C++11 support

Developers
2011-10-19
2013-05-13
  • Maxim Yanchenko

    Maxim Yanchenko - 2011-10-19

    Any plans/ETA to incorporate the changes in C++11 Standard Library?

    Compilers for C++11 are more and more wide-spread, and there are libraries that already assume that underlying STL provides C++11 features when C++11 compatibility is turned on in a compiler (e.g. recent versions of TBB expect new std::exception class with the exception propagation support).

    So STLPort without C++11 support becomes more and more problematic.

     
  • Petr Ovtchenkov

    Petr Ovtchenkov - 2011-10-19

    In mainstream.

     
  • Maxim Yanchenko

    Maxim Yanchenko - 2011-10-20

    Great news!
    When the release is planned?

     
  • Petr Ovtchenkov

    Petr Ovtchenkov - 2011-10-20

    'Release' in time when VCS available is a bit fiction.

    Things, that I used everyday, are fine. For me. Support for compilers with absent C++11 removed, or be removing ASAP. Are you Windows player? Then bad news for you: I don't write for Windows, so even things that I use everyday, may be (indeed should be) broken on VS. Conclusion: if you will wait for "good, stable release", and only then (may be) see on code, then you will wait a very long time.

     
  • Maxim Yanchenko

    Maxim Yanchenko - 2011-10-20

    Well, VCSes exist everywhere, notion of release is orthogonal to VCS - it's a (kind of) guarantee of stability and is a reference point. See Boost for example.

    A 'release' means that many people/teams downloaded something which is tagged as release, tested it, reported bugs, and after a couple of bug-fix releases we can be sure it's stable and relatively bug-free because this particular version was widely tested (on a known set of platforms and compilers) and nobody reported any bugs for a couple of months already.
    Who can tell the same about the top of VCS which changes every day? Not even you, I'm afraid.

    It's a matter of third-party libraries management as well - people usually can't put in production code something from the current top of VCS. Just because the top of VCS can have some experimental non-tested code that was committed a few minutes ago. In many companies they are just not allowed to use in production code anything that is not recognized as stable, at least by its developer.

    At least, if you had had a 'stable' branch, it would be a justification to take a snapshot from it, but I don't see this either, I see just one 'master' head.

     
  • Maxim Yanchenko

    Maxim Yanchenko - 2011-10-21

    Thanks for the links. Again, I see only 'master' and old numbered tags.
    Does 'master' contain only stuff that is tested, or there can be experimental commits?
    Also, what is the compiler(s) and platform(s) you use for development and testing?

    > Thanks for useful response. You opinion is very important for us.
    I really hope it will be useful, as I see that lack of releases hurts STLPort reception and widespread use.
    E.g. from http://libcxx.llvm.org/ (LLVM is quite popular now): "….STLPort…lack C++'0x support….projects are apparently abandoned: STLport 5.2.1 was released in Oct'08…"
    People just don't know that some work is going on in VCS, they just see that there were no releases for 3 years and treat STLPort as abandoned.

     
  • Petr Ovtchenkov

    Petr Ovtchenkov - 2011-10-21

    > Again, I see only 'master' and old numbered tags.

    Do you need a git lessons?

    > I really hope it will be useful,

    It was a trite comments. Nothing new. Good patch series are much more useful (thanks nikolaynnov@, solganik@, wojciech.meyer@, … ).

    > E.g. from http://libcxx.llvm.org/

    If LLVM will be in scope of my interests, then STLport will support it. All good patches was considered and accepted.But  I don't see attempts for contribution from LLVM users yet.

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks