November 2025 - Boost - lists.boost.org

[release] Boost 1.90.0 beta 1 released
by Marshall Clow 14 Nov '25

14 Nov '25

Boost release 1.90.0 beta 1 is now available at: https://archives.boost.io/beta/1.90.0.beta1/source/ The SHA256 checksums are: 8072831c2c54c7e0e99c69e934b5694574565dd435de19a0a43f7bf7dabad71b boost_1_90_0_b1.zip d1754f544e83418c2b5fe18245666830088948cb3629d2b346d56f31c71138e5 boost_1_90_0_b1.tar.bz2 c94235eb00edc732253d3ae8000a489ffc5773f34cd98bd7645d54cb4ae2482a boost_1_90_0_b1.7z 0f55a646eaed4daf110eeec7c2c70e2c7058efa77eef31137248a6e2cf1412da boost_1_90_0_b1.tar.gz Please try the beta and report any problems you encounter. Thanks to everyone who contributed to the release. -- The release managers

5 5

Boost.Decimal Review Result.
by John Maddock 10 Nov '25

10 Nov '25

First off a big thanks to all the folks who participated in the review process, and of course to Matt and Chris for submitting the library. Then the headline news: with 8 favorable reviews, and two equivocal ones, Boost.Decimal is ACCEPTED into Boost. There were two main contentious issues: 1) Cohort support: several reviewers asked for cohort support and one thought that printing the actual cohort should be the default. I know that Matt has been working on support for cohort IO, on the basis that the notion of cohorts take some getting used to, and that libraries should generally make the easy stuff easy, I' happy for this to NOT be the default. 2) Binary-decimal conversion. This was discussed in some depth during the review, so I won't try to repeat here. Matt has agreed to document the issue, and to make binary-decimal conversion non-straightforward (ie strongly discouraged), I'm also satisfied that the interface can be extended (in a non-breaking way) to support multiple conversion methods at a later date should there prove to be demand. Given the difficulty of implementing those methods correctly, this sounds like a reasonable compromise to me at present. I've filed all the issues raised by reviewers as issues so folks can track their progress, they are: * Cohort IO support: https://github.com/cppalliance/decimal/issues/1218 * Document what a normalized number is: https://github.com/cppalliance/decimal/issues/1217 * Document samequantum behavior: https://github.com/cppalliance/decimal/issues/1216 * Do we need issignalling, is the signature correct: https://github.com/cppalliance/decimal/issues/1215 * Better document {fmt} support: https://github.com/cppalliance/decimal/issues/1214 * Financial examples need explanatory text: https://github.com/cppalliance/decimal/issues/1213 * Document what bit conversion do: https://github.com/cppalliance/decimal/issues/1212 * Explain generic programming example, remove/rename float_equal function: https://github.com/cppalliance/decimal/issues/1211 * Improve charconv docs: https://github.com/cppalliance/decimal/issues/1210 * Clean up badly named "promotion" section in the docs: https://github.com/cppalliance/decimal/issues/1209 * Improve literal docs (and float_equal again): https://github.com/cppalliance/decimal/issues/1208 * Improve construction docs: https://github.com/cppalliance/decimal/issues/1207 * Improve examples with better hand holding rather than bare code: https://github.com/cppalliance/decimal/issues/1206 * Improve readability of binary constants: https://github.com/cppalliance/decimal/issues/1205 * Probably remove the bitwise operators: https://github.com/cppalliance/decimal/issues/1204 * Simplify numeric_limits, provide out-of-line data members: https://github.com/cppalliance/decimal/issues/1203 * Misc doc typos and corrections: https://github.com/cppalliance/decimal/issues/1202 * Correct <format> detection: https://github.com/cppalliance/decimal/issues/1201 * {fmt} unicode support: https://github.com/cppalliance/decimal/issues/1200 * Arm64 detection: https://github.com/cppalliance/decimal/issues/1199 * Codecov markup: https://github.com/cppalliance/decimal/issues/1198 * assert in charconv example: https://github.com/cppalliance/decimal/issues/1197 * Improve example descriptions: https://github.com/cppalliance/decimal/issues/1196 * More precise header usage in examples: https://github.com/cppalliance/decimal/issues/1195 * Public to_string function is undocumented: https://github.com/cppalliance/decimal/issues/1194 * Mistaken using declaration in docs (make the documentation code runable as part of the tests): https://github.com/cppalliance/decimal/issues/1193 * Testing only code in public headers: https://github.com/cppalliance/decimal/issues/1192 * charconv docs do not compile (T undefined):https://github.com/cppalliance/decimal/issues/1191 * Binary - decimal conversions: https://github.com/cppalliance/decimal/issues/1188 * Convenience string constructors: https://github.com/cppalliance/decimal/issues/1187 * Make sure decTest tests pass: https://github.com/cppalliance/decimal/issues/1186 * Rounding docs: https://github.com/cppalliance/decimal/issues/1185 Once again, many thanks to everyone involved, best, John.

