A Dimarts, 18 d'octubre de 2011, Peter Breitenlohner vàreu escriure: > On Mon, 17 Oct 2011, Albert Astals Cid wrote: > > Well, that is your choice, you can still link against a static poppler > > build instead of embedding code the code in your repo. > > Hi Albert, > > yes we could. However, we prefer the repo to be as self consistent as > possible. > > > Anyway: > > > > * patch-01-PDF_PARSER_ONLY > > Applied, does not make much sense having PDF_PARSER_ONLY in poppler > > Thanks > > > * patch-10-exclude-body-of-Page_getDefaultCTM > > Having PDF_PARSER_ONLY in poppler makes not much sense, so rejected > > The idea was to apply either this or the previous patch but not both. > > > * patch-04-mingw > > Having PDF_PARSER_ONLY in poppler makes not much sense, so rejected. > > Maybe you can actually fix the problem instead of workarounding it? > > Indeed. All we need to do is adding '-lgdi32'. That library is useless for > our purposes but so what. > > > * patch-03-POPPLER_VERSION > > Is this for when you use the embedded copy or for when you link against > > the external lib? > > Actually for both. For xpdf there was xpdfVersion defined in xpdf/config.h, > but you have decided to make the poppler version info available only via the > c++ wrapper library and corresponding headers.
To be honest we do not want/like/encourage people using poppler core internals so adding this would be a "breach" of those ideals. OTOH it is a small-ish change and I remember some other people wanted it. Anyone else has an opinion on this? Carlos? Pino? Adrian? > > Having POPPLER_VERSION (possibly under a different name) available in > poppler-config.h would be helpful when not using libpoppler-cpp. > > > * patch-02-warnings > > Which compiler/flags give you the need for this? > > The compiler is gcc-3.4.6. When compiling poppler, your configure yields > CXXFLAGS = -Wall -Wno-write-strings -Woverloaded-virtual > -Wnon-virtual-dtor \ -Wcast-align -fno-exceptions -fno-check-new > -fno-common -O2 -ansi $(PTHREAD_CFLAGS) in the various Makefiles, resulting > in these warnings: These seem to be the same flags we use and I'm not getting those warnings with gcc 4.6.1 > and many many more of the same kind. Of course, gcc-3.4.6 is not exactly > recent. However, looking at your code there are many instances with > explicit typecasts, and just a few without such casts causing these > warnings. Given that I use a noticeable newer gcc maybe this is actually a "gcc bug"? Sincerely I would prefer avoiding adding casts. Albert > > But as I said in the original mail, it would be nice if these warnings could > be avoided, but this is by no means necessary. > > Regards > Peter Breitenlohner <[email protected]> > _______________________________________________ > poppler mailing list > [email protected] > http://lists.freedesktop.org/mailman/listinfo/poppler _______________________________________________ poppler mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/poppler
