Date: Fri, 7 Jul 2023 17:56:22 +0200 From: Rainer Hurling <rhurlin@gwdg.de> To: Nuno Teixeira <eduardo@freebsd.org> Cc: Raphael Kubo da Costa <rakuco@freebsd.org>, <ports-committers@freebsd.org>, <dev-commits-ports-all@freebsd.org>, <dev-commits-ports-main@freebsd.org> Subject: Re: git: b56d0bc41af7 - main - devel/libcutl: Explain why USE_CXXSTD=c++11 is used Message-ID: <34ea7315-6e9b-ebe4-6be8-5e0518cf969c@gwdg.de> In-Reply-To: <CAFDf7UJ90-hVX3WePtD--tHb0HEeUrq6Y0mRezZ8kYYoDQL69g@mail.gmail.com> References: <202307050850.3658oPOA049562@gitrepo.freebsd.org> <CAFDf7UJRc4bd3XVq%2Bb7Za2cLN7h_hkN1mpPTGQiSajV1p9A=0g@mail.gmail.com> <ebcdf01a-2b73-39c3-b6e4-25659e53b1ca@gwdg.de> <e56d377d0985b86be2291b64bb3c686742283dde.camel@FreeBSD.org> <6cf6a5eb00dfac8358dfcb1e1d5b69f293c3e85a.camel@FreeBSD.org> <4701f862-7a0b-afab-48fd-a62b6f5ec308@gwdg.de> <CAFDf7UKJh%2BscZ-zkmwWUNeQbVB%2B42TQSN8-m102iChi8wg0WoA@mail.gmail.com> <4cc5665f-a4ef-a4ef-226a-a9a59b29f3a0@gwdg.de> <CAFDf7ULoQfRo4S45Fpwm9Ve4h0E%2Bd82z5SYrFopmcKFLOShrdw@mail.gmail.com> <45470f65-48c3-9a18-b7bd-26ce17160afd@gwdg.de> <CAFDf7UJ90-hVX3WePtD--tHb0HEeUrq6Y0mRezZ8kYYoDQL69g@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Am 07.07.23 um 13:29 schrieb Nuno Teixeira: > PORTREVISION=11: USE_CXXSTD=gnu++14 > http://freebsdrpi4.hopto.org/data/13Saarch64-main/2023-07-07_12h12m14s/logs/eboard-1.1.3_11.log <http://freebsdrpi4.hopto.org/data/13Saarch64-main/2023-07-07_12h12m14s/logs/eboard-1.1.3_11.log>; > > PORTREVISION=10: USE_CXXSTD=c++14 > http://freebsdrpi4.hopto.org/data/13Saarch64-main/2023-07-07_09h44m48s/logs/eboard-1.1.3_10.log <http://freebsdrpi4.hopto.org/data/13Saarch64-main/2023-07-07_09h44m48s/logs/eboard-1.1.3_10.log>; > > Both revs 10 and 11 runs OK. yes, both seems fine. Even the hashes of the packages are the same. So no dump of portrevision necessary ;) Which version, gnu++14 or c++14, should we use here in terms of being as universal as possible? Probably gnu++14, since clang treats both the same ... > I've take a look at logs and it seems that glib2 deprecations could be > the reason for c++17 build failures. > Could you take a look at: > https://cgit.freebsd.org/ports/commit/?id=e95c4bea425b3e3c49b63e0cf4eff24af40c574b <https://cgit.freebsd.org/ports/commit/?id=e95c4bea425b3e3c49b63e0cf4eff24af40c574b>; > it also includes link to mailing full discussion. > > This is what we should do when possible, fix the code as said in > previous messages. Yes, it is likely that glib2 is the culprit here. Two warnings with direct glib reference stand out to me: - 'GTypeDebugFlags' is deprecated - GTimeVal' is deprecated: Use 'GDateTime' instead My C++ knowledge and my knowledge about glib are unfortunately not enough to provide a patch :( Many thanks for the poudriere runs under arm. Very appreciated! > > Rainer Hurling <rhurlin@gwdg.de <mailto:rhurlin@gwdg.de>> escreveu no > dia sexta, 7/07/2023 à(s) 10:53: > > Am 07.07.23 um 10:59 schrieb Nuno Teixeira: > > Hello! > > > > I've bumped PORTREVISION to 10 so we can identify better each > experiment > > build: > > > > PORTREVISION=10: USE_CXXSTD=c++14 > > build status: > > > http://freebsdrpi4.hopto.org/build.html?mastername=13Saarch64-main&build=2023-07-07_09h44m48s <http://freebsdrpi4.hopto.org/build.html?mastername=13Saarch64-main&build=2023-07-07_09h44m48s>; <http://freebsdrpi4.hopto.org/build.html?mastername=13Saarch64-main&build=2023-07-07_09h44m48s <http://freebsdrpi4.hopto.org/build.html?mastername=13Saarch64-main&build=2023-07-07_09h44m48s>>; > > > > and when built I will do a run test. > > > > Oops, yes! > > > Shoud I test gnu++14 too? > > I accidentally provided a patch with c++14 that I had used for another > test of my own. Meant was gnu++14, so please feel free to test that one > too. Thanks! > > > > > > > > > > > Rainer Hurling <rhurlin@gwdg.de <mailto:rhurlin@gwdg.de> > <mailto:rhurlin@gwdg.de <mailto:rhurlin@gwdg.de>>> escreveu no > > dia sexta, 7/07/2023 à(s) 06:42: > > > > Hi Nuno, > > > > Am 06.07.23 um 23:51 schrieb Nuno Teixeira: > > > Hello Rainer, > > > > > > Still, one more question: is it safe to remove > > USES=compiler:c++11-lang > > > for architectures other than amd64 and i386 and just use > > > USE_STDCXX=gnu++14. I have no way to test for arm, etc. > > > > > > > > > I've asked the same for compiler:c11 on mailing today. > > > My doubt is: is it neccessary using c11, c++11, etc when > available > > > compilers support it or have capabilities for it? > > > > > > I can do poudriere testport and run test on a rpi4 running 13 > > stable, > > > just send me diff. > > > > Thanks for the offer. Here is a diff against games/eboard > with the to > > changes: > > > > diff --git a/games/eboard/Makefile b/games/eboard/Makefile > > index 3f78f9e91874..ce399415cba8 100644 > > --- a/games/eboard/Makefile > > +++ b/games/eboard/Makefile > > @@ -20,10 +20,10 @@ LIB_DEPENDS= > > libfontconfig.so:x11-fonts/fontconfig \ > > libfreetype.so:print/freetype2 \ > > libpng.so:graphics/png > > > > -USES= compiler:c++11-lang gnome gstreamer perl5 > pkgconfig > > shebangfix tar:bzip2 > > +USES= gnome gstreamer perl5 pkgconfig shebangfix > tar:bzip2 > > # eboard 1.1.13 is not compatible with C++17, which is the > > default since > > # clang 16. This means ports depending on it also need to > use at > > most > > C++14. > > -USE_CXXSTD= c++11 > > +USE_CXXSTD= c++14 > > USE_GITHUB= yes > > GH_ACCOUNT= fbergo > > USE_GNOME= cairo gdkpixbuf2 gtk20 > > > > > > It would be very nice if you could also look for arm, if the > package is > > still binary compatible compared to the unpatched version :) > > > > Thanks a lot ! > > > > Best wishes, > > Rainer > > > > > > > > Cheers, > > > > > > -- > > > Nuno Teixeira > > > FreeBSD Committer (ports) > > > > > > > > -- > > Nuno Teixeira > > FreeBSD Committer (ports) > > > > -- > Nuno Teixeira > FreeBSD Committer (ports)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?34ea7315-6e9b-ebe4-6be8-5e0518cf969c>