Bug 3267

Summary: apr and apr-util config scripts produce incorrect paths
Product: [Build System, Metadata & Runtime] OE-Core Reporter: Paul Eggleton <bluelightning>
Component: coreAssignee: Saul Wold <sgw>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: Medium CC: meta.mr.watcher, meta.watcher
Version: 1.3   
Target Milestone: 1.4   
Hardware: All   
OS: Multiple   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: ---

Description Paul Eggleton 2012-10-10 08:46:52 UTC
The apr-1-config and apu-1-config scripts from apr and apr-util respectively have incorrect paths pointing into the workdir for the recipe, for example:

$ /media/large/tmp/sysroots/qemuarm/usr/bin/crossscripts/apr-1-config --apr-libtool
/media/large/tmp/work/armv5te-poky-linux-gnueabi/apr-1.4.6-r1/apr-1.4.6/libtool
$ /media/large/tmp/sysroots/qemuarm/usr/bin/crossscripts/apr-1-config --includedir
/media/large/tmp/work/armv5te-poky-linux-gnueabi/apr-1.4.6-r1/apr-1.4.6/include

What should be produced:
$ /media/large/tmp/sysroots/qemuarm/usr/bin/crossscripts/apr-1-config --apr-libtool
/media/large/tmp/sysroots/qemuarm/usr/share/build-1/libtool
$ /media/large/tmp/sysroots/qemuarm/usr/bin/crossscripts/apr-1-config --includedir
/media/large/tmp/sysroots/qemuarm/usr/include

The reason this is broken is that OE-Core's version of binconfig.bbclass does not implement OE_BINCONFIG_EXTRA_MANGLE, which is set by both of these recipes to change the config scripts so that they use the correct paths. We need to find the patch in OE-Classic that added this variable and bring it into OE-Core.

(This apparently does not break building subversion, which is the only thing in OE-Core that uses apr/apr-util; for apache2 in meta-webserver I have a workaround which is to make a copy of the script and then mangle it, so this is not an urgent issue).
Comment 1 Saul Wold 2012-12-04 01:16:33 UTC
Paul, is this the patch you are referring to from Holger?

http://git.openembedded.org/openembedded/commit/?id=030992bdd415c8892cf2925cf6fbe888226be046

If so, I can merge it and test
Comment 2 Paul Eggleton 2012-12-04 09:34:20 UTC
Looks like it, yes.