages needed to allow the gpe package set
to boot.
* Emdebian::Tools, (create_patches) : skip generated files.
[*.debhelper, *.substvars, files]
* em_installtdeb: support special modes via debian/xcontrol
* Support unpack method in basic emsandbox rootfs
* embootstrap: continue after some errors
* Support unpack method in gtk emsandbox rootfs
* Simplify basic empdebuild chroot and copy in hostname files
* empbuilderlib: add support for creating skeleton versions of
/etc/group and /etc/passwd.
* empbuilderlib: add an approximate time set func for use when no
network is available.
* emsecondstage: configure cdebconf before others so that the database
is ready
* embootstrap: use the new support in empbuilderlib for group, passwd
and date
* Add --arch support to em_installtdeb
* Depend on new support in libdebian-dpkgcross-perl
* empdebuild: force the -S argument to apt-cross to prevent
intermittent failures
* empbuilderlib: parse the .list contents from dpkg
* embootstrap: ensure buildplace is clean before starting a fresh
debootstrap
* emdebuild: use dpkg-cross support for detect arch to distinguish
between arm and armel binaries
* Emdebian::Tools. add parse_xcontrol function for easy queries of
debian/xcontrol
* pbuilder/emdebian.crossd gpe gtk : Tidy up the scripts with routine
calls to the empbuilderlib library
* pbuilder/empbuilderlib: Add some common routines to help make the
rootfs bootable
* pbuilder/emsecondstage: bring into line with other changes in the
* empbuilderlib.3 : Add a manpage for the expanded set of common
functions in empbuilderlib
* Suggest file-roller for inspection of rootfs tarballs
* emdebhelper.mk: sync with recent changes in CDBS for debhelper
support


Reasoning
OK, so why upload emdebian-tools to Emdebian when it depends on the update to dpkg-cross in Debian?
Well, emdebian-tools is becoming a bit of a behemoth. I mean a package that has only existed for 12months and has had over 30 releases, increasing in size from 30kb to over 500kb (without counting dependencies) and which encompasses a gammut of scripts with different roles is just begging for a functional package split.
Package splits take time - not just the time in NEW but all the testing to ensure that each split can be installed separately without falling over "Unable to find foo in @INC" etc.
At the same time, emdebian-tools is both rapidly developing and becoming remarkably stable (which gives some indication of just how diverse it has become).
So I'm heading for a 1.0.0 release before Lenny and to achieve that, I want to split it:
  1. emdebian-tools - umbrella package consisting of the core build scripts (perl) - these are becoming quite stable - that depends on all the others.

  2. emdebian-archive-keyring - tiny package that just implements SecureApt for the Emdebian repository.

  3. emdebian-rootfs - the rapidly developing (shell and perl based) root filesystem management scripts, based on debootstrap and pbuilder.

  4. libemdebian-tools-perl - a standard perl module of common functions for the other packages.


This allows people to use the toolchains with the smallest package, prepare root filesystems from prebuilt binary packages with just three of the four and only those who build packages will need all four.
So the plan, overall, is to keep updating emdebian-tools in Emdebian unstable during the time that emdebian-tools 0.9.0 goes through NEW in Debian. Then, when 0.9.x is in testing and any issues from the split are resolved, 1.0.0 can be uploaded to Debian.

