Re: [28266] trunk/dports/x11/gtk2/Portfile
On Aug 26, 2007, at 20:41, source_changes@macosforge.org wrote:
Revision: 28266 http://trac.macosforge.org/projects/macports/changeset/28266 Author: rhwood@macports.org Date: 2007-08-26 18:41:47 -0700 (Sun, 26 Aug 2007)
Log Message: ----------- Possible fix for Mac OS X 10.3.x build failures
Modified Paths: -------------- trunk/dports/x11/gtk2/Portfile
Modified: trunk/dports/x11/gtk2/Portfile =================================================================== --- trunk/dports/x11/gtk2/Portfile 2007-08-27 00:19:23 UTC (rev 28265) +++ trunk/dports/x11/gtk2/Portfile 2007-08-27 01:41:47 UTC (rev 28266) @@ -53,7 +53,7 @@ return -code 1 "\nYou must first build cairo with the quartz variant enabled. Please\nuninstall (or deactivate) the cairo port and reinstall by running:\n\n\"port install cairo +quartz\"\n" } } - if {[file exists ${prefix}/bin/cups-config]} { + if {${os.platform} == "darwin" && [rpm-vercomp ${os.version} 8.0] < 0 && [file exists ${prefix}/bin/cups-config]} { return -code 1 "\nThe cups-headers port may prevent building this port. Please uninstall\n(or deactivate) cups-headers and restart the build.\n" } } @@ -71,6 +71,11 @@
platform darwin 6 { patchfiles-append patch-gtk-xdgmime- xdgmimemagic.c }
+platform darwin 7 { + depends_build-append \ + port:cups-headers +} + variant quartz { configure.args-append --with-gdktarget=quartz }
post-patch { reinplace "s|xdg_data_dirs = \"/usr|xdg_data_dirs = \"${prefix}/share:/usr|g" ${worksrcpath}/gtk/xdgmime/xdgmime.c
I'm a little confused... It sounds like the port now a) depends on cups-headers on darwin 7, and b) when darwin < 8, complains that cups- headers may conflict, and advises uninstalling it. Aren't those two contradictory?
On 26 Aug 2007, at 23:51, Ryan Schmidt wrote:
On Aug 26, 2007, at 20:41, source_changes@macosforge.org wrote: [snip]
} - if {[file exists ${prefix}/bin/cups-config]} { + if {${os.platform} == "darwin" && [rpm-vercomp ${os.version} 8.0] < 0 && [file exists ${prefix}/bin/cups-config]} { return -code 1 "\nThe cups-headers port may prevent building this port. Please uninstall\n(or deactivate) cups-headers and restart the build.\n" }
[snip]
I'm a little confused... It sounds like the port now a) depends on cups-headers on darwin 7, and b) when darwin < 8, complains that cups-headers may conflict, and advises uninstalling it. Aren't those two contradictory?
You're right, the "<" should be ">=" thanks for catching that. Now if only man portfile said more than "Compare two RPM-format versions for equality." when describing rpm-vercomp Randall Wood rhwood@mac.com http://shyramblings.blogspot.com "The rules are simple: The ball is round. The game lasts 90 minutes. All the rest is just philosophy."
Randall Wood wrote:
I'm a little confused... It sounds like the port now a) depends on cups-headers on darwin 7, and b) when darwin < 8, complains that cups-headers may conflict, and advises uninstalling it. Aren't those two contradictory?
You're right, the "<" should be ">=" thanks for catching that. Now if only man portfile said more than "Compare two RPM-format versions for equality." when describing rpm-vercomp
Currently it says: rpm-vercomp versionA versionB Compare two RPM-format versions for equality. The source code has the additional information: "RPM compatible version comparison * If A is newer than B, return an integer > 0 * If A and B are equal, return 0 * If B is newer than A, retun an integer < 0" (the spelling error has probably been there a while, and no idea why function is named comp and not cmp) The function is called "rpmvercmp" in RPM, and it is found in <rpm/rpmevr.h> nowadays... http://rpm5.org/docs/api/group__rpmds.html (it was re-implented using EVRcmp in 4.4.7) --anders PS. The return values are the same as for strcmp(3).
participants (3)
-
Anders F Björklund
-
Randall Wood
-
Ryan Schmidt