After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 728972 - Missing symbol in webextensions.so
Missing symbol in webextensions.so
Status: RESOLVED INCOMPLETE
Product: epiphany
Classification: Core
Component: General
3.12.x (obsolete)
Other Mac OS
: Normal normal
: ---
Assigned To: Epiphany Maintainers
Epiphany Maintainers
Depends on:
Blocks:
 
 
Reported: