<div dir="ltr">On Wed, Dec 23, 2015 at 12:10 PM, Oded Gabbay <<a href="mailto:oded.gabbay@gmail.com" target="_blank">oded.gabbay@gmail.com</a>> wrote: <div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Wed, Dec 23, 2015 at 1:05 PM, Andrea Canciani <<a href="mailto:ranma42@gmail.com">ranma42@gmail.com</a>> wrote: > On Tue, Dec 22, 2015 at 2:19 PM, Oded Gabbay <<a href="mailto:oded.gabbay@gmail.com">oded.gabbay@gmail.com</a>> wrote: >> >> On Thu, Jun 4, 2015 at 12:02 PM, Andrea Canciani <<a href="mailto:ranma42@gmail.com">ranma42@gmail.com</a>> >> wrote: >> > The patch does not regress the mingw-based build (basically that used in >> > <a href="http://cairographics.org/end_to_end_build_for_win32/" rel="noreferrer" target="_blank">http://cairographics.org/end_to_end_build_for_win32/</a> ) >> > I did not manage to get the build working on the cmd.exe shell. Do you >> > have >> > any instructions (or possibly indications about what is the correct >> > environment) for that? >> > >> > AFAICT the change still relies on a Unix-like make command. >> > A bigger (but possibly even more interesting) change would be an >> > NMAKE-based >> > build, so that the command-line tools available in Visual Studio are >> > sufficient for building the whole project. >> > >> > Andrea >> > >> >> Hi Andrea, >> >> I know it has been almost 6 months since you last looked at this >> patch, but I would like to ask if you can give it an r-b and then I >> will push it to master. > > > Actually, this is a good time for me to work on pixman, as I have some more > free time :) Thanks Andrea, much appreciated. <div><div class="h5"> > >> >> If not, what needs to be done to give it an r-b ? > > > The main thing blocking me from an r-b tag originally was that I had been > unable to get it working. > I set up an environment to test it and found that it mostly behaves as > intended, with the main glitch being that the "clean" target can spit out > undesired error logs. > > I pushed a branch which is based on Simon's patch here: > <a href="http://cgit.freedesktop.org/~ranma42/pixman/log/?h=win32-build" rel="noreferrer" target="_blank">http://cgit.freedesktop.org/~ranma42/pixman/log/?h=win32-build</a> > > The commits include: > - cleanup of the win32 build system (mostly unrelated to the original > patch) > ea8a070 build: Remove use of BUILT_SOURCES from Makefile.win32 > 8989759 build: Avoid phony `pixman` target in test/Makefile.win32 > - shell-generic compatibility > db72f26 build: Do not use `mkdir -p` on Windows > - cmd.exe compatibility > 6223be5 build: Use `del` instead of `rm` on `cmd.exe` shells > > The last commit is based on the original patch, with some minor differences > to avoid path separator handling (instead I changed the current directory) > and to avoid error messages from `del` (when it finds no file to delete, it > logs a warning that apparently cannot be silenced by any of the flags). > > While doing this I discovered that on Win32 fence-image-self-test and > cover-test are currently failing (regardless of these patches and apparently > also with any combination of MMX, SSE2, SSSE3). > I will try to investigate ASAP. </div></div></blockquote><div> </div><div>The tests are reported as failing because the 77 return value is recognized as SKIP by the automake-generated test handling code, but it is assumed to be a generic error by the Win32 build system. </div><div>I will try to provide a patch for this soon. </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5"> </div></div>Thanks. > > I believe that the first two commits should be "obviously" ok, while the > other two are not as trivial. > The whole patchset has been tested on Win32, both in an MSYS environment and > on cmd.exe (in both cases, using GNU make). > > Andrea > > PS: it looks like pixman has adopted Reviewed-by, Signed-off-by, etc. > I assume that the first three patches should be marked signed off by me, > while I am unsure about the tagging for the last patch (I guess it should be > credited to Simon). Yes, you are correct. Could you please send the patches as a patch series to the pixman ML, so we could do a quick review process and merge it ? </blockquote><div> </div><div>Thank oyu for taking the time to review/ack the patches. If nobody says otherwise, I will push the commits tonight. </div><div> </div><div>Andrea </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> If you want to credit the last patch to Simon, just make sure he is the author of the patch (you can change the author git commit --amend --author). Then, send it as part of your series. Thanks, Oded <div class="HOEnZb"><div class="h5">> > >> Unfortunately, I don't have a Windows machine to check this. >> >> Thanks, >> >> Oded >> >> > >> > On Mon, Jun 1, 2015 at 8:44 AM, Andrea Canciani <<a href="mailto:ranma42@gmail.com">ranma42@gmail.com</a>> >> > wrote: >> >> >> >> The ping was good and the changes make sense. Currently I am replying >> >> from >> >> the phone and I will not have access to my computer until the 3rd, but >> >> I am >> >> flagging this mail so that I will test and review the patch asap. >> >> >> >> Sorry for the delay >> >> Andrea >> >> >> >> On Jun 1, 2015 8:09 AM, "Siarhei Siamashka" >> >> <<a href="mailto:siarhei.siamashka@gmail.com">siarhei.siamashka@gmail.com</a>> >> >> wrote: >> >>> >> >>> On Wed, 22 Apr 2015 04:01:31 +0200 >> >>> Simon Richter <<a href="mailto:Simon.Richter@hogyros.de">Simon.Richter@hogyros.de</a>> wrote: >> >>> >> >>> > This finds out whether the standard shell for make is cmd.exe, and >> >>> > adjusts >> >>> > the build process accordingly. >> >>> > --- >> >>> > Makefile.win32.common | 14 ++++++++++++-- >> >>> > 1 file changed, 12 insertions(+), 2 deletions(-) >> >>> >> >>> Thanks for the patch. >> >>> >> >>> But it is a bit difficult for non-Windows folks to see if >> >>> it makes any sense or not. So it would be best if some other >> >>> Windows user could review and confirm that it works. >> >>> >> >>> I have added Andrea Canciani (the author of the win32 makefiles) >> >>> to CC. >> >>> >> >>> > diff --git a/Makefile.win32.common b/Makefile.win32.common >> >>> > index 777f94c..335886f 100644 >> >>> > --- a/Makefile.win32.common >> >>> > +++ b/Makefile.win32.common >> >>> > @@ -1,5 +1,15 @@ >> >>> > LIBRARY = pixman-1 >> >>> > >> >>> > +ifeq ($(shell echo ""),) >> >>> > +# POSIX style shell >> >>> > +mkdir_p = mkdir -p $1 >> >>> > +rm = $(RM) $1 >> >>> > +else >> >>> > +# DOS/Windows style shell >> >>> > +mkdir_p = if not exist $(subst /,\,$1) md $(subst /,\,$1) >> >>> > +rm = del $(subst /,\,$1) >> >>> > +endif >> >>> > + >> >>> > CC = cl >> >>> > LD = link >> >>> > AR = lib >> >>> > @@ -47,10 +57,10 @@ endif >> >>> > >> >>> > >> >>> > $(CFG_VAR)/%.obj: %.c $(libpixman_headers) >> >>> > - @mkdir -p $(CFG_VAR) >> >>> > + @$(call mkdir_p,$(@D)) >> >>> > @$(CC) -c $(PIXMAN_CFLAGS) -Fo"$@" $< >> >>> > >> >>> > clean: inform >> >>> > - @$(RM) $(CFG_VAR)/*.{exe,ilk,lib,obj,pdb} $(BUILT_SOURCES) || >> >>> > exit 0 >> >>> > + @-$(call rm,$(CFG_VAR)/*.exe $(CFG_VAR)/*.ilk $(CFG_VAR)/*.lib >> >>> > $(CFG_VAR)/*.obj $(CFG_VAR)/*.pdb} $(BUILT_SOURCES)) >> >>> > >> >>> > .PHONY: inform clean >> >>> >> >>> >> >>> >> >>> -- >> >>> Best regards, >> >>> Siarhei Siamashka >> > >> > >> > >> > _______________________________________________ >> > Pixman mailing list >> > <a href="mailto:Pixman@lists.freedesktop.org">Pixman@lists.freedesktop.org</a> >> > <a href="http://lists.freedesktop.org/mailman/listinfo/pixman" rel="noreferrer" target="_blank">http://lists.freedesktop.org/mailman/listinfo/pixman</a> >> > > > </div></div></blockquote></div> </div></div>