- Boost - lists.boost.org

Introducing compile_assert C++ Proposal P4021
by Jonathan Grant 02 Nov '26

02 Nov '26

Hello I'd like to introduce my P4021 - compile_assert, compile-time diagnostics for expressions that can be proven at compile time. ie they are not const, constexpr, consteval. https://isocpp.org/files/papers/P4021R2.pdf There's a github repo with a testsuite, showing lots of simple examples. It's a useful tool to detect issues at compile time, like nullptr dereferences, divide by zero, array subscript errors. It diagnoses invariants that are proven false, as well as invariants that are not-proven. Helps make software more robust. Handy for Functional Safety. It does this by identifying code that is reachable, but should not be. Could something like this be added to Boost? I also hope tools like Claude could utilise compile_assert() to evaluate open source software. Regards Jonathan

2 3

Is it time to raise our required CMake version?
by Peter Dimov 02 Nov '26

02 Nov '26

At the moment, our required CMake version is conservatively set to 3.8, and all libraries are recommended to support it. This may be set too low for 2027. The minimum usable Ubuntu on Github Actions is 20.04, which has CMake 3.16. Boost installation requires 3.13. Boost.Python requires 3.14. (These are preexisting requirements, from the very beginning.) Opinions? Does anyone need CMake < 3.13? < 3.16?

5 6

Last week in github.com/boostorg (19-25 October 2026)
by Boost Fiscal Sponsorship Committee 02 Nov '26

