Bug 866 - perl-native conflicts with host perl within autotools
Summary: perl-native conflicts with host perl within autotools
Status: RESOLVED FIXED
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: devtools / tool chain (show other bugs)
Version: unspecified
Hardware: x86 x86_64
: Medium major
Target Milestone: 1.1
Assignee: Nitin Kamble
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2011-03-14 04:37 UTC by Paul Eggleton
Modified: 2011-03-21 07:29 UTC (History)
4 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Paul Eggleton 2011-03-14 04:37:07 UTC
I'm getting what looks to be a reproducible problem between gnu-config-native, autotools and perl-native. It seems that if perl-native has been built when autoconf-native/automake-native get built then they attach themselves to perl-native; unfortunately this results in gnu-configize failing during autoreconf (during configure of of libtool-native in the example I have found) due to a version mismatch between host perl on my system (5.12.3) and the perl-native libraries in the sysroot (5.12.2).  

If you then clean perl-native and try again it just complains of perl being missing from the sysroot, with no attempt to rebuild perl-native either. This is I think at least part of the issue I was seeing in bug #768.

Steps to reproduce:

1. bitbake -c cleanall perl-native libtool-native gnu-config-native autoconf-native automake-native
2. bitbake perl-native
3. bitbake automake-native
4. bitbake libtool-native

If you perform the above but skip step 2, then all is well.
Comment 1 Saul Wold 2011-03-14 15:16:34 UTC
This will require invasive changes and is too late for 1.0.
Add to release notes.
Comment 2 Paul Eggleton 2011-03-15 03:30:25 UTC
There are a couple of patches in OE that will likely solve this and it's not too invasive:

http://lists.linuxtogo.org/pipermail/openembedded-core/2011-March/000462.html