Search Results: "Dirk Eddelbuettel"

19 November 2025

Dirk Eddelbuettel: #055: More Frequent r2u Updates

Welcome to post 55 in the R4 series. r2u brings CRAN packages for R to Ubuntu. We mentioned it in the R4 series within the last year in posts #54 about faster CI, #48 about the r2u keynote at U Mons, #47 reviewing r2u it at its third birthday, #46 about adding arm64 support, and #44 about the r2u for mlops talk. Today brings news of an important (internal) update. Following both the arm64 builds as well as the last bi-annual BioConductor package update (and the extension of BioConductor coverage to arm64), more and more of our build setup became automated at GitHub. This has now been unified. We dispatch builds for amd64 packages for jammy (22.04) and noble (24.04) (as well as for the arm64 binaries for noble ) from the central build repository and enjoy the highly parallel build of the up to fourty available GitHub Runners. In the process we also switched fully to source builds. In the past, we had relied on p3m.dev (formerly known as ppm and rspm) using its binaries. These so-called naked binaries are what R produces when called as R CMD INSTALL --build. They are portable with the same build architecture and release, but do not carry packaging information. Now, when a Debian or Ubuntu .deb binary is built, the same step of R CMD INSTALL --build happens. So our earlier insight was to skip the compilation step, use the p3m binary, and then wrap the remainder of a complete package around it. Which includes the all-important dependency information for both the R package relations (from hard Depends / Imports / LinkingTo or soft Suggests declarations) as well as the shared library dependency resolution we can do when building for a Linux distribution. That served us well, and we remain really grateful for the p3m.dev build service. But it also meant were dependending on the clock and cadence of p3m.dev. Which was not really a problem when it ran reliably daily, and early too, included weekends, and showed a timestamp of last updates. By now it is a bit more erratic, frequently late, skips weekends more regularly and long stopped showing when it was last updated. Late afternoon releases reflecting the CRAN updates ending one and half-days earlier is still good, it s just not all that current. Plus there was always the very opaque occurrencem where maybe one in 50 packages or so would not even be provided as a binary so we had to build it anyway the fallback always existing, and was used for both BioConductor (no binaries) and arm64 (no binaries at first, this now changed). So now we just build packages the standard way, albeit as GitHub Actions. In doing so we can ignore p3m.dev, and rather follow the CRAN clock and cadence (as for example CRANberries does), and can update several times a day. For example early this morning (Central time) we ran update for the then-new 28 source packages resulting in 28 jammy and 36 noble binary packages; right now in mid-afternoon we are running another build for 37 source packages resuling in 37 jammy and 47 noble packages. (Packages without a src/ directory and hence no compilation can be used across amd64 and arm64; those that do have src/ are rebuilt for arm64 hence the different sets of jammy and noble packages as only the latter has arm64 now.) This gets us packages from this morning into r2u which p3m.dev should have by tomorrow afternoon or so. And with that r2u remains Fast. Easy. Reliable. Pick all three! and also a little more predictable and current in its delivery. What s not to like?

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 November 2025

Dirk Eddelbuettel: digest 0.6.38 on CRAN: Several Updates

Release 0.6.38 of the digest package arrived at CRAN today and has also been uploaded to Debian. digest creates hash digests of arbitrary R objects. It can use a number different hashing algorithms (md5, sha-1, sha-256, sha-512, crc32, xxhash32, xxhash64, murmur32, spookyhash, blake3,crc32c, xxh3_64 and xxh3_128), and enables easy comparison of (potentially large and nested) R language objects as it relies on the native serialization in R. It is a mature and widely-used package (with 86.8 million downloads just on the partial cloud mirrors of CRAN which keep logs) as many tasks may involve caching of objects for which it provides convenient general-purpose hash key generation to quickly identify the various objects. This release, the first in about fifteen months, updates a number of items. Carl Pearson suggested, and lead, a cleanup of the C API in order to make more of the functionality accessibly at the source level of other packages. This is ongoing / not yet complete but lead to several nice internal cleanups, mostly done by Carl. Several typos were corrected, mostly in Rd files, by Bill Denney, who also improved the test coverage statistics. Thierry Onkelinx and I improved the sha1 functionality. Sergey Fedorov improved an endianness check that matter for his work on PowerPC. I updated the blake3 hasher, expanded the set of ORCID IDs for listed contributors, updated the continuous integration setup, reinstated code coverage reports, refreshed / converted the documentation site setup, and made general updates and edits to the documentation. The release was prepared a week ago, and held up a few days until an affected package was updated: it requested raw returns where none were previously delivered (for xxhash64) but now are so it needed to not reqest them. It was then seen that another package made some assumptions about our DESCRIPTION file; this has been addressed at its end via a pull request we submitted (that remains unmerged). This delayed processing at CRAN for a few days and as it happens, hours after the packages was updated at CRAN today I heard from the ever-so-tireless Brian Ripley about an SAN issue on arm64 only. He kindly provided a fix; it needed a cast. Checking this on amd64 against our Rocker-based ASAN and UBSAN containers (where is remains impossible to replicate, this issue is apparently known for some arm64 issues) another micro-issue (of a missing final argument NULL missing in one .Call()) was detected. I plan to fix both of these in a follow-up release next week. My CRANberries provides a summary of changes to the previous version. For questions or comments use the issue tracker off the GitHub repo. For documentation (including the changelog) see the documentation site. If you like this or other open-source work I do, you can now sponsor me at GitHub.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

11 November 2025

Dirk Eddelbuettel: duckdb-mlpack 0.0.4: Added random forest and logistic regression

A new release of the budding duckdb extension for mlpack, the C++ header-only library for machine learning, was merged into the duckdb community extensions repo today, and has been updated at its duckdb mlpack extension page. This release 0.0.4 adds two new methods (random forests, and regularized logistic regression), reworked the interface a little to now consistently provide fit (or train) and predict methods, adds a new internal state variable mlpack_verbose which can trigger (or suppress) verbose mode directly from SQL, expanded the documentation and added more unit tests. For more details, see the repo for code, issues and more, and the extension page for more about this duckdb community extension.

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.

4 November 2025

Dirk Eddelbuettel: RcppCNPy 0.2.14 on CRAN: Minor Maintenance

Another (again somewhat minor) maintenance release of the RcppCNPy package arrived on CRAN just now. RcppCNPy provides R with read and write access to NumPy files thanks to the cnpy library by Carl Rogers along with Rcpp for the glue to R. The changes are all minor chores. As R now checks usage of packages in demos, we added the rbenchmark to Suggests: in DESCRIPTION. We refreshed the main continuous integration script for a minor update, and also replaced one URL in a badge to avoid a timeout during checks at CRAN. So nothing user-facing this time! Full details are below.

Changes in version 0.2.14 (2024-11-03)
  • The rbenchmark package is now a Suggests: as it appears in demo
  • The continuous integration setup now uses r-ci with its embedded setup step
  • The URL used for the GPL-2 is now the R Project copy

CRANberries also provides a diffstat report for the latest release. As always, feedback is welcome and the best place to start a discussion may be the GitHub issue tickets page. If you like this or other open-source work I do, you can now sponsor me at GitHub.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

3 November 2025

Dirk Eddelbuettel: duckdb-mlpack 0.0.3: macOS binaries, unit tests, more outputs

A littler two weeks a short post announced the duckdb-mlpack as ML quacks : combining the powerful C++ machine learning library mlpack with the amazing analytical database engine duckdb. About a week ago another short post covered first extensions. We actually followed-up with release 0.0.3 days later, and never posted about it so this short note catches up. In release 0.0.3, we provide macOS binaries: following a known issue with one of the components, we apply a simple patch to enable the build. Next up are wasm and windows, if you know your way around these platforms please get in touch. Release 0.0.3 also added first unit tests, serializes the coefficients from the (regularized) linear regression into the output table. See see two previous posts linked above for details and background, the repo for code, issues and more, and the extension page for more about this duckdb community extension.

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.

27 October 2025

Dirk Eddelbuettel: #054: Faster r-ci Continuous Integration via r2u Container

Welcome to post 54 in the R4 series. The topic of continuous integration has been a recurrent theme here at the R4 series. Post #32 introducess r-ci, while post #41 brings r2u to r-ci, but does not show a matrix deployment. Post #45 describes the updated r-ci setup that is now the default and contains a macOS and Ubuntu matrix, where the latter relies on r2u to keep things fast, easy, reliable . Last but not least more recent post #52 shares a trick for ensuring coverage reports. Following #45, use of r-ci at for example GitHub Actions has seen steady use and very reliable performance. With the standard setup, a vanilla Ubuntu setup is changed into one supported by r2u. This requires downloading and installating a few Ubuntu packages, and has generally been fairly quick on the order of around fourty seconds. Now, the general variability of run-times for identical tasks in GitHub Actions is well documented by the results of the setup described in post #39 which still runs weekly. It runs the identical SQL query against a remote backend using two different package families. And lo and behold, the intra-method variability on unchanged code or setup and therefore due solely to system variability is about as large as the inter-method variability. In short, GitHub Actions performance varies randomly with significant variability. See the repo README.md for chart that updates weekly (and see #39 for background). Of late, this variability became more noticeable during standard GitHub Actions runs where it would regularly take more than two minutes of setup time before actual continuous integration work was done. Some caching seems to be in effect, so subsequent runs in the same repo seem faster and often came back to one minute or less. For lightweight and small packages, loosing two minutes to setup when the actual test time is a minute or less gets old fast. Looking around, we noticed that container use can be combined with matrix use. So we have now been deploying the following setup (not always over all the matrix elements though)
jobs:
  ci:
    strategy:
      matrix:
        include:
          -   name: container, os: ubuntu-latest, container: rocker/r2u4ci  
          -   name: macos,     os: macos-latest   
          -   name: ubuntu,    os: ubuntu-latest  