02 Nov '26

Hi all, Activity on the develop branches of the Boost libraries between 19 and 25 October. This report has been generated automatically with AI. THE NUMBERS - 260 commits on develop (21 CI-related) - 46 pull requests opened against develop, 57 merged - 14 issues opened, 28 closed - 35 libraries with activity, 30 contributors Most active libraries: - Asio: 72 commits PRs: 0 opened, 0 merged; issues: 0 opened, 0 closed - Graph: 17 commits PRs: 16 opened, 15 merged; issues: 1 opened, 0 closed - Container: 37 commits PRs: 0 opened, 0 merged; issues: 0 opened, 1 closed - Fiber: 18 commits PRs: 0 opened, 7 merged; issues: 0 opened, 1 closed - IOStreams: 16 commits PRs: 1 opened, 0 merged; issues: 0 opened, 4 closed - Interval: 7 commits PRs: 4 opened, 4 merged; issues: 2 opened, 3 closed - Intrusive: 20 commits PRs: 0 opened, 0 merged; issues: 0 opened, 0 closed - Decimal: 5 commits PRs: 2 opened, 5 merged; issues: 2 opened, 5 closed Most active contributors: - Christopher Kohlhoff (72 commits, 1 library) - Ion Gaztañaga (64 commits, 4 libraries) - Mateusz Jakub Fila (17 commits, 1 library) - Tobias Stensbeck (16 commits, 1 library) - Peter Dimov (14 commits, 2 libraries) - Jean-Louis Leroy (10 commits, 1 library) - Arnaud Becheler (9 commits, 2 libraries) - Andrea Cassioli (7 commits, 1 library) HIGHLIGHTS Asio Christopher Kohlhoff pushed 72 commits on 21 October. New features include asio::eager, a completion token that starts an operation at once and returns a handle to await later; asio::trampoline, an executor adapter that batches nested submissions; a dependency-free spawn() for x86-64 and AArch64 on Linux and macOS (Boost.Context stays the default); a "spin" reactor backend that avoids epoll_wait-style syscalls; and Perfetto-format handler tracking. channel and concurrent_channel move to the asio namespace, and ready() is deprecated in favour of can_receive(). Fixes cover a possible deadlock in the reactor, a hang in cancel_after/cancel_at with reentrant cancellation and duplicate work counting. poll() now runs the scheduler's task at most once per call. The rest is mostly performance work. Intrusive, Container and Move Ion Gaztañaga reviewed exception guarantees across the three libraries. In Intrusive, move assignment now copies the predicates and moves the nodes instead of swapping, a potentially breaking change; BOOST_INTRUSIVE_MOVE_ASSIGNMENT_SWAPS restores the old behaviour. Every Intrusive container got a friend swap, and swaps and move constructors got conditional noexcept. Container got many noexcept corrections and allocator fixes: private_adaptive_pool::allocate not throwing bad_alloc, a division by zero in alignment-only adaptive pools, an overflow in the aligned_allocate fallback, a string bug when moving to the internal buffer, and devector reallocating despite enough capacity. Move fixed is_constructible on GCC 8-11 in C++03 mode. Graph Arnaud Becheler fixed binary serialization of adjacency_list, whose layout changed by accident in 1.92: archives written by 1.92 can't be read by other releases, and 1.93 goes back to the 1.91 layout. He also extended stateful visitors to planar face traversal, make_maximal_planar, make_connected and make_biconnected_planar, and moved the distributed parts of BGL into GraphParallel. Andrea Cassioli, Varun Sharma and Sayan Samanta removed more legacy compatibility code. New PRs: #664, deprecating labeled and named graphs; #656, Kerem Okyay's fix for #655, a colour map from boykov_kolmogorov_max_flow that is not a minimum cut. Charconv Michiel van Slobbe fixed five writes past the end of a buffer in to_chars (#314 to #318). Several paths wrote a minus sign or a zero without checking the room left: into an empty range, for negative 128-bit integers and for negative zero. Some of them then reported success. General format with a large precision overflowed a stack buffer by one byte for some negative values. Decimal Shen-Ta Hsieh fixed a series of defects he had reported: fmod, remainder and remquo are now exact for exponents far apart (#1517), scalbn and scalbln handle large exponents, acosh and atanh give NaN outside their domain, fma no longer drops a small addend in a tie, and fdim is right for infinities and signaling NaNs. New PRs: #1529, for #1528, equal infinities with payload bits hashing differently. Interval Square roots of point intervals could miss the true root. Sonny Guardian stopped GCC with -frounding-math from reusing the downward-rounded root for the upper end (#61, fixes #60). Sayan Samanta made 32-bit MSVC use x87 fsqrt or SSE, because its runtime sqrt ignores the rounding mode (#64), and Andrey Semashev optimised the SSE2 path. Sayan Samanta also kept GCC from moving x87 arithmetic across the rounding-mode change (#62, fixes #41). Math Nick Thompson added two global optimisers, multilevel coordinate search (closes #1577) and TuRBO, a trust-region Bayesian method, and removed Eigen from CMA-ES. Jacob Hass merged Gauss-Laguerre and Gauss-Jacobi quadrature with anti-Gauss error estimates (#1517, 33 comments). New issues: #1585, documentation for the new quadratures; #1586, hermite, jacobi and laguerre zeros as special functions. JSON Dmitry Arkhipov merged a refactor of how users customise conversion categories (#1156) and ranked described classes lowest ahead of reflection (#1205). New PRs: #1211, splitting a parse_into test that needed over 3GB to build; #1213, for #1212, an option to stop relying on the no_destroy attribute. New issues: #1212, a linker error with Clang on Windows against an MSVC-built Boost (6 comments); #1214, documenting the config macros. Fiber Mateusz Jakub Fila merged CUDA and HIP work from August and September: the HIP integration now uses its own namespace, callbacks use LaunchHostFunc where available, and cudaLaunchHostFunc_v2 flags can be passed (#350). Helena Pankov's futex support for macOS 14.4 and later (#330), first written in April 2025, was merged too. Also last week * UUID and System: Peter Dimov added std::format support to UUID (fixes #201) and made System's to_string_formatter conform to std::formattable. * Beast: Mohammad Nejati's permessage-deflate offer no longer sends server_max_window_bits=15, which some servers reject (fixes #2671), and sends client_max_window_bits without a value. * MSM: Christian Granzin removed the experimental status of backmp11, now the recommended back-end, and removed its active state switch policies and deprecated process_result values. * OpenMethod: Jean-Louis Leroy fixed or annotated the open Coverity defects and merged named_static_rtti for class identity across shared libraries (#127). New issues: #137, static validation of closed overrider sets. * IOStreams: Tobias Stensbeck made the C++11 and CMake 3.8 requirements explicit and fixed test warnings. New PRs: #195, a buffer overrun in range_adapter, from @aouxwoux. * Lockfree: Michiel van Slobbe stopped Clang turning spsc_queue's index wrap into a division on every push and pop (#144). * Config and Predef: Matt Turner added Alpha to the architecture checks. * Wave: Martin Medler added a C++14 language mode. New PRs: #276, a C++26 mode. * New issues and PRs: Process PR #571 (OpenBSD descriptor cleanup), Redis PR #441 (moved-from flat_tree), ProgramOptions PR #161 (C++20 module), GIL issue #817 (Iterator and MPL dependencies). CI MAINTENANCE * Boost.CI: Alexander Grund added Coveralls as a coverage option, made the OpenCppCoverage install more reliable and fixed the Xcode SDK setup; Jean-Louis Leroy fixed Homebrew Clang on macos-26 (#353). * IOStreams: Tobias Stensbeck overhauled the workflow: containers for old compilers, retired runners dropped, GCC 15/16, Clang 20-22, MSVC 14.5 and macOS 26 added, Travis and AppVeyor removed. * Container, Intrusive, Move and Interprocess: Ion Gaztañaga added CMake-based test runs to GitHub Actions. * Interval: Sayan Samanta fixed link failures with Homebrew Clang on macos-26. * UUID: Peter Dimov added libc++ testing. Have a good week!

8 18

Last week in github.com/boostorg (26 October - 1 November 2026)
by Boost Fiscal Sponsorship Committee 02 Nov '26

02 Nov '26

Hi all, Activity on the develop branches of the Boost libraries between 26 October and 1 November. This report has been generated automatically with AI. THE NUMBERS - 227 commits on develop (15 CI-related) - 108 pull requests opened against develop, 92 merged - 55 issues opened, 80 closed - 55 libraries with activity, 35 contributors Most active libraries: - Graph: 63 commits PRs: 22 opened, 31 merged; issues: 5 opened, 33 closed - Math: 8 commits PRs: 10 opened, 8 merged; issues: 17 opened, 18 closed - Container: 60 commits PRs: 0 opened, 0 merged; issues: 0 opened, 0 closed - MySQL: 13 commits PRs: 15 opened, 13 merged; issues: 5 opened, 2 closed - Decimal: 10 commits PRs: 2 opened, 3 merged; issues: 7 opened, 8 closed - Redis: 8 commits PRs: 6 opened, 8 merged; issues: 0 opened, 1 closed - Interval: 7 commits PRs: 5 opened, 4 merged; issues: 2 opened, 1 closed - Wave: 10 commits PRs: 2 opened, 2 merged; issues: 2 opened, 2 closed Most active contributors: - Ion Gaztañaga (65 commits, 4 libraries) - Arnaud Becheler (38 commits, 1 library) - Rubén Pérez (20 commits, 3 libraries) - Andrea Cassioli (14 commits, 1 library) - Christopher Kohlhoff (10 commits, 1 library) - Nick Thompson (9 commits, 2 libraries) - Shen-Ta Hsieh (8 commits, 1 library) - Sayan Samanta (7 commits, 1 library) HIGHLIGHTS Graph Arnaud Becheler (38 commits) fixed many old bugs, among them a heap overflow when copying an undirected graph into a CSR graph (#670), a read of a destroyed edge bundle in CSR add_edges (#672), strong_components breaking after including reverse_graph.hpp (#116) and tiernan_all_cycles rejecting adjacency_list (#182). connected_components now uses a colour map about 16 times smaller. labeled_graph and named_graph are deprecated, to be removed in Boost 1.95 (#664), and building the library now requires C++14. Varun Sharma added serialization for adjacency_matrix and the CSR graph (#650), and Kerem Okyay fixed the colour map of boykov_kolmogorov_max_flow (#655). Most of the 33 closed issues are feature requests gathered into tracking issue #679. New PRs: #697, parallel edges in biconnected_components. New issues: #678, bellman_ford_shortest_paths missing negative cycles on overflow. Container and Move Ion Gaztañaga made 60 commits in Container, nearly all bug fixes. memmove was being used for types with user-defined move operations, small_vector lost its internal buffer when an assignment changed the allocator, synchronized_pool_resource observers didn't lock the mutex, and monotonic_buffer_resource::release didn't restore the initial buffer size. Flat containers never found unused storage for sorting and merging, so they always used the slower bufferless path, and their erase_if was not linear. basic_string got fixes to operator>>, getline and deduction guides. In Move, iter_value is now SFINAE-friendly. Math Nick Thompson packaged the Python bindings as a wheel (Python's ready-to-install package format, installed with pip) and used them to cross-check Boost.Math against SciPy and mpmath. That found the defects he fixed in #1604 and #1607, among them wrong z-test p-values, a heap overflow in naive_monte_carlo, and hangs and wrong results in several special functions. He also added nested Leja quadrature with a dimension-adaptive sparse grid, hermite_zeros, laguerre_zeros and jacobi_zeros, and the DYCORS optimiser. New PRs: #1611, a WebAssembly build; #1614, end points of open Catmull-Rom curves. New issues: #1612, hypergeometric_0F1 losing about 20 ulps. Asio Christopher Kohlhoff added batched datagram I/O: send_multiple, receive_multiple and their _to/_from and asynchronous forms, using sendmmsg and recvmmsg where available, or a single datagram per call elsewhere. Immediate completion now respects the handler's associated executor and works with a TS executor as the I/O executor. Other fixes cover the error category of IOCP operations and the assembly fiber code (MSVC trampoline, CET guard). MySQL and Redis Rubén Pérez fixed two race conditions in MySQL's connection_pool: one with immediate completions and one where a cancelled async_get_connection could lose the notification that a connection had become free (#536). He also moved Redis to the stable asio::channel. In Redis, @aouxwoux fixed undefined behaviour in the move operations of flat_tree and its size tracking (#441, fixes #440), and Marcelo Zimbres fixed the protobuf examples. Decimal Shen-Ta Hsieh fixed the text output in #1536: to_chars rounded some values the wrong way and didn't match printf %g, snprintf could write past its buffer, operator<< ignored stream flags, and std::format didn't compile in C++20. std::hash now gives equal infinities the same hash (#1529). Matt Borland fixed a fmt FMT_COMPILE build failure with wide characters (#1538). Wave Martin Medler made __cplusplus correct for every language standard (#264) and added C++14 and C++26 modes. Jeff Trull made macros that expand to defined and __has_include work and fixed Windows link errors in the samples and tool. Alexander Grund removed a wrong dependency on Xpressive from build.jam (#277). New PRs: #280, C++23 #elifdef and #elifndef. New issues: #281, how the language level is checked in the code. ICL Six new issues report wrong results for descending intervals and overflow: cardinality, length and distance give wrong values or undefined behaviour (#63, #64, #65, #66), cardinality's return type is too small (#61), and touches is poorly specified (#62). New PRs: #60 from Treston Cuzzort, for #41 and #63. Geometry No commits, four new PRs: #1499 (Pulkit Chauhan), undefined behaviour in sin_cos_degrees below -45 degrees; #1500 (@gastald88), a buffer that disappears (#1428); #1502 (@Osyotr), a macro making rtree packing deterministic across compilers; #1505 (Marcel Jacobse), simplify_ring when buffering a ring. New issues: #1503, CMake test configuration missing the Interprocess dependency. Regex Two new PRs: #287 from Treston Cuzzort, for #241, so that sets such as [\W\D] match the union of the negated classes; and #289 from @ConnorQi01, for the new issue #288, keeping the POSIX C headers usable from C. Also last week * Beast: Mohammad Nejati made the parser reject malformed Transfer-Encoding values (fixes #3123), handled repeated Transfer-Encoding fields, and refreshed the example root certificates, leaving out the emSign roots after the eMudhra compromise. New issues: #3124, repeated Connection fields. * JSON: Dmitry Arkhipov fixed to_number rejecting exactly -2^63 for int64_t (#1216) and made reliance on the no_destroy attribute optional (#1213). New PRs: #1215, documentation updates. * Interval: Sayan Samanta stopped compilers from merging the rounded operations across rounding-mode changes (#65) and fixed interval<long double> with constants on PowerPC GCC (#69). New PRs: #71, fmod of long double intervals (#67). * Histogram: Henry Schreiner added a pick command to reduce, selecting arbitrary bins of an unordered axis (#429). Nathan Chandrasekar fixed unsigned integer axes putting values below the range into overflow (#447). * ProgramOptions: Vladimir Prus fixed error::what() looping forever when an invalid option contains %canonical_option% (#164). * Assert, OpenMethod, URL, StaticString and Unordered: Chris Glover's change that lets projects using add_subdirectory skip re-checking globbed headers on every build was merged in Assert and opened in the others. * New issues and PRs: Treston Cuzzort opened fixes in 15 more libraries, mostly for old issues, among them DateTime #254 (duration scalars narrowed to int), Format #112 and Chrono #84. Random PR #171 (Sebastiano Vigna) adds the MWCX generators. OpenMethod and TypeIndex have new issues about older CMake versions. CI MAINTENANCE * Graph: Arnaud Becheler added a sanitizer workflow with a read-only token (#551, #680) and moved off the macOS 14 images retired on 2 November. * MySQL: Rubén Pérez moved the Windows builds to new container images, added an MSVC 14.5 build and nightly runs, and opened #537 for gcc-16 and clang 21 and 22. * Interval: Sayan Samanta fixed the QEMU multiarch jobs on Ubuntu 26.04 and the icpx release links. * Wave: Jeff Trull moved to the Boost.CI reusable workflow and enabled MSVC jobs, replacing AppVeyor. * Geometry: Tinko Bartels opened #1504 to fix silently failing Clang jobs and other test problems. Have a good week!

