Re: increasing CXX_STANDARD for kdevelop
Milian Wolff <[email protected]> Wed, 14 Oct 2020 13:59:17 +0200
| Newsgroups | gmane.comp.kde.devel.kdevelop |
|---|---|
| Message-ID | <3910424.ZkMNLspcJH@milian-workstation> |
On Mittwoch, 14. Oktober 2020 13:13:39 CEST Kevin Funk wrote: > On Wednesday, 14 October 2020 12:41:20 CEST Milian Wolff wrote: > > On Mittwoch, 14. Oktober 2020 12:36:58 CEST Kevin Funk wrote: > > > On Wednesday, 14 October 2020 11:28:17 CEST Milian Wolff wrote: > > > > On Mittwoch, 14. Oktober 2020 06:57:15 CEST Igor Kushnir wrote: > > > > > On 2020-10-13 23:36, Milian Wolff wrote: > > > > > > Hey all, > > > > > > > > > > > > I'd like to increase the CXX_STANDARD for kdevelop, seems like we > > > > > > are > > > > > > still > > > > > > compiling with `-std=gnu++11`. Any objections to raise this to > > > > > > `-std=c++14`? > > > > > > > > > > > > And, actually, while at it - I would personally even prefer to > > > > > > jump > > > > > > straight ahead to `-std=c++17`. Any objections to that? > > > > > > > > > > > > I rarely have time for KDevelop, but whenever I do take the time > > > > > > I'm > > > > > > put > > > > > > off by the (imo) archaic C++ support we have :) C++17 is a lot of > > > > > > fun, > > > > > > esp. with `if constexpr` and some other utilites. > > > > > > > > > > > > Cheers > > > > > > > > > > Hi Milian, > > > > > > > > > > I'd love that change too. Never had a chance to practice C++17 yet. > > > > > I > > > > > can see two downsides though: > > > > > > > > > > 1. If the KDE Frameworks and the majority of KDE applications are > > > > > mostly > > > > > tested with C++11 (not sure about that), then there might be > > > > > compatibility issues and bugs exposed by switching KDevelop to a > > > > > different C++ standard. > > > > > > > > From personal experience, I would say this is a non-issue. The ABI is > > > > independent of the compiler version being used e.g. > > > > > > > > > 2. Some of the compilers KDevelop aims to support may not support > > > > > C++17 > > > > > perfectly. We'll need to specify which versions of which compilers > > > > > should be able to build KDevelop, then use standard feature support > > > > > tables like this - > > > > > https://en.cppreference.com/w/cpp/compiler_support#cpp17 - whenever > > > > > we > > > > > try out a new feature. > > > > > > > > Right, does anyone know a table of compiler versions as shipped with > > > > the > > > > major distributions? I seem to remember I once saw something like > > > > that. > > > > > > Heya, > > > > > > Repology.com maybe? > > > > > > E.g.: > > > https://repology.org/project/clang/versions > > > > That looks good - thanks! > > > > Based on the feedback I got so far it seems like we should be able to > > update to C++17. Based on the question below I'll then make that step in > > the next days. > > > > > Also check whether the compilers on the CentOS image we're using for > > > AppImage building has proper C++17 support. Though I'm sure there's also > > > a > > > more recent devtoolset one can use on CentOS to overcome these issues > > > (haven't checked it). > > > > Where do I find documentation on the AppImage build setup for KDevelop? > > All here: > https://invent.kde.org/kdevelop/kdevelop/-/tree/master/appimage > > Though FLHerne was in the process of updating to CentOS 7 at least (we're @ > CentOS 6.10 right now). Not sure that was successful, Francis? Ah, thanks! Is this integrated into craft/CI anywhere, or can I hack on this at will? I would then automate a few steps, based on the experience I have with AppImage generation in docker in heaptrack and hotspot. Cheers -- Milian Wolff [email protected] http://milianw.de
signature.asc
(application/pgp-signature, 833 B)
-----BEGIN PGP SIGNATURE----- iQIzBAABCAAdFiEEezawi1aUvUGg3A1+8zYW/HGdOX8FAl+G6BUACgkQ8zYW/HGd OX/7phAAslL6GAkXAP4bAF+dlGURYNKlrFQEUfgTXNFgFaOqu7WCSGKWRu0PkqPj Di1ylyrI/zPAm7W5Pm2Z2Mx84mVCB9jYP4SZ5hu4uXAyISH1ejCGwtCIMNrEm6Om 8LSYaaZsd7ek9HiSvbp9YnuRADERIpYlRuMcn8+I+G4FOqZzF4Auj2yzzOBfhnU+ Q0A/8iUH7ENUxa1RRBTcUR956bAQAXLnfhI8wcYhc0bnBzUqfo6wPSS0S0JVRSjS HiyKA7+5MRrO0sw5J42MU1gVlZ/BbkTe82F9HpGbDtmL7uHeU5c/rozH9g7e8EdL jNs15Mg+W3/w7fl7HeoFbZdm+9oAl6UADHSQKzXmwaUctkqLP9gs0W+rl0CSOCHq Wo4GEnRypW3JoV+4J4RjE4ko/2iUyaGxV0WzBPEK0YveKnSZyMByY3E1ZvKMR7Fc U05Nict/t+nmWWFrhsaKy4OF/PQSDNDxcIh3XsQkWAaOd7FCSHX0ROm9HdoG/zl4 yf9e6xXqifClNrhAqV4zHNL9Y3womxYXTzz4amfd3mX0KdZfBp++ij5LHE49kUQF 2REolkByQ4tbTBW4csDFMtRdStlexqSqk26z8lOmcsqidRfY7IVcSumVTTYbQJX1 icesJXUsA6t74Hstg5iIQ96RkpH+poMY6MyTPDGTCO0OVzYPxE8= =O3hE -----END PGP SIGNATURE-----