5 4

[container] C++23 / C++26 Heterogeneous overloads (P2077+P2363)
by Ion Gaztañaga 10 Nov '25

10 Nov '25

Hi, In case anyone is missing them, those overloads for to [flat_][multi]set/[multi]map were recently pushed to the develop branch. Since the window for major code changes in the master branch closed at the end of October, if anyone thinks it would be valuable to have them in Boost 1.90, I'd need to ask permission to Marshall. Otherwise they will be generally available in Boost 1.91 (scheduled for April 2026). Best, Ion

1 0

[release] Boost 1.90.0 Beta 1 Release Candidate 1 is available
by Marshall Clow 09 Nov '25

09 Nov '25

The first release candidates for the 1.90.0 Beta 1 release are now available at: <https://archives.boost.io/beta/1.90.0.beta1/source/> The SHA256 checksums are as follows: 0f55a646eaed4daf110eeec7c2c70e2c7058efa77eef31137248a6e2cf1412da boost_1_90_0_b1_rc1.tar.gz c94235eb00edc732253d3ae8000a489ffc5773f34cd98bd7645d54cb4ae2482a boost_1_90_0_b1_rc1.7z 8072831c2c54c7e0e99c69e934b5694574565dd435de19a0a43f7bf7dabad71b boost_1_90_0_b1_rc1.zip d1754f544e83418c2b5fe18245666830088948cb3629d2b346d56f31c71138e5 boost_1_90_0_b1_rc1.tar.bz2 As always, the release managers would appreciate it if you download the candidate of your choice and give building it a try. Please report both success and failure, and anything else that is noteworthy. -- The Release managers

6 6

[move] Why not use std::addressof when available for boost::move_detail::addressof rather than undocumented __builtin_addressof
by travisp＠mailhaven.com 06 Nov '25

06 Nov '25

Hi all, I noticed that a change was made to boost::move_detail::addressof implementation to address some potential GCC compiler alignment warnings: https://github.com/boostorg/move/commit/7637429b6c361321cf0512014949f86be32… The new implementation uses GCC's __builtin_addressof (which remains purposefully undocumented, https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96634 ), or uses a clever series of casts. And it doesn't take advantage of the potential existence of std::addressof. Is there any reason why not use std::addressof when it is available (since C++11)? (Aside: I ran across this because I have an unusual situation, where the new clever casts are an issue (but the previous clever implementation was alright, but presumably caused strict-alignment warnings), and using __builtin_addressof is an issue too. In all cases, std::addressof will work fine for me and seems like it would be preferable for all, when available.) -Travis

2 1

[release] Master branch closes TODAY for the 1.90.0 beta release
by Marshall Clow 05 Nov '25

05 Nov '25

As always, you can find the calendar here: https://www.boost.org/calendar/ — The release managers

1 0

moving TypeTraits to the boost::type_traits namespace
by Gero Peterhoff 05 Nov '25

05 Nov '25

I think we should finally move TypeTraits to the boost::type_traits namespace. This has only advantages: - clear separation from other boost libraries - the boost namespace is no longer contaminated - when providing additional TypeTraits, it is immediately clear in which namespace they must be implemented These changes would of course be elaborate, but they would bring advantages: consolidation + general review of existing TypeTraits: - are they still necessary / could they be removed? better to use std::TypeTraits - could they be implemented in a more contemporary way? Gero

2 1

boost.parser: Hitting the wall; what am I doing wrong?
by Richard 03 Nov '25

03 Nov '25

Hi Boosters, I've hit a wall while using boost.parser to parse a fractal formula file format. At this point, when I attempt to add a new rule to my grammar, msvc goes to compile the file and just sits there. It isn't hammering CPU, memory or disk, just sitting there. I let it run for like 30 minutes before I aborted it. There's no error or warning diagnostic. The details are in this issue on boost.parser: <https://github.com/boostorg/parser/issues/288> I believe separate compilation of the rules implementing my grammar would work around this problem (and give me better turnaround time on edits as I'm not changing the entire grammar as I improve it), but I was unable to get separate compilation of rules to work with boost.parser. Is separate compilation even possible? -- Richard -- "The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline> The Terminals Wiki <http://terminals-wiki.org> The Computer Graphics Museum <http://ComputerGraphicsMuseum.org> Legalize Adulthood! (my blog) <http://LegalizeAdulthood.wordpress.com>

1 0

The Master branch will be closing for the 1.90.0 beta release on Wednesday (5-Nov)
by Marshall Clow 02 Nov '25

02 Nov '25

As always, you can find the calendar here: https://www.boost.org/calendar/ — The release managers

1 0