1 0

capy: thread_pool use
by Brook Milligan 01 Nov '26

01 Nov '26

I am considering a use of capy that looks like the following: for (auto i = 0; i < m; ++i) { capy::thread_pool pool(0); for (auto j = 0; j < n; ++j) { capy::run_async(pool.get_executor())(task[j]()); } pool.join(); } I am wondering about the overhead of constructing/deconstructing a heap of thread pools in the outer loop. Does it make sense to pre-construct m of them prior to the loop and then use the ith one within the loop? Is there a better pattern to use for this? Is this the most efficient way to execute a bunch of independent tasks in parallel? This is the first time I am working through capy, so all suggestions are welcome. Thanks a lot. Cheers, Brook

3 5

Non-trivial build hang due to CONFIGURE_DEPENDS in CMake files
by Chris Glover 31 Oct '26

31 Oct '26

Hi, I was recently diagnosing a small build hang and traced it back to the use of CMake GLOB + CONFIGURE_DEPENDS in a handful of Boost libraries. Most of the use cases seem to be specifically for Visual Studio project generation which most libraries handle conditionally, but some do not. One use case (Boost.Url) is GLOBing sources. In my opinion CMake GLOB should not be used at all; instead explicitly listing all sources, including headers, should be the preferred pattern. Since that's a large change probably requiring some discussion, for now I've gone ahead and submitted the following PRs that just continue the pattern used elsewhere. https://github.com/boostorg/assert/pull/49 https://github.com/boostorg/unordered/pull/357 https://github.com/boostorg/openmethod/pull/138 https://github.com/boostorg/static_string/pull/83 https://github.com/boostorg/json/pull/1217 https://github.com/boostorg/url/pull/1015 Thank you, -- chris

2 7

[taar] Interest in an HTTP application library on Asio/Beast
by Reza Jahanbakhshi 31 Oct '26