Package split issues in general
Something that bugs me with package splits is this stable vs development issue within the same source. I think it would be sensible to only upload emdebian-tools (the stable bit) if it changes but to upload emdebian-rootfs as often as necessary whilst still building from the same source.
This issue will also bite when Emdebian deploys other functional package splits and seeks to improve apt itself to handle many, many more binary packages.
It's also a similar idea to allowing translators to upload TDebs separately from normal package updates.
Why should I have to split the source in $RCS in order to upload package splits separately? Other upstream teams won't do that so Debian needs a solution.
Why should users have to download emdebian-tools 1.0.1 which is absolutely identical to 1.0.0 just because emdebian-rootfs needed lots of updates in 1.0.1. (Remember, emdebian-tools will depend on emdebian-rootfs so when emdebian-rootfs is updated, all users of emdebian-rootfs and all users of emdebian-tools will get the update. Yet I don't see why emdebian-tools itself needs to be touched.)
It is plain daft.
I've seen similar issues in other environments where, for various very good reasons, users need to limit the bandwidth and duration of apt downloads. For Emdebian, we also need to limit the total amount of data downloaded by apt-get upgrade simply because free space on the device can be severely restricted - my target device will have about 7Mb free space and that is regarded as "under-utilised". Downloading an entire package just to update /usr/share/foo/changelog.gz in such circumstances is just plain stupid. I know we can do better than that.
Apt needs to become much more intelligent and have a way of only downloading the changes in packages instead of the entire .deb - something along the lines of an rsync method. To get that working, the actual package names need to be munged so that rsync can compare foo_1.2.3-1_all.deb against foo_1.2.3-2_all.deb by having both symlinked as foo_current_all.deb. (dpkg-name can restore the original name if needed.)
Such a method must be possible and would have enormous benefits for packages like OOo where each upload can't be changing 100% of the code each time, yet we all have to download Mb after Mb of content that is identical to what dpkg is about to replace.
Am I just ranting or does this strike a chord on the lazyweb in general?
The slowest part of many desktop experiences of Debian could well be the time taken to download updates - not everyone has multi Mb broadband.

Gunnar Wolf: Dreamhost: Honest about mistakes. And that's _good_!

I have been maintaining several minor sites hosted at Dreamhost for about a year. And since over one month ago, my personal website is with them as well. And I must say, I am very pleased with them. No, not (well, not only) because they run Debian on their servers, nor because they are probably the cheapest game in town (I paid something like US$200 for a basically unlimited package , for three years), but because of their degree of responsability and personal service.
Responsability? Aren't they well-known for their network outages? Why, yes, of course - Today's example is paramount: Somebody edited the wrong firewall entry, and all of Dreamhost became unavailable. In general terms, Dreamhost has a great blog-like structured page where they inform customers of every network or server problem they have - No, you don't have to dig in to understand why your site is down: They bring it up to you. Upfront. And in a familiar, very non-formal style.
Whenever I have submitted an issue to their request tracker, I get prompt reply. Does it always solve the situation? no, by far. I'm often told to, basically, go screw myself if I really need such feature... But they are straightforward with that, they are good, nice BOFHs (if such thing ever existed), and they don't present you with corporate-minded studies backing up their solution. Yes, I know that in their servers, it's plainly their way or the highway. But hey, that's what I paid for, right?
That is what wins my heart. Yes, Dreamhost is no good for many, many tasks - including, for example, anything that requires a real RDBMS (forgodssake, they offer MySQL but not PostgreSQL, damnit! WTF!?), nor any legendary five-nines reliability. But they are great for the vast majority of the Internet sites' needs. They even exceed what a simple person like me would ever dream of.
So, my hat off to you guys. Again.
(No, and I'm not getting paid or discounted on services because of this blog post. Although maybe I should! ;-) )

Chris Lamb: Timeline of the Debian project

I’ve put together a partial timeline of the Debian project using the SIMILE Project’s DHTML timeline component. This is definately work in progress: I would welcome additional events (by email or merge request please, not via blog comments) regarding: The timeline is available to view here. Update: Assistance fixing the (non-Webkit) Konqueror issue would also be appreciated.

Steve Kemp: Some people get by with a little understanding

Since my last example of fixing a bug received some interesting feedback (although I notice no upload of the package in question ..) we'll have another go. Looking over my ~/.bash_history file one command I use multiple times a day is make. Happily GNU make has at least one interesting bug open: I verified this bug by saving the Makefile in the report and running make:
skx@gold:~$ make
make: file.c:84: lookup_file: Assertion  *name != '\0'' failed.
Aborted
(OK so this isn't a segfault; but an assertion failure is just as bad. Honest!) So I downloaded the source to make, and rebuilt it. This left me with a binary with debugging symbols. The execution was much more interesting this time round:
skx@gold:~$ ./make
*** glibc detected ***
  /home/skx/./make: double free or corruption (fasttop): 0x00000000006327b0 ***
======= Backtrace: =========
/lib/libc.so.6[0x2b273dbdd8a8]
/lib/libc.so.6(cfree+0x76)[0x2b273dbdf9b6]
/home/skx/./make[0x4120a5]
/home/skx/./make[0x4068ee]
/home/skx/./make[0x406fb2]
...
[snip mucho texto]
And once I'd allowed core-file creation ("ulimit -c 9999999") I found I had a core file to help debugging. Running the unstripped version under gdb showed this:
(gdb) up
#5  0x00000000004120a5 in multi_glob (chain=0x1c, size=40) at read.c:3106
3106			    free (memname);
So it seems likely that this free is causing the abort. There are two simple things to do here: To get started I did the first of these: Commenting out the free() call did indeed fix the problem, or at least mask it (at the cost of a memory leak):
skx@gold:~$ ./make
make: *** No rule to make target  Erreur_Lexicale.o', needed by  compilateur'.  Stop.
So, now we need to go back to read.c and see why that free was causing problems. The function containing the free() is "multi_glob". It has scary pointer magic in it, and it took me a lot of tracing to determine the source of the bug. In short we need to change this:
free (memname);
To this:
free (memname);
memname = 0;
Otherwise the memory is freed multiple times, (once each time through the loop in that function. See the source for details). Patch mailed.

Raphael Geissert: where to put such a script?...

$ notAptable --help
notAptable [--version] [--help] [-s/var/lib/dpkg/status] [-aptf/path/to/aptavail/equiv/file] [-aptx'apt-cache dumpavail'] [-ipackagename]
List packages which are installed on the system but are not known by apt.
Options:
-s Path to file to use when grepping for installed packages
-aptf Path to file to use when grepping for the apt-able packages
-aptx Same as above but instead of being a file the specified command is executed
-i Exclude a package that would otherwise be listed
When not overriden by an option the default files/data is up to grep-dctrl's defaults
This script is said to report packages that were removed from the archive
(except those listed in /home/raphael/.listNotAptable.ignore)

Copyright 2008 by Raphael Geissert <atomo64@gmail.com>


With this simple script I wrote I got rid of about 45 packages (libraries, programs, etc, etc) what are no longer in the archive.
My question is: on what package should this script be shipped? debian-goodies? devscripts?

Hope somebody else finds this script useful and puts it in their weekly cronjob as I did :)
Get it from http://alioth.debian.org/~atomo64-guest/notAptable until it is shipped in some package.

Steve Kemp: Book Us A Ticket On The Next Space Shuttle

Tomorrow I turn 32, so I've got to be all mature and responsible and stuff now. Maybe. This means no more song lyrics will be used for blog titles. Instead I shall switch to film quotes. Films are more mature than songs, right? There are three candidates standing for the DPL this year. I'm glad I didn't stand, but I came pretty close to doing so. I just think I'd be unlikely to receive the votes, and busy enough to fail even if I did get picked. It seems to me that every year people promise to do too many things; fixing NEW, fixing the NM process & etc.. Me? I'd have only one goal: Open up the keyring-handling. Nothing more. Nothing less. (Sure I'd blow all the Debian money on new toys for people but that'll be our little secret ;) Anyway since there was a bit of interest I've uploaded a new steam engine video and I've started to document some of my collection: More updates later. Really I need to sit down, clean my toys, and then get some good pictures taken. Maybe next month I'll find the time. (Ha!)

loldebian - Can I has a RC bug?: Drow me a zirkuz plz


Drow me a zirkuz plz

Wouter Verhelst: Reverse hard disk failure

There are two ways of telling a story: The first way is just reciting what happens. I did foo. Then I did bar. That didn't work, so I did quux.. The second way is reciting what mistakes you shouldn't have made. Lessons learned: don't do foo, because that will result in bar, and a visit to the hospital. No, I won't be updating the package, as I promised. In the interest of, well, keeping things interesting, here's a tale of what I did last week, in reverse mode: The good news is that I now have a 160G hard disk rather than a 60G one. Whee.

Enrico Zini: ldap-userdb-fields

Fields used by an LDAP Unix user database Some notes about LDAP for Unix user management: Meaning of fields for objectClass posixAccount: Meaning of fields for objectClass posixGroup: For objectClass inetOrgPerson: what you put there can be used as if it were a vCard by mail programs and contact lists. Now, the name of people could potentially be split in cn,givenName,sn,gecos and displayName (possibly more): how would normal user tools deal with the redundancy? To show a gecos field, pam_ldap will search for a gecos field first, then automatically fallback on building a gecos field out of the other suitable info it finds. To show a name, sane programs try displayName first and if it's not present they guess using the rest. Then there is the issue of how to chose the dn to identify users, groups and so on. Users usually go in uid=$USERNAME,ou=People,$SUFFIX, while groups would usually go in cn=$GROUPNAME,ou=Groups,$SUFFIX. Should you need to create the People and Groups organizational units, this could be the proper bit of LDIF:
dn: ou=$NAME,$SUFFIX
ou: $NAME
objectClass: organizationalUnit
To add fields that are not already part of a schema, one needs to create their own schema. To do that, one needs to first obtain (free of charge) a Private Enterprise Number that is used in various places in the schema definition. Making up your own one means risking conflicts if you eventually grow larger. But it is rarely needed, because for most things there are already schemas available. Many thanks to Wouter and noshadow for allowing me to crudely extract all this content from their brains.

Ingo Juergensmann: Armeb removed from Buildd.Net

I've removed armeb as an architecture from Buildd.Net, because it's simply dead. When there's ever the need for it again, support can be easily re-enabled.

Biella Coleman: Dead Media Archive

When you join an academic department, you are just one body and mind among many others doing what you do: teach, write, go to conferences, answer a lot of student emails, advise, etc. However, what you don’t get to do is really see your colleagues very often. While we come together far too frequently for faculty meetings, once you are done, you just want to scuttle out of the room (and as quickly as possible) to continue what seems to be an endless stream of work. But over the course of months and years, there are times and moments when you do come across your colleagues and their work: you may assign some of their work in your class class, go see them speak at an event, or learn about their work over coffee. This weekend during a PhD prospective weekend, I finally got to learn about what is probably one of the coolest projects to come out of my department, The Dead Media Archive run by Ben Kafka and Alex Galloway for their Dead Media Archeology Class. Loosely inspired by a mailing list by the same name started years ago by Bruce Sterling (which eventually suffered its own death), they are collecting information and analyzing the significance of all sorts of Dead Media, some of which were quite present and important for our economy and communication systems, others which flickered much more briefly as a bright idea but never really caught on. As part of their class assignments, students have to write in-depth descriptions, histories, and analysis’ of these objects and they are fantastic; so if you are into dead media, do check out the archive.

Aigars Mahinovs: Standard software not in standard?

I was consulting a small company with a couple Debian servers the other day and I found that they did not have some packages that I expected to be there. Now thinking about it on every server that I install or take over the first thing I do is install a bunch of packages, such as: sudo, mc, wajig, localepurge and a bunch of others that I can’t remember at the first moment, but that I re-discover each time I find them missing. I assume that other people have discovered other great non-standard tools that I am missing out on. What software do other people consider standard on their servers?

Bernhard R. Link: Why I am putting dpkg on hold on my unstable boxes right now...

A hijack of an important package makes me always uneasy. But reading the next two messages to the debian-dpkg list, which I can only read as "oh, I broke make dist. But why should I follow sane practises, all hail my workflow" and "oh, I brake Changelog, that's on purpose because I consider it a bad idea, if you want it make write some scripts to generate it" does nothing to make me believe such a person can be a good maintainer.

David Welton: Bypassing Android's &quot;Resources&quot;

Android relies heavily on something called "resources" to access portions of the system that aren't directly in the Java code. For instance, if you had an ogg file that you wanted to use to play a sound with, you would add it as res/raw/foo.ogg. Then, during compilation, the aapt tool transforms all your resources into a Java file called R.java, which is basically a series of integers that can be used to reference the actual resources. For instance, to set up our hypothetical ogg file, we would do:
MediaPlayer mp = MediaPlayer.create(context, R.raw.foo);
That's all well and good as far as it goes, and if you don't have any reason not to use it, you should. However, with Hecl, one of my goals is to make it very easy to create applications. This means: I'm not there yet, on the Android platform, which is a bit more complex than Java ME (where it is indeed so easy to create new Hecl .jars that it's possible to do it on the fly via this web site: http://builder.hecl.org), but I am making progress. One thing that a user might want to do is add their own files to the .apk file that actually gets installed on the phone. Since it's just a zip file, that's pretty easy:
zip -r Hecl.apk res/raw/foo.ogg
However, without a resource, there is no way to access that file, so the question is: how can we get ahold of it somehow? The first piece of the puzzle is Java's ZipFile class, which lets us open up a zip compressed file, such as the .apk file that contains the ogg file we're interested in in this case. The critical passage is getting ahold of the actual filename on the Android filesystem of the code that's currently running. It turns out to not be too difficult, if a bit obscure:
String filename = this.getPackageManager().getApplicationInfo(packagename, 0).sourceDir;
With that information in hand, we can feed it to ZipFile, and then access the contents at run time. It's not a perfect nor complete solution, yet. So many of the Android API's require a resource ID, so it takes some extra coding to do things from the InputStream that we can get out of the zip file, but there may be ways to make this easier. For instance, we could simply dump the contents of raw/ into an application specific data directory the first time an application runs, in order to at least be able to access the files as files, rather than InputStreams connected to ZipEntries. In any case, at least it's possible to access files added to the .apk without requiring the compilation step. I'm not sure how many other use cases there might be for direct access to the .apk file, but if other people need to do this, hopefully this information will be helpful.

Vincent Fourmond: ctioga's non-linear axes

Much as been done on ctioga since the last release, not even two weeks ago: small bug fixes, tangents improvements, the rebirth of a --subplot option, a saner framework for plot styles... But the most significant improvement in my opinion is the possibility to make non-linear axes: you just need to provide a two-ways transformation between the standard coordinates and the axis coordinates, and ctioga does the rest:
ctioga -N --xpdf --math --math-xrange 2:8 \
  'sin(x)' --xaxis bottom --new-axis biniou top \
  --axis-function biniou 1/x 1/x \
  --lines-color biniou Gray --no-title
As you can see, I did profit to add the possibility of background lines corresponding to the axes' major ticks. The only real trick that is left is to find a decent way to partition the target segment so that they ticks are "natural" (not 0.34759, for instance) and reasonably equally spaced. I thought other plotting programs had dealt with this, but in fact, that doesn't seem to be the case: I'll have to do that myself.

Kartik Mistry: /etc/init.d/early-blr-life


* Initial life at any place is not easy. Away from /home, missing 2Ks, dealing with Uncertainty Principle daily and almost 2/3 km walking is making not so good effect. My body is adjusting with good amount of rice and rice products - I love that!! * Work is great till now (I mean I have done almost 0 work!). Installed Sid on my laptop. No issues till so far.

Kartik Mistry: In BLR


* Just say ‘Hi’ to me, if you are in BLR. I’ll be here :) Even, after journey of 22 hours, I am very much fresh like lime soda. Work is not yet started, but getting into it. I have joined Comat, where very familiar guys - Jace, Parthan and Kushal rocks. Thanks guys for warm welcome!

Kartik Mistry: fortunes-debian-hints


* Fortune is nice. Combination of Fortune and Debian can be much nicer! If you have encounter ‘tips and tricks’ while working with Debian system, you might want to contribute some useful tips to Debian users. Have a look at wiki page of FortunesDebianHints Hints can be from pointing useful package to short command that can be useful in daily activity for system admins. Look at previous hints for example. You can also comment here or drop me mail to me.

Next.

Previous.