| Summary: | Revision for the internal jhbuild is too old | ||||||
|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Mario Sanchez Prada <mario> | ||||
| Component: | Tools / Tests | Assignee: | Mario Sanchez Prada <mario> | ||||
| Status: | RESOLVED FIXED | ||||||
| Severity: | Normal | CC: | d-r, gustavo, mrobinson, webkit.review.bot | ||||
| Priority: | P2 | ||||||
| Version: | 528+ (Nightly build) | ||||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Attachments: |
|
||||||
Created attachment 195116 [details]
Patch proposal
Here comes the patch
Thanks, Mario. Comment on attachment 195116 [details]
Patch proposal
Thanks for the review. Setting cq+, since I'm not able to push it myself, for some weird reason
Comment on attachment 195116 [details] Patch proposal Clearing flags on attachment: 195116 Committed r147023: <http://trac.webkit.org/changeset/147023> All reviewed patches have been landed. Closing bug. |
Today I found WebKitGTK is using a pretty old version of jhbuild for the internal libs, which has an important bug that makes the process to fail during the "make install" stage, in case it founds a symlink to a directory after moving it from the temporary place under _jhbuild/ to the final destination. The problem comes from this part of the code in jhbuild (revision 1eedc423f75c605224b430579e4c303292199507). In file jhbuild/jhbuild/modtypes/__initi__.py: if os.path.isdir(src_path): if os.path.exists(dest_path): if not os.path.isdir(dest_path): os.unlink(dest_path) os.mkdir(dest_path) else: os.mkdir(dest_path) num_copied += self._process_install_files(installroot, src_path, prefix) os.rmdir(src_path) That will fail if src_path is a symlink to a dir, since os.path.isdir() follows links, but os.rmdir() does not. So I think we should update the minimum required revision to some point where that issue is fixed, and I suggest ddb8b4e64bede55212c7590e2104daff92b72ce0 for that, since it's properly fixed there.