    runs-on: $  matrix.os  
    container: $  matrix.container  
GitHub Actions is smart enough to provide NULL for container in the two other cases, so container: $ matrix.container is ignored there. But when container is set as here for the ci-enhanced version of r2u (which adds a few binaries commonly needed such as git, curl, wget etc needed for CI) then the CI jobs runs inside the container. And thereby skips most of the setup time as the container is already prepared. This also required some small adjustments in the underlying shell script doing the work. To not disrupt standard deployment, we placed these into a release candidate / development version one can op into via an new variable dev_version
      - name: Setup
        uses: eddelbuettel/github-actions/r-ci@master
        with:
          dev_version: 'TRUE'
Everything else remains the same and works as before. But faster as much less time is spent on setup. You can see the actual full yaml file and actions in my repositories for rcpparmadillo and rcppmlpack-examples. Additional testing would be welcome, so feel free to deploy this in your actions now. Otherwise I will likely carry this over and make it the defaul in a few weeks time. It will still work as before but when the added container: line is used will run much faster thanks to rocker/r2u4ci being already set up for CI.

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.

26 October 2025

Dirk Eddelbuettel: duckdb-mlpack 0.0.2: mlpack is now a duckdb community extension

A couple of days ago in a short post, I announced duckdb-mlpack as ML quacks : combining the powerful C++ machine learning library mlpack with the amazing analytical database engine duckdb. See that post for more background. The duckdb-mlpack package is now a community extension joining an impressive list of existing extensions. This means duckdb builds and distributes duckdb-mlpack for all supported platforms allowing users to just install the resulting (signed) binary. (We currently only support Linux in both arm64 and amd64, adding macOS should be straightforward once we sort one build issue out. Windows and WASM should work too, with a little love and polish, as both duckdb and mlpack support them.) Given the binary build, a simple
INSTALL mlpack FROM community;
LOAD mlpack;
installs and loads the package. By the duckdb convention the code is stored per-user and per-version, so the first line needs to be executed only once per duckdb release used. The second line is then per session. We also extended the capabilities of duckdb-mlpack. While still a MVP stressing minimal viable product, the two supported methods adaBoost and (regularized) linear regression both serialize and store their model object permitting rapid prediction on new data as shown in the adaBoost example:
-- Perform adaBoost (using weak learner 'Perceptron' by default)
-- Read 'features' into 'X', 'labels' into 'Y', use optional parameters
-- from 'Z', and prepare model storage in 'M'
CREATE TABLE X AS SELECT * FROM read_csv("https://eddelbuettel.github.io/duckdb-mlpack/data/iris.csv");
CREATE TABLE Y AS SELECT * FROM read_csv("https://eddelbuettel.github.io/duckdb-mlpack/data/iris_labels.csv");
CREATE TABLE Z (name VARCHAR, value VARCHAR);
INSERT INTO Z VALUES ('iterations', '50'), ('tolerance', '1e-7');
CREATE TABLE M (json VARCHAR);

-- Train model for 'Y' on 'X' using parameters 'Z', store in 'M'
CREATE TEMP TABLE A AS SELECT * FROM mlpack_adaboost("X", "Y", "Z", "M");

-- Count by predicted group
SELECT COUNT(*) as n, predicted FROM A GROUP BY predicted;

-- Model 'M' can be used to predict
CREATE TABLE N (x1 DOUBLE, x2 DOUBLE, x3 DOUBLE, x4 DOUBLE);
-- inserting approximate column mean values
INSERT INTO N VALUES (5.843, 3.054, 3.759, 1.199);
-- inserting approximate column mean values, min values, max values
INSERT INTO N VALUES (5.843, 3.054, 3.759, 1.199), (4.3, 2.0, 1.0, 0.1), (7.9, 4.4, 6.9, 2.5);
-- and this predict one element each
SELECT * FROM mlpack_adaboost_pred("N", "M");
Ryan and I have some ideas for where to go from here, ideally towards autogenerating bindings for most (if not all) methods as is done for the mlpack language bindings. Anybody interested and willing to help should reach out to us.

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.

20 October 2025

Dirk Eddelbuettel: RcppArmadillo 15.2.0-0 on GitHub: New Upstream, Simpler OpenMP

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) 1270 other packages on CRAN, downloaded 42 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 650 times according to Google Scholar. This versions updates to the 15.2.0 upstream release made today. It brings a few changes over Armadillo 15.0 (see below for more). It follows the most recent RcppArmadillo 15.0.2-2 release and the Armadillo 15 upstream transition with its dual focus on moving on from C++11 and deprecation of a number of API access points. As we had a few releases last month to manage the transition, we will sit this upgrade out and not upload to CRAN in order to normalize our update cadence towards the desired about six in six months (that the CRAN Policy asks for). One can of course install as usual directly from the GitHub repository as well as from r-universe which also offers binaries for all CRAN platforms. The transition to Armadillo 15 appears to be going slowly but steadily. We had well over 300 packages with either a need to relax the C++11 setting and/or update away from now-deprecated API access points. That number has been cut in half thanks to a lot of work from a lot of package maintainers which is really appreciated! Of course, a lot remains to be done. Issues #489 and #491 contain the over sixty PRs and patches I prepared for all packages with at least one reverse dependency. Most (but not all) have aided in CRAN updates, some packages are still outstanding in terms of updates. As before meta-issue #475 regroups all the resources for the transition. If you, dear reader, have a package that is affected and I could be of assistance please do reach out. The other change we made is to greatly simplify the detection and setup of OpenMP. As before, we rely on configure to attempt compilation of a minimal OpenMP-using program in order to pass the success or failure onto Armadillo as a can-or-cannot use OpenMP. In the year 2025 one of the leading consumer brands still cannot ship an OS where this works out of the box, so we try to aide there. For all others systems, R actually covers this pretty well and has a reliable configuration variable that we rely upon. Just as we recommend for downstream users of the package. This setup should be robust, but is a change so by all means if you knowingly rely on OpenMP please test and report back. The detailed changes since the last CRAN release follow.

