Re: [Mingw-w64-public] [PATCH 2/2] winstorecompat: Provide RtlAddFunctionTable on 64bits platforms only
On Mon, 6 Aug 2018, Liu Hao wrote: 在 2018-08-06 01:38, Martin Storsjö 写道: On Sun, 5 Aug 2018, NightStrike wrote: On Wed, Jul 11, 2018, 1:20 PM Hugo Beauzée-Luyssen wrote: On Wed, Jul 11, 2018, at 7:12 PM, Martin Storsjö wrote: I originally wanted to do so but disliked the "iso C forbids empty translation unit" warning, however if that's fine by you then I'll remove the configure/makefile part! Builds should be warning free, from all tools and all configurations. This is a quality regression I'm seeing across the entire project. I personally think the warning about empty translation units shouldn't be enabled, as it is included in `-Wpedantic` only, which shouldn't be turned on unless by some VERY pedantic people. This warning can be silenced by putting an unused typedef in the TU: ``` typedef __attribute__((__unused__)) int this_translation_unit_is_not_empty; ``` Yes, and in this particular case, I worked around it by moving the ifdef guards after the system headers, so the translation unit never really is empty. The other warning I was referring to, that I would have deemed acceptable, only seems to be printed by macOS ar/ranlib, e.g. like this: $ cat test.c typedef int foo; $ clang test.c -c -o test.o $ ar rcs libtest.a test.o /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/ranlib: file: libtest.a(test.o) has no symbols That isn't even prefixed with "warning", just an extra verbose printout, and only in the macOS tools, not with GNU binutils. // Martin -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
Re: [Mingw-w64-public] [PATCH 2/2] winstorecompat: Provide RtlAddFunctionTable on 64bits platforms only
在 2018-08-06 01:38, Martin Storsjö 写道: > On Sun, 5 Aug 2018, NightStrike wrote: > >> On Wed, Jul 11, 2018, 1:20 PM Hugo Beauzée-Luyssen >> wrote: >> >>> On Wed, Jul 11, 2018, at 7:12 PM, Martin Storsjö wrote: >>> I originally wanted to do so but disliked the "iso C forbids empty >>> translation unit" warning, however if that's fine by you then I'll >>> remove >>> the configure/makefile part! >>> >> >> Builds should be warning free, from all tools and all configurations. >> This >> is a quality regression I'm seeing across the entire project. > > I personally think the warning about empty translation units shouldn't be enabled, as it is included in `-Wpedantic` only, which shouldn't be turned on unless by some VERY pedantic people. This warning can be silenced by putting an unused typedef in the TU: ``` typedef __attribute__((__unused__)) int this_translation_unit_is_not_empty; ``` -- Best regards, LH_Mouse -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
Re: [Mingw-w64-public] [PATCH 2/2] winstorecompat: Provide RtlAddFunctionTable on 64bits platforms only
On Sun, 5 Aug 2018, NightStrike wrote: On Wed, Jul 11, 2018, 1:20 PM Hugo Beauzée-Luyssen wrote: On Wed, Jul 11, 2018, at 7:12 PM, Martin Storsjö wrote: > On Wed, 11 Jul 2018, Hugo Beauzée-Luyssen wrote: > > > RtlAddFunctionTable is only available on those. Moreover, so are the > > (P)RUNTIME_FUNCTION types > > No, it's available for armv7 as well (I just checked the msvc import > libs). So the correct logic probably is "HAVE_NON_I386" or something like > that. (We have a corresponding macro F_NON_I386 for .def files.) > Ok, I'll send an updated version :) > Also, to keep things simple, you could skip the build system part and just > use an ifdef in the source file; ar usually gives a small warning for > object files without symbols, but it shouldn't hurt much imo. But I won't > object to that part if you really want to have the condition in the > makefile. > I originally wanted to do so but disliked the "iso C forbids empty translation unit" warning, however if that's fine by you then I'll remove the configure/makefile part! Builds should be warning free, from all tools and all configurations. This is a quality regression I'm seeing across the entire project. The version actually committed was warning free - in particular, I changed Hugo's patch to get rid of that warning he mentioned (it was trivial to do so), and ar didn't produce the warning I thought it might. I do generally try to make sure that all my changes add no warnings. // Martin -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
Re: [Mingw-w64-public] [PATCH 2/2] winstorecompat: Provide RtlAddFunctionTable on 64bits platforms only
On Wed, Jul 11, 2018, at 7:12 PM, Martin Storsjö wrote: > On Wed, 11 Jul 2018, Hugo Beauzée-Luyssen wrote: > > > RtlAddFunctionTable is only available on those. Moreover, so are the > > (P)RUNTIME_FUNCTION types > > No, it's available for armv7 as well (I just checked the msvc import > libs). So the correct logic probably is "HAVE_NON_I386" or something like > that. (We have a corresponding macro F_NON_I386 for .def files.) > Ok, I'll send an updated version :) > Also, to keep things simple, you could skip the build system part and just > use an ifdef in the source file; ar usually gives a small warning for > object files without symbols, but it shouldn't hurt much imo. But I won't > object to that part if you really want to have the condition in the > makefile. > I originally wanted to do so but disliked the "iso C forbids empty translation unit" warning, however if that's fine by you then I'll remove the configure/makefile part! Thanks, -- Hugo Beauzée-Luyssen h...@beauzee.fr -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
Re: [Mingw-w64-public] [PATCH 2/2] winstorecompat: Provide RtlAddFunctionTable on 64bits platforms only
On Wed, 11 Jul 2018, Hugo Beauzée-Luyssen wrote: RtlAddFunctionTable is only available on those. Moreover, so are the (P)RUNTIME_FUNCTION types No, it's available for armv7 as well (I just checked the msvc import libs). So the correct logic probably is "HAVE_NON_I386" or something like that. (We have a corresponding macro F_NON_I386 for .def files.) Also, to keep things simple, you could skip the build system part and just use an ifdef in the source file; ar usually gives a small warning for object files without symbols, but it shouldn't hurt much imo. But I won't object to that part if you really want to have the condition in the makefile. // Martin -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
Re: [Mingw-w64-public] [PATCH 2/2] winstorecompat: Provide RtlAddFunctionTable on 64bits platforms only
On 07/11/2018 12:06 PM, Hugo Beauzée-Luyssen wrote: > RtlAddFunctionTable is only available on those. Moreover, so are the > (P)RUNTIME_FUNCTION types Patches OK. signature.asc Description: OpenPGP digital signature -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot___ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
