Ah, I see. Yes, libjpeg is built from source on the build machines, but only temporarily. It could be "installed" in /tmp, for instance, and removed as part of the build process, after PIL is built, for instance. I just used SAGE_LOCAL without thinking too hard about it.
I'm not familiar yet with the spkg infrastructure, so perhaps it's impossible to fit a step like this into the Sage build process. On Nov 13, 1:31 pm, William Stein <wst...@gmail.com> wrote: > On Sun, Nov 13, 2011 at 1:23 PM, Bill Janssen <bill.jans...@gmail.com> wrote: > > My suggestion doesn't require binary build machines to have libjpeg > > installed. In fact, you don't want it installed. > > Your suggestion for Sage binaries is: > > "1. Build libpng and libjpeg, static-only, and install to SAGE_LOCAL/ > lib/. ... > " > > This of course requires a build machine to have libjpeg built on it, > say as part of the Sage build process. So that Sage works in a > consistent manner across platforms, releases, and people building > binaries, this should only happen if libjpeg is added to Sage itself > as an spkg. I'm not necessarily opposed to that happening, since jpeg > is a common format. > > -- William > > > > > > > > > > > And there seems to > > be no technical reason to not statically link PIL against libjpeg.a. > > > But there *is* a reason to not *dynamically* link against > > libjpeg.dylib, libpng.dylib, or libtiff.dylib. OS X includes versions > > of these libraries in /System/Library/Frameworks/ > > ApplicationServices.framework/Versions/A/Frameworks/ImageIO.framework/ > > Versions/A/Resources/. They have different names, but the names > > differ only in case, which isn't significant to the dynamic linker on > > OS X. And Sage sets DYLD_LIBRARY_PATH. This means that any system > > program (like Preview.app, which Imaging-1.1.7 uses for im.show()), > > will fail if started from inside Sage, as the newer Sage-supplied > > dylibs will be different from the (usually obsolete) versions the > > system libraries are dynamically linked against. Unless the image > > processing libraries are statically linked where needed in Sage, in > > which case it works fine. > > > On Nov 13, 12:58 pm, William Stein <wst...@gmail.com> wrote: > >> On Sun, Nov 13, 2011 at 12:47 PM, Bill Janssen <bill.jans...@gmail.com> > >> wrote: > >> > What's the objection to linking PIL statically against libjpeg? It > >> > would remove a lot of uncertainty in the build. PIL seems to be the > >> > only place libjpeg is used. > > >> We should not require binary build machines to have libjpeg installed. > > >> If we include libjpeg, then there is no reason to statically link. > > >> -- William > > >> > On Nov 13, 12:44 pm, William Stein <wst...@gmail.com> wrote: > >> >> On Sun, Nov 13, 2011 at 12:10 PM, Bill Janssen <bill.jans...@gmail.com> > >> >> wrote: > >> >> > Yes, I think that's right, Justin. > > >> >> > I just tried this: > > >> >> > 1. Build libpng and libjpeg, static-only, and install to SAGE_LOCAL/ > >> >> > lib/. > >> >> > 2. Build Imaging-1.1.7 (after setting CPPFLAGS and LDFLAGS to point > >> >> > to SAGE_LOCAL), and install to SAGE_LOCAL/lib. > >> >> > 3. Remove my /opt/local/lib/ directory. > > >> >> > Now Image.open() and .show() work fine. > > >> >> > I think this should become part of the Sage build process, if it isn't > >> >> > already. > > >> >> No way. Either we include libjpeg properly with Sage (and link it > >> >> dynamically), or we make sure that our build machines don't have stuff > >> >> in /opt that produces bad binaries. > > >> >> I think including libjpeg at some point is reasonable, assuming that > >> >> it is legal these days. (Is it?) > > >> >> -- William > > >> >> > That way you won't wind up with "improper builds". You can > >> >> > delete the PNG and JPEG libraries after building PIL, as they're now > >> >> > statically linked into the PIL libraries. > > >> >> > So, why not build libpng and libjpeg dynamically? Because Sage also > >> >> > sets DYLD_LIBRARY_PATH (usually a bad indicator on Macs), and the > >> >> > dynamic version would override the versions various system frameworks > >> >> > (like ImageIO) are supposed to link against. Might it make more sense > >> >> > to use DYLD_FALLBACK_LIBRARY_PATH for this purpose? > > >> >> > Bill > > >> >> > On Nov 13, 11:50 am, "Justin C. Walker" <jus...@mac.com> wrote: > >> >> >> On Nov 13, 2011, at 09:59 , Bill Janssen wrote: > > >> >> >> > You're implying that 4.7.1 or 4.7.2 fix this issue in some way? I > >> >> >> > don't see anything in the release notes which would cause me to > >> >> >> > suspect that. > > >> >> >> Unless I'm still too caffeine-deprived, I think the issue is that > >> >> >> you are one of the first to try the Mac OS X 10.5 binary release of > >> >> >> 4.7, at least with something involving this particular library. > > >> >> >> There is no (i.e., doesn't seem to be an) issue for those who are > >> >> >> using either a different binary or have built from source (at least, > >> >> >> your primordial example works for me) (once I change my name to > >> >> >> "wjanssen" :-}). > > >> >> >> This doesn't seem to be a problem in Sage that needs to be fixed. > >> >> >> It seems to be a problem with one (or perhaps more than one) binary > >> >> >> that has been built improperly. > > >> >> >> If someone else doesn't verify this first, I'll try to get to it > >> >> >> tomorrow. > > >> >> >> Justin > > >> >> >> -- > >> >> >> Justin C. Walker, Curmudgeon at Large > >> >> >> Director > >> >> >> Institute for the Enhancement of the Director's income > >> >> >> ----------- > >> >> >> -- > >> >> >> They said it couldn't be done, but sometimes, > >> >> >> it doesn't work out that way. > >> >> >> - Casey Stengel > >> >> >> -- > > >> >> > -- > >> >> > To post to this group, send email to sage-support@googlegroups.com > >> >> > To unsubscribe from this group, send email to > >> >> > sage-support+unsubscr...@googlegroups.com > >> >> > For more options, visit this group > >> >> > athttp://groups.google.com/group/sage-support > >> >> > URL:http://www.sagemath.org > > >> >> -- > >> >> William Stein > >> >> Professor of Mathematics > >> >> University of Washingtonhttp://wstein.org > > >> > -- > >> > To post to this group, send email to sage-support@googlegroups.com > >> > To unsubscribe from this group, send email to > >> > sage-support+unsubscr...@googlegroups.com > >> > For more options, visit this group > >> > athttp://groups.google.com/group/sage-support > >> > URL:http://www.sagemath.org > > >> -- > >> William Stein > >> Professor of Mathematics > >> University of Washingtonhttp://wstein.org > > > -- > > To post to this group, send email to sage-support@googlegroups.com > > To unsubscribe from this group, send email to > > sage-support+unsubscr...@googlegroups.com > > For more options, visit this group > > athttp://groups.google.com/group/sage-support > > URL:http://www.sagemath.org > > -- > William Stein > Professor of Mathematics > University of Washingtonhttp://wstein.org -- To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to sage-support+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-support URL: http://www.sagemath.org