31 Oct '26

Hello all, I would like to gauge interest in Boost.Taar, a C++20 header-only library for writing HTTP services on top of Boost.Asio, Boost.Beast, Boost.JSON and Boost.URL. How it started In 2021 a colleague challenged me: "In Python you can serve HTTP in one line (python -m http.server). Can you do that in C++?" With Beast you can build anything, but even a small JSON endpoint means writing your own routing, argument parsing, error mapping and connection loop. Taar is my answer to that challenge. It went through several shapes before I settled on the current design in 2024, and it has grown into a library since. The gap it fills Beast is deliberately low level: it gives you HTTP messages, parsers and streams, and leaves routing, parameter extraction, conversions, sessions, static files and protocol handling to the application. Every non-trivial Beast server rewrites those layers. Outside Boost there are frameworks (Crow, Drogon, Oat++, cpp-httplib, userver), but they bring their own networking, threading and JSON models and do not compose with Asio executors, coroutines, Boost.JSON or Boost.URL. Boost has nothing at the application level. Taar is that layer, built only from Boost components and in their idiom: - the server and every connection are Asio coroutines; handlers may be plain functions or awaitables; it runs on any Asio executor, with a strand per connection on a multi-threaded io_context; - a composable matcher DSL selects routes (method, path templates with placeholders, headers, cookies, version); - handlers are ordinary callables: arguments are extracted from the path, query, headers, cookies or body and converted to the parameter types, and return values are converted to responses (both through tag_invoke customisation points); - HTTP protocols sit on top of the same callables: a JSON-RPC 2.0 endpoint, a Model Context Protocol (MCP) server so C++ applications can expose tools and resources to AI agents, and an OpenAPI 3.1 document generated from the registered routes. The layering is strict: Asio/Beast/JSON/URL at the bottom, the HTTP framework above them, and JSON-RPC, MCP and OpenAPI as optional layers on the HTTP framework that do not affect users who only need REST. What it looks like The answer to the challenge, a static file server for the current directory: taar::session::http session; session.register_request_handler( method == http::verb::get && target == "/{*path}", taar::handler::htdocs {"."}); boost::asio::io_context io; taar::cancellation_signals signals; boost::asio::co_spawn(io, taar::server::tcp("0.0.0.0", "8080", session, signals), boost::asio::detached); io.run(); And one function served three ways: int add(int a, int b) { return a + b; } // REST: GET /add/2/3 -> 5, documented in the generated OpenAPI file session.register_request_handler( method == http::verb::get && target == "/add/{a}/{b}", taar::handler::rest(&add, path_arg("a"), path_arg("b")), openapi.describe("Add two numbers")); // JSON-RPC 2.0: {"method": "add", "params": {"a": 2, "b": 3}} rpc.method("add", &add, rpc_arg("a"), rpc_arg("b")); // MCP: a tool whose input schema is derived from the parameter types mcp.tool("add", "Add two numbers", &add, mcp_arg("a"), mcp_arg("b")); Beyond that - static files with ETag/Last-Modified and conditional requests, byte ranges, precompressed variants, cache rules and caches; - streaming responses (chunked, server-sent events) and streaming request bodies, multipart forms; - per-connection state, cookie or header keyed sessions, interceptors and route guards, CORS, TLS, connection timeouts, graceful shutdown; - one error model: taar::failure with nested causes, taar::result<T> (boost::system::result with a rich error) that throws on access, mapped per surface to HTTP statuses, JSON-RPC errors and MCP results; - MCP over Streamable HTTP and stdio, with sessions, progress, cancellation, server-to-client requests and resource subscriptions. Status Version 1.0 is released. The library has about 480 test cases (unit and end-to-end over TCP and TLS). C++20 is the minimum; CI runs the tests as C++20 on GCC 13, Clang 17 and 18 (libstdc++ and libc++), MSVC 2022 and Apple Clang, as C++23 on newer GCC, Clang, MSVC and Apple Clang, as C++26 on GCC 15 and Clang 22, and under ASan, UBSan and TSan. It builds with b2 in the Boost superproject and with CMake (including the Boost CMake integration tests), and has QuickBook documentation with a reference and quick reference. Questions I would value the list's opinion on - Is MCP appropriate inside a Boost library given how young the specification is? It is an optional layer and versioned by protocol revision, but I would like to hear concerns. - I plan to replace much of the type introspection with C++26 reflection once compilers ship it (parameter names, schemas without BOOST_DESCRIBE, annotations for descriptions and constraints), keeping the C++20 path. Is that a reasonable direction? - The name: "taar" is Persian for a strand or thread, as in a spider's web, and is also the name of a long-necked lute; a library of strands for the web seemed fitting. Is it acceptable as a Boost name? Links - Repository: https://github.com/rjahanbakhshi/boost.taar - Documentation: https://rjahanbakhshi.github.io/boost.taar/ - Examples: https://github.com/rjahanbakhshi/boost.taar/tree/master/example - Known gaps and roadmap: https://github.com/rjahanbakhshi/boost.taar/blob/master/ROADMAP.md If there is interest, I would be grateful for feedback on the design and API, and I am looking for a review manager. Thank you, Reza Jahanbakhshi

8 16

Unmaintained libraries: a tentative list
by Joaquin M López Muñoz 29 Oct '26

29 Oct '26

It is generally known within the community that some of the libraries in Boost are not being properly maintained: issues pile up without a prompt (or any) response from the nominal maintainers, eroding the libraries' usefulness and, ultimately, their viability. There are a number of stalwart contributors (Peter Dimov, Andrey Semashev, and others) who work to make sure CI tests stay green throughout the project, but beyond that they can't possibly address outstanding issues and evolution requests for all libraries. Any initiative to try to remedy this situation begins with identifying its true extent. My colleagues at the FSC and I have gone through all Boost libraries and identified those that we regard as unmaintained, as evidenced by the presence of many important issues that go unresolved without anyone addressing them. This is the list: accumulators, align, callable_traits, circular_buffer, compute, concept_check, contract, convert, date_time, foreach, format, function_types, functional, fusion, hana, hof, integer, lambda, local_function, logic, multi_array, numeric_conversion, odeint, parameter, phoenix, polygon, pool, preprocessor, property_tree, proto, ptr_container, range, rational, scope_exit, statechart, stl_interfaces, tokenizer, tti, tuple, type_erasure, ublas, units, variant, vmd, xpressive, yap Any such assessment is a little fuzzy by definition, so the cutoff point for what's considered maintained or not is subject to discussion. I'd like to request your feedback: * If you think some entry in this list should be removed (because you're the maintainer or you know the library is actually maintained and we got it wrong), please say so. * Conversely, tell us if you feel some Boost library not in the list should be included. * Please share any ideas on how to improve the situation so that we get more people actively maintaining more libraries. For starters, unmaintained libraries could be labeled as such on the website to attract people potentially interested in taking over (and to warn potential users about long-term adoption risks). Thank you! Best, Joaquín M López Muñoz Fiscal Sponsorship Committee

16 27

Boost 1.93.0 closed for new libraries and breaking changes
by Marshall Clow 22 Oct '26

22 Oct '26

Current status: Release branch is closed for new libraries and breaking changes to existing libraries. Still open for bug fixes and other routine changes to all libraries without release manager review. Next milestone: Boost 1.93.0 will close for major changes on Wednesday, October 28 Calendar is here: https://www.boost.org/calendar/ — The Release Managers

1 0

Re: Unmaintained libraries: a tentative list (Jason Rice)
by Jason Rice 21 Oct '26

21 Oct '26

Regarding Boost.Hana, I was given Collaborator status by the original author which I think he implied it as an honorary role. I would be willing to maintain it if he gave explicit permission to do so. I would be especially interested, if it meant allowing to move to C++26 when reflection arrives. Its interface involves functional programming concepts rooted in category theory which make it interesting and relevant, but compile times can be untenably slow for non-trivial use. On a more general note, perhaps libraries that would be otherwise removed could be allowed to be updated possibly with breaking changes to require newer versions of the standard so they have value and continued interest from collaborators or potential collaborators. Jason Rice > _______________________________________________ > Boost mailing list -- boost(a)lists.boost.org > To unsubscribe send an email to boost-leave(a)lists.boost.org > https://lists.boost.org/mailman3/lists/boost.lists.boost.org/ >

2 1