GNOME Bugzilla – Bug 621963
gtk-doc scanner fails to compile with the Sun Studio compiler
Last modified: 2010-06-21 18:06:01 UTC
Created attachment 163986 [details] [review] patch fixing problem. gst-plugins-bad 0.10.19 fails to build the gtk-docs on Solaris with the Sun Studio compiler because gstdoc-scangobj program adds the GCC specific -Wall to the compile command. Is this really needed. If so, can the addition of this flag be added conditionally if gcc is being used, but not otherwise. Attached is a patch that removes the offending argument so gst-plugins-bad can build on Solaris.
We have a WARNING_CFLAGS now, would exporting and using those make sense? (I suppose they would work, though testing that seems like a good idea, too.)
Yes, this is bogus. I'm assigining it to core as blocker bug, since it affects all modules and it will get noticed/fixed more quickly there (given that core/base/good are to be released next).
Created attachment 163989 [details] [review] alternative patch, uses WARNING_CFLAGS instead of hard-coding -Wall Alternative patch that sues WARNING_CFLAGS instead. Don't know if we need to pass warning flags here at all though. Also removed the -g.
Pushed a simpler version of the above, maintaining the warning flags: commit 277d83e3e02205f63e922cbd41572111c41ac5b6 Author: Tim-Philipp Müller <tim.muller@collabora.co.uk> Date: Fri Jun 18 09:54:20 2010 +0100 gstdoc-scangobj: fix build with C compilers that don't support -Wall Don't unconditionally pass -Wall to the compiler. Use WARNING_CFLAGS instead if it is defined. Fixes build with Sun Studio Compiler. Fixes #621963.
I meant to give you credit for the original patch Brian, but pushed a bit too fast. Sorry about that.
No, worry. My original patch wasn't much of a patch, it just removed the offending argument from the file to highlight what the problem was. You did most of the work. I'm just happy the issue was addressed promptly.