June 2024 - Boost - lists.boost.org

CppCast: Boost, The Beman Project and Beyond
by Niall Douglas 16 Jun '24

16 Jun '24

Given that some on here were expressing surprise at recent goings on, I thought in case nobody had seen it yet listening to this might be useful: https://cppcast.com/boost-the_beman_project_and_beyond/ Our own Zach Laine explains some of what's been going on behind the scenes. Niall

2 1

[Newbie] Problem receiving data with asio::serial_port
by Stefano Mora 14 Jun '24

14 Jun '24

Hi all, I’m writing a Qt application running on both Windows and MacOs platforms. The application is based on CMake project that includes boost-asio package 1.84.0 by vcpkg packager. The application must communicate with a board by USB cable = virtual serial port. I’m handling the communication with a small wrap around the asio::serial_port object, found in sample https://github.com/fedetft/serial-port/tree/master/2_with_timeout The problem is on MacOS platform; Windows version works correctly. The problem is that after some data exchanges (they are correct) the program does not receive a reply (with the TIMEOUT exception) and, after that, it receives only system_error = 89. This is a chunk of my log: ... ( I send a command and I receive the reply) [ReceiveLog] start [TP] start SendCommand 'log export' [WriteCommand] Byte[] Write - cmd: 'log export' [WaitForResponse] ms 1000 RX [34] : ' log export -> XModem receive file' [SendCommand] response: ' log export -> XModem receive file' [ReceiveLog] response1: 'XModem receive file' matching: 'XModem receive file' [ReceiveLog] ok -> XMODEM (now I send a byte to start the data from the board - 3 attempts) > Write BEGIN DATA RECEIVE .. [WaitForXModemResponse] ms 4000 [WaitForXModemResponse] TIMEOUT TO diff: 4 > Write BEGIN DATA RECEIVE .. [WaitForXModemResponse] ms 4000 [ToSerial][readCompleted] error system:89 [ToSerial][readCompleted] byteTransf 0 [WaitForXModemResponse] Boost System_error > Write BEGIN DATA RECEIVE .. [WaitForXModemResponse] ms 4000 [ToSerial][readCompleted] error system:89 [ToSerial][readCompleted] byteTransf 0 [WaitForXModemResponse] Boost System_error [TP][ReceiveLog] start XMODEM: 0 (now I sent an ACK, just to try ...) [WaitForXModemResponse] ms 1000 [ToSerial][readCompleted] error system:89 [ToSerial][readCompleted] byteTransf 0 [WaitForXModemResponse] Boost System_error [dlThread] ReceiveLog ret: 0 The read function is simply set the timeout and read 133 bytes: ... _serialPort.setTimeout( boost::posix_time::millisec(timeout)); char data[XMODEM_PACKET_LENGTH]; try { _serialPort.read( data, XMODEM_PACKET_LENGTH); ... } catch(timeout_exception){ ... } catch(boost::system::system_error){ ...} Please what error 89 means (in MACos context)? Any idea on how to exit from this situation? Thanks! Regards

2 8

CMake and Boost
by Robert Ramey 12 Jun '24

12 Jun '24

I'm trying to figure out how to switch to CMake to build and test my libraries. I have some experience with this as I've made such files on sort of an experimental/prototype basis. However it's been some time since I messed with this so a forgot a lot. Now I'm getting serious so I have some questions. I've looked over the official CMake documentation. It seems complete but some stuff is not well explained. I've spent some time studying https://github.com/boostorg/cmake at I'm beginning to see what I have to do. Never-the-less, I still have a few questions: 1) The first lines show mkdir __build cd __build cmake .. cmake --build . Does this presume that one starts from the library directory? Does this syntax work for out of source tree builds? That is, would the following work? mkdir /temp/__build cd /temp/__build cmake .. cmake --build . (very confusing usage of leading __) ahhhh - I see a problem. I'm thinking of starting from within the directory of a particular library. I'm presuming "modular boost". Is this supported? 2) BOOST_INCLUDE_LIBRARY I would like my CMakeLists.txt file to explicitly list the direct antecedent libraries (dependencies - bad word usage). So I'd expect to see something like add_directory(../filesystem) add_directory(../regex) ... So I could just cd to .../libs/serialization cmake --build and have it build the antecedent if and only if required. 3) CMAKE_INSTALL_INCLUDEDIR I would prefer not to "install" (copy) the include files but just use the ones from their original locations. 4) BUILD_TESTING I would prefer to have my library source to look like: ... libs/serialization CMakeLists.txt // includes add_directory for subdirectorys build, test, example, profile and ... // includes add_directory for antecedent libraries like add_directory(../libs/filesystem) include // directory with include files for users of serialization library build(or src) // source files for library build CMakeLists.txt // to build library test CMakeLists.txt example CMakeLists.txt profile CMakeLists.txt Maybe the add_directory would be conditioned on a CMAKE variable. So if I invoke CMake --build test_polymorphic_binary_archive It would recursively (re)build everything necessary and nothing not necessary If I didn't specify a list of targets, it would recursively (re)build everything that needs (re)building. This seems totally reasonable and natural to me - it work like make - but better. It would dovetail with efforts to "modularize" boost. Users would only have to download libraries they actually use - if that's what they want. I realized I've glossed over a couple of details, in particular recursive inclusion of header files and likely some other stuff. I've also experimented with CMake enough to have a feeling that what I want to do is not possible. And this is the short version of what I'd like to do. Any thoughts on this?