Changes in RcppArmadillo version 15.2.0-0 (2025-10-20) (GitHub Only)
  • Upgraded to Armadillo release 15.2.0 (Medium Roast Deluxe)
    • Added rande() for generating matrices with elements from exponential distributions
    • shift() has been deprecated in favour of circshift(), for consistency with Matlab/Octave
    • Reworked detection of aliasing, leading to more efficient compiled code
  • OpenMP detection in configure has been simplified

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.

17 October 2025

Dirk Eddelbuettel: ML quacks: Combining duckdb and mlpack

A side project I have been working on a little since last winter and which explores extending duckdb with mlpack is now public at the duckdb-mlpack repo. duckdb is an excellent small (as in runs as a self-contained binary ) database engine with both a focus on analytical payloads (OLAP rather than OLTP) and an impressive number of already bolted-on extensions (for example for cloud data access) delivered as a single-build C++ executable (or of course as a library used from other front-ends). mlpack is an excellent C++ library containing many/most machine learning algorithms, also built in a self-contained manner (or library) making it possible to build compact yet powerful binaries, or to embed (as opposed to other ML framework accessed from powerful but not lightweight run-times such as Python or R). The compact build aspect as well as the common build tools (C++, cmake) make these two a natural candidate for combining them. Moreover, duckdb is a champion of data access, management and control and the complementary machine learning insights and predictions offered by mlpack are fully complementary and hence fit this rather well. duckdb also has a very robust and active extension system. To use it, one starts from a template repository and its use this template button, runs a script and can then start experimenting. I have now grouped my initial start and test functions into a separate repository duckdb-example-extension to keep the duckdb-mlpack one focused on the extend to mlpack aspect. duckdb-mlpack is right an MVP , i.e. a minimally viable product (or demo). It just runs the adaboost classifier but does so on any dataset fitting the rectangular setup with columns of features (real valued) and a final column (integer valued) of labels. I had hope to use two select queries for both features and then labels but it turns a table function (returning a table of data from a query) can only run one select *. So the basic demo, also on the repo README is now to run the following script (where the SELECT * FROM mlpack_adaboost((SELECT * FROM D)); is the key invocation of the added functionality):
#!/bin/bash

cat <<EOF   build/release/duckdb
SET autoinstall_known_extensions=1;
SET autoload_known_extensions=1; # for httpfs

CREATE TEMP TABLE Xd AS SELECT * FROM read_csv("https://mlpack.org/datasets/iris.csv");
CREATE TEMP TABLE X AS SELECT row_number() OVER () AS id, * FROM Xd;
CREATE TEMP TABLE Yd AS SELECT * FROM read_csv("https://mlpack.org/datasets/iris_labels.csv");
CREATE TEMP TABLE Y AS SELECT row_number() OVER () AS id, CAST(column0 AS double) as label FROM Yd;
CREATE TEMP TABLE D AS SELECT * FROM X INNER JOIN Y ON X.id = Y.id;
ALTER TABLE D DROP id;
ALTER TABLE D DROP id_1;
CREATE TEMP TABLE A AS SELECT * FROM mlpack_adaboost((SELECT * FROM D));

SELECT COUNT(*) as n, predicted FROM A GROUP BY predicted;
EOF
to produce the following tabulation / group by:
./sampleCallRemote.sh
Misclassified: 1
 
    n     predicted  
  int64     int32    
 
     50           0  
     49           1  
     51           2  
 
$
(Note that this requires the httpfs extension. So when you build from a freshly created extension repository you may be ahead of the most recent release of duckdb by a few commits. It is easy to check out the most recent release tag (or maybe the one you are running for your local duckdb binary) to take advantage of the extensions you likely already have for that version. So here, and in the middle of October 2025, I picked v1.4.1 as I run duckdb version 1.4.1 on my box.) There are many other neat duckdb extensions. The core ones are regrouped here while a list of community extensions is here and here. For this (still more minimal) extension, I added a few TODO items to the README.md: Please reach out if you are interested in working on any of this.

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.

14 October 2025

Dirk Eddelbuettel: qlcal 0.0.17 on CRAN: Regular Update

The seventeenth release of the qlcal package arrivied at CRAN today, once again following a QuantLib release as 1.40 came out this morning. qlcal delivers the calendaring parts of QuantLib. It is provided (for the R package) as a set of included files, so the package is self-contained and does not depend on an external QuantLib library (which can be demanding to build). qlcal covers over sixty country / market calendars and can compute holiday lists, its complement (i.e. business day lists) and much more. Examples are in the README at the repository, the package page, and course at the CRAN package page. This releases mainly synchronizes qlcal with the QuantLib release 1.40. Only one country calendar got updated; the diffstat looks larger as the URL part of the copyright got updated throughout. We also updated the URL for the GPL-2 badge: when CRAN checks this, they always hit a timeout as the FSF server possibly keeps track of incoming requests; we now link to version from the R Licenses page to avoid this.

Changes in version 0.0.17 (2025-07-14)
  • Synchronized with QuantLib 1.40 released today
  • Calendar updates for Singapore
  • URL update in README.md

Courtesy of my CRANberries, there is a diffstat report for this release. See the project page and package documentation for more details, and more examples.

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.

12 October 2025

Dirk Eddelbuettel: RcppSpdlog 0.0.23 on CRAN: New Upstream

Version 0.0.23 of RcppSpdlog arrived on CRAN today (after a slight delay) and has been uploaded to Debian. RcppSpdlog bundles spdlog, a wonderful header-only C++ logging library with all the bells and whistles you would want that was written by Gabi Melman, and also includes fmt by Victor Zverovich. You can learn more at the nice package documention site. This release updates the code to the version 1.16.0 of spdlog which was released yesterday morning, and includes version 1.12.0 of fmt. We also converted the documentation site to now using mkdocs-material to altdoc (plus local style and production tweaks) rather than directly. I updated the package yesterday morning when spdlog was updated. But the passage was delayed for a day at CRAN as their machines still times out hitting the GPL-2 URL from the README.md badge, leading to a human to manually check the log assert the nothingburgerness of it. This timeout does not happen to me locally using the corresponding URL checker package. I pondered this in a r-package-devel thread and may just have to switch to using the R Project URL for the GPL-2 as this is in fact recurrning. The NEWS entry for this release follows.

Changes in RcppSpdlog version 0.0.23 (2025-10-11)
  • Upgraded to upstream release spdlog 1.16.0 (including fmt 12.0)
  • The mkdocs-material documentation site is now generated via altdoc

Courtesy of my CRANberries, there is also a diffstat report detailing changes. More detailed information is on the RcppSpdlog page, or the package documention site.

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.

10 October 2025

Dirk Eddelbuettel: RcppArmadillo 15 CRAN Transition: Offering Office Hours

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) 1273 other packages on CRAN, downloaded 41.8 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 651 times according to Google Scholar. Armadillo 15 brought changes. We mentioned these in the 15.0.2-1 and 15.0.2-1 release blog posts: (The second point is a consequence of the first. Prior to C++14, deprecation notes were issue via a macro, and the macro was set up by Conrad in the common way of allowing an override, which we took advantage of in RcppArmadillo effectively shielding downstream package. In C++14 this is now an attribute, and those cannot be suppressed.) We tested this then-upcoming change extensively: Thirteen reverse dependency runs expoloring different settings and leading to the current package setup where an automatic fallback to the last Armadillo 14 release offers fallback for hardwired C++11 use and Armadillo 15 others. Given the 1200+ reverse deoendencies, this took considerable time. All this was also quite extensively discussed with CRAN (especially Kurt Hornik) and documented / controlled via a series of issue tickets starting with overall issue #475 covering the subissues: The sixty pull requests (or emailed patches) followed a suggestion by CRAN to rank-order packages affected by their reverse dependencies sorted in descending package count. Now, while this change from Armadillo 14 to 15 was happening, CRAN also tightened the C++11 requirement for packages and imposed a deadline for changes. In discussion, CRAN also convinced me that a deadline for the deprecation warning, now unmasked, was viable (and is in fairness commensurate with similar, earlier changes triggered by changes in the behaviour of either gcc/g++ or clang/clang++). So we now have two larger deadline campaigns affecting the package (and as always there are some others). These deadlines are coming close: October 17 for the C++11 transition, and October 23 for the deprecation warning. Now, as became clear preparing the sixty pull requests and patches, these changes are often relatively straightforward. For the former, remove the C++11 enforcement and the package will likely build without changes. For the latter, make the often simple (e.g. swith from arma::is_finite to std::isfinite) change. I did not encounter anything much more complicated yet. The number of affected packages approximated by looking at all packages with a reverse dependency on RcppArmadillo and having a deadline can be computed as
suppressMessages(library(data.table))
D <- setDT(tools::CRAN_package_db())
P <- data.table(Package=tools::package_dependencies("RcppArmadillo", reverse=TRUE, db=D)[[1]])
W <- merge(P, D, all.x=TRUE)[is.na(Deadline)==FALSE,c(1:2,38,68)]
W

W[, nrevdep := length(tools::package_dependencies(Package, reverse=TRUE, recursive=TRUE, db=D)[[1]]), by=Package]
W[order(-nrevdep)]
and has been declining steadily from over 350 to now under 200. For that a big and heartfelt Thank You! to all the maintainers who already addressed their package and uploaded updated packages to CRAN. That rocks, and is truly appreciated. Yet the number is still large. And while issues #489 and #491 show a number of pending packages that have merged but not uploaded (yet?) there are also all the other packages I have not been able to look at in detail. While preparing sixty PRs / patches was viable over a period of a good week, I cannot create these for all packages. So with that said, here is a different suggestion for help: All of next week, I will be holding open door open source office hours online two times each day (11:00h to 13:00h Central, 16:00h to 18:00h Central) which can be booked via this booking link for Monday to Friday next week in either fifteen or thirty minutes slots you can book. This should offer Google Meet video conferencing (with jitsi as an alternate, you should be able to control that) which should allow for screen sharing. (I cannot hookup Zoom as my default account has organization settings with a different calendar integration.) If you are reading this and have a package that still needs helps, I hope to see you in the Open Source Office Hours to aid in the RcppArmadillo package updates for your package. Please book a slot!

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.

9 October 2025

Dirk Eddelbuettel: xptr 1.2.0 on CRAN: New(ly Adopted) Package!

Excited to share that xptr is back on CRAN! The xptr package helps to create, check, modify, use, share, external pointer objects. External pointers are used quite extensively throughout R to manage external resources such as datanbase connection objects and alike, and can be very useful to pass pointers to just about any C / C++ data structure around. While described in Writing R Extensions (notably Section 5.13), they can be a little bare-bones and so this package can be useful. It had been created by Randy Lai and maintained by him during 2017 to 2020, but then fell off CRAN. In work with nanoarrow and its clean and minimal Arrow interface xptr came in handy so I adopted it. Several extensions and updates have been added: (compiled) function registration, continuous integration, tests, refreshed and extended documentation as well as a print format extension useful for PyCapsule objects when passing via reticulate. The package documentation site was switched to altdoc driving the most excellent Material for MkDocs framework (providing my first test case of altdoc replacing my older local scripts; I should post some more about that ). The first NEWS entry follow.

Changes in version 1.2.0 (2025-10-03)
  • New maintainer
  • Compiled functions are now registered, .Call() adjusted
  • README.md and DESCRIPTION edited and update
  • Simple unit tests and continunous integration have been added
  • The package documentation site has been recreated using altdoc
  • All manual pages for functions now contain \value sections

For more, see the package page, the git repo or the documentation site.

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.

8 October 2025

Dirk Eddelbuettel: RPushbullet 0.3.5: Mostly Maintenance

RPpushbullet demo A new version 0.3.5 of the RPushbullet package arrived on CRAN. It marks the first release in 4 1/2 years for this mature and feature-stable package. RPushbullet interfaces the neat Pushbullet service for inter-device messaging, communication, and more. It lets you easily send (programmatic) alerts like the one to the left to your browser, phone, tablet, or all at once. This releases reflects mostly internal maintenance and updates to the documentation site, to continuous integration, to package metadata, and one code robustitication. See below for more details.

Changes in version 0.3.5 (2025-10-08)
  • URL and BugReports fields have been added to DESCRIPTION
  • The pbPost function deals more robustly with the case of multiple target emails
  • The continuous integration and the README badge have been updated
  • The DESCRIPTION file now use Authors@R
  • The (encrypted) unit test configuration has been adjusted to reflect the current set of active devices
  • The mkdocs-material documentation site is now generated via altdoc

Courtesy of my CRANberries, there is a diffstat report relative to previous release. More detailed information is on the repo where comments and suggestions are welcome.

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.

3 October 2025

Dirk Eddelbuettel: #053: Adding llvm Snapshots for R Package Testing

Welcome to post 53 in the R4 series. Continuing with posts #51 from Tuesday and #52 from Wednesday and their stated intent of posting some more here is another quick one. Earlier today I helped another package developer who came to the r-package-devel list asking for help with a build error on the Fedora machine at CRAN running recent / development clang. In such situations, the best first step is often to replicate the issue. As I pointed out on the list, the LLVM team behind clang maintains an apt repo at apt.llvm.org/ making it a good resource to add to Debian-based container such as Rocker r-base or the offical r-base (the two are in fact interchangeable, and I take care of both). A small pothole, however, is that the documentation at the top of apt.llvm.org site is a bit stale and behind two aspects that changed on current Debian systems (i.e. unstable/testing as used for r-base). First, apt now prefers files ending in .sources (in a nicer format) and second, it now really requires a key (which is good practice). As it took me a few minutes to regather how to meet both requirements, I reckoned I might as well script this. Et voil the following script does that:
#!/bin/sh