4 5

[container] Draft development plan for Boost.Container
by Ion Gaztañaga 11 Jun '24

11 Jun '24

Hi to all, Boost.Container was accepted in 2011 when container classes were spun off from Boost.Interprocess to form a new library. At that time it offered new features like emulated move semantics for "classic" STL-like containers and some new classes like flat_map and stable_vector. During these years more containers/main features were added: - The scoped allocator Model was added in Boost 1.50 (2012) - static_vector was added in Boost 1.54 (2013) - small_vector was added in Boost 1.58 (2015) - Polymorphic Memory Resources were added in Boost 1.60 (2015) - devector was added in Boost 1.75 (2020) Lately I've been collecting some ideas about how to improve the library and keep it interesting/useful both for other Boost libraries and Boost users. I've somewhat formalized a bit some of those ideas and I'd like to share them here in order to obtain feedback from the developers. Here's the document: https://docs.google.com/document/d/1VdMTheFUczC946dP3VKPseOoUq3sVKjU73oMvK8… Note that this document does not outline a schedule for all those ideas, it's an initial description of possible developments that should be evaluated/prioritized, as developing a substantial subset of the proposals would be a huge task. Thanks in advance for your attention and comments. Best, Ion

9 27

b2 headers failing on windows?
by John Maddock 10 Jun '24

10 Jun '24

In latest CI runs, see: https://github.com/boostorg/math/actions/runs/9427754774/job/25973291514?pr… Any ideas? Thanks! John.

6 8

Boost Downloads
by Sam Darwin 07 Jun '24

07 Jun '24

Summary: Boost downloads will be hosted at a new location, with the URL format https://archives.boost.io/release/1.85.0/source/boost_1_85_0.tar.gz Details: During the last couple of years the Boost downloads have been hosted on JFrog Artifactory, and before that on Bintray. JFrog has been very generous in providing the bandwidth and hosting services. Recently, bandwidth increased into the 150TB/month range. JFrog requested the download be moved behind a CDN, or migrated to another service entirely. At the C++ Alliance, we have set up load balanced AWS EC2 instances, and configured a Fastly CDN, on the domain archives.boost.io. Fastly offered a discounted bulk rate based on the large amount of traffic. The boost.org website now shows links to the new CDN. If you are using any automated scripts that point to jfrog.io, the existing JFrog links will continue to function until December 2024. - Sam Darwin

4 5

Docs in release tarballs
by Ruben Perez 07 Jun '24

07 Jun '24

Hi all, For someone that hasn't been that long involved in Boost, it feels strange to distribute the documentation in the same release tarball as the code, specially considering the usual complains about "Boost size". I'm sure there are reasons behind it, and I'd like to learn about them. This is a candid question, out of being a newbie here. Thanks, Ruben

8 14

Re: [boost] Boost Downloads
by Tim Haines 06 Jun '24

06 Jun '24

> Summary: > > Boost downloads will be hosted at a new location, with the URL format > https://archives.boost.io/release/1.85.0/source/boost_1_85_0.tar.gz > > Details: > > During the last couple of years the Boost downloads have been hosted on > JFrog Artifactory, and before that on Bintray. JFrog has been very generous > in providing the bandwidth and hosting services. > > Recently, bandwidth increased into the 150TB/month range. JFrog requested > the download be moved behind a CDN, or migrated to another service entirely. > > At the C++ Alliance, we have set up load balanced AWS EC2 instances, and > configured a Fastly CDN, on the domain archives.boost.io. Fastly offered a > discounted bulk rate based on the large amount of traffic. > > The boost.org website now shows links to the new CDN. > > If you are using any automated scripts that point to jfrog.io, the existing > JFrog links will continue to function until December 2024. > > > - Sam Darwin I maintain the Boost spack package[1]. The versions currently available there go back to 1.39.0. However, the versions available on archives.boost.io only go back to 1.63.0. Are the older versions going to be available there or will they remain on the sourceforge hosting? Thanks. - Tim [1] https://github.com/spack/spack/blob/develop/var/spack/repos/builtin/package… P.S. I am terribly sorry if this messes up the thread. I get my messages in digest form and could not for the life of me figure out how to reply to your email directly to the list.

2 1

[JSON] Detect duplicate keys
by abbott＠dima.unige.it 04 Jun '24

04 Jun '24

Hi, Feature request: I would like the JSON parse to inform me that duplicate keys were encountered. I believe they are currently silently handled (with a last-one-wins strategy). Thanks, John Abbott

4 3

Mailman3
by Sam Darwin 04 Jun '24

04 Jun '24

Hi, Since the current Boost webserver and mailing list are quite out-of-date, it makes sense to consider the upgrade path. Mailman3 is a complete rewrite of the mailing list software. It's designed with multiple components: hyperkitty archives, postorius, and mailman-core. Dev instances have been installed. If you have any curiosity about testing mailman3 using your local email client, or sending feedback about the mailman3 experience, you can subscribe at https://lists.preview.boost.org. Send/receive emails. If you notice problems using mailman3, open an Issue at the Github repository cppalliance/boost-mailman. Now is the best time to discover issues, rather than after deployment. Lists at https://lists.preview.boost.org won't be maintained permanently. Any emails or accounts there will eventually be deleted. It's only a staging type environment. Working with the web developers we are planning to modify the appearance of the front-end UI, adding css styles, links to the boost website, and so on. Keep an eye on the test lists for updates. In the future when "going live", the idea will be to schedule (extended) planned downtime, and have a new server fully replace "lists.boost.org". Which specific boost mailing lists should be kept? It would be easier to avoid migrating all lists, if some of them are not active. Best Regards, Sam Darwin C++ Alliance

1 0