## Update does not hurt but is not strictly needed
#apt update --quiet --quiet
#apt upgrade --yes

## wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key   sudo tee /etc/apt/trusted.gpg.d/apt.llvm.org.asc
## or as we are root in container
wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key > /etc/apt/trusted.gpg.d/apt.llvm.org.asc

cat <<EOF >/etc/apt/sources.list.d/llvm-dev.sources
Types: deb
URIs: http://apt.llvm.org/unstable/
# for clang-21 
#   Suites: llvm-toolchain-21
# for current clang
Suites: llvm-toolchain
Components: main
Signed-By: /etc/apt/trusted.gpg.d/apt.llvm.org.asc
EOF

test -d ~/.R   mkdir ~/.R
cat <<EOF >~/.R/Makevars
CLANGVER=-22
# CLANGLIB=-stdlib=libc++
CXX=clang++\$(CLANGVER) \$(CLANGLIB)
CXX11=clang++\$(CLANGVER) \$(CLANGLIB)
CXX14=clang++\$(CLANGVER) \$(CLANGLIB)
CXX17=clang++\$(CLANGVER) \$(CLANGLIB)
CXX20=clang++\$(CLANGVER) \$(CLANGLIB)
CC=clang\$(CLANGVER)
SHLIB_CXXLD=clang++\$(CLANGVER) \$(CLANGLIB)
EOF

apt update
apt install --yes clang-22
Once the script is run, one can test a package (or set of packages) against clang-22 and clang++-22. This may help R package developers. The script is also generic enough for other development communities who can ignore (or comment-out / delete) the bit about ~/.R/Makevars and deploy the compiler differently. Updating the softlink as apt-preferences does is one way and done in many GitHub Actions recipes. As we only need wget here a basic Debian container should work, possibly with the addition of wget. For R users r-base hits a decent sweet spot.

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.

2 October 2025

Dirk Eddelbuettel: tinythemes 0.0.4 at CRAN: Micro Maintenance

tinythemes demo A tiniest of tiny violins micro maintenance release 0.0.4 of our tinythemes arrived on CRAN today. tinythemes provides the theme_ipsum_rc() function from hrbrthemes by Bob Rudis in a zero (added) dependency way. A simple example is (also available as a demo inside the package) contrasts the default style (on the left) with the one added by this package (on the right): This version adjusts to the fact that hrbrthemes is no longer on CRAN so the help page cannot link to its documentation. No other changes were made. The set of changes since the last release follows.

Changes in tinythemes version 0.0.4 (2025-10-02)
  • No longer \link to now-archived hrbrthemes

Courtesy of my CRANberries, there is a diffstat report relative to previous release. More detailed information is on the repo where comments and suggestions are welcome.

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.

1 October 2025

Dirk Eddelbuettel: #052: Running r-ci with Coverage

Welcome to post 52 in the R4 series. Following up on the post #51 yesterday and its stated intent of posting some more here The r-ci setup (which was introduced in post #32 and updated in post #45) offers portable continuous integration which can take advantage of different backends: Github Actions, Azure Pipelines, GitLab, BitBuckets, and possibly as it only requires a basic Ubuntu shell after which it customizes itself and runs via shell script. Portably. Now, most of us, I suspect still use it with GitHub Actions but it is reassuring to know that one can take it elsewhere should the need or desire arise. One thing many repos did, and which stopped working reliably is coverage analysis. This is made easy by the covr package, and made fast, easy, reliable (as we quip) thanks to r2u. But transmission to codecov stopped working a while back so I had mostly commented it out in my repos, rendering the reports more and more stale. Which is not ideal. A few weeks ago I gave this another look, and it turns that that codecov now requires an API token to upload. Which one can generate in the user -> settings menu on their website under the access tab. Which in turn then needs to be stored in each repo wanting to upload. For Github, this is under settings -> secrets and variables -> actions as a repository secret . I suggest using CODECOV_TOKEN for its name. After that the three-line block in the yaml file can reference it as a secret as in the following snippet, now five lines, taken from one of my ci.yaml files:
      - name: Coverage
        if: $  matrix.os == 'ubuntu-latest'  
        run: ./run.sh coverage
        env:
          CODECOV_TOKEN: $  secrets.CODECOV_TOKEN  
It assigns the secret we stored on the website, references it by prefix secrets. and assigns it to the environment variable CODECOV_TOKEN. After this reports flow as one can see on repositories where I re-enabled this as for example here for RcppArmadillo.

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.

30 September 2025

Dirk Eddelbuettel: #051: A Neat Little Rcpp Trick

Welcome to post 51 in the R4 series. A while back I realized I should really just post a little more as not all post have to be as deep and introspective as for example the recent-ish two cultures post #49. So this post is a neat little trick I (somewhat belatedly) realized somewhat recently. The context is the ongoing transition from (Rcpp)Armadillo 14.6.3 and earlier to (Rcpp)Armadillo 15.0.2 or later. (I need to write a bit more about that, but that may require a bit more time.) (And there are a total of seven (!!) issue tickets managing the transition with issue #475 being the main parent issue, please see there for more details.) In brief, the newer and current Armadillo no longer allows C++11 (which also means it no longer allowes suppression of deprecation warnings ). It so happens that around a decade ago packages were actively encouraged to move towards C++11 so many either set an explicit SystemRequirements: for it, or set CXX_STD=CXX11 in src/Makevars .win . CRAN has for some time now issued NOTEs asking for this to be removed, and more recently enforced this with actual deadlines. In RcppArmadillo I opted to accomodate old(er) packages (using this by-now anti-pattern) and flip to Armadillo 14.6.3 during a transition period. That is what the package does now: It gives you either Armadillo 14.6.3 in case C++11 was detected (or this legacy version was actively selected via a compile-time #define), or it uses Armadillo 15.0.2 or later. So this means we can have either one of two versions, and may want to know which one we have. Armadillo carries its own version macros, as many libraries or projects do (R of course included). Many many years ago (git blame points to sixteen and twelve for a revision) we added the following helper function to the package (full source here, we show it here without the full roxygen2 comment header)
// [[Rcpp::export]]
Rcpp::IntegerVector armadillo_version(bool single)  

    // These are declared as constexpr in Armadillo which actually does not define them
    // They are also defined as macros in arma_version.hpp so we just use that
    const unsigned int major = ARMA_VERSION_MAJOR;
    const unsigned int minor = ARMA_VERSION_MINOR;
    const unsigned int patch = ARMA_VERSION_PATCH;

    if (single)  
        return Rcpp::wrap(10000 * major + 100 * minor + patch) ;
      else  
        return Rcpp::IntegerVector::create(Rcpp::Named("major") = major,
                                           Rcpp::Named("minor") = minor,
                                           Rcpp::Named("patch") = patch);
     
 
It either returns a (named) vector of the standard major , minor , patch form of the common package versioning pattern, or a single integer which can used more easily in C(++) via preprocessor macros. And this being an Rcpp-using package, we can of course access either easily from R:
> library(RcppArmadillo)
> armadillo_version(FALSE)
major minor patch 
   15     0     2 
> armadillo_version(TRUE)
[1] 150002
>
Perfectly valid and truthful. But cumbersome at the R level. So when preparing for these (Rcpp)Armadillo changes in one of my package, I realized I could alter such a function and set the S3 type to package_version. (Full version of one such variant here)
// [[Rcpp::export]]
Rcpp::List armadilloVersion()  
    // create a vector of major, minor, patch
    auto ver = Rcpp::IntegerVector::create(ARMA_VERSION_MAJOR, ARMA_VERSION_MINOR, ARMA_VERSION_PATCH);
    // and place it in a list (as e.g. packageVersion() in R returns)
    auto lst = Rcpp::List::create(ver);
    // and class it as 'package_version' accessing print() etc methods
    lst.attr("class") = Rcpp::CharacterVector::create("package_version", "numeric_version");
    return lst;
 
Three statements each to and return the value. And now in R we can operate more easily on this (using three dots as I didn t export it from this package):
> RcppDE:::armadilloVersion()
[1]  15.0.2 
> RcppDE:::armadilloVersion() >= "15.0.0"
[1] TRUE
>
An object of class package_version inheriting from numeric_version can directly compare against a (human- but not normally machine-readable) string like 15.0.0 because the simple S3 class defines appropriate operators, as well as print() / format() methods as the first expression shows. It is these little things that make working with R so smooth, and we can easily (three statements !!) do so from Rcpp-based packages too. The underlying object really is merely a list containing a vector:
> unclass(RcppDE:::armadilloVersion())
[[1]]
[1] 15  0  2

> 
but the S3 glue around it makes it behave nicely. So next time you are working with an object you plan to return to R, consider classing it to take advantage of existing infrastructure (if it exists, of course). It s easy enough to do, and may smoothen the experience at the R side.

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.

25 September 2025

Dirk Eddelbuettel: tint 0.1.6 on CRAN: Maintenance

A new version 0.1.6 of the tint package arrived at CRAN today. tint provides a style not unlike Tufte for use in html and pdf documents created from markdown. The github repo shows several examples in its README, more as usual in the package documentation. This release addresses a small issue where in pdf mode, pandoc (3.2.1 or newer) needs a particular macros defined when static (premade) images or figure files are included. No other changes.

Changes in tint version 0.1.6 (2025-09-25)
  • An additional LaTeX command needed by pandoc (>= 3.2.1) has been defined for the two pdf variants.

Courtesy of my CRANberries, there is a diffstat report relative to previous release. More information is on the tint page. For questions or comments use the issue tracker off 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.

21 September 2025

Dirk Eddelbuettel: rcppmlpackexamples 0.0.1 on CRAN: New Package

mlpack is a fabulous project providing countless machine learning algorithms in clean and performant C++ code as a header-only library. This gives both high performance and the ability to run the code in resource-constrained environment such as embedded systems. Bindings to a number of other languages are available, and an examples repo provides examples. The project also has a mature R package on CRAN which offers the various algorithms directly in R. Sometimes, however, one might want to use the header-only C++ code in another R package. How to do that was not well documented. A user alerted me by email to this fact a few weeks ago, and this lead to both an updated mlpack release at CRAN and this package. In short, we show via three complete examples how to access the mlpack code in C++ code in this package, offering a re-usable stanza to start a different package from. The only other (header-only) dependencies are provided by CRAN packages RcppArmadillo and RcppEnsmallen wrapping, respectively, the linear algebra and optimization libraries used by mlpack. Courtesy of my CRANberries, there is also a a new package note (no diffstat report yet). More detailed information is on the rcppmlpackexamples page, or 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.

Next.

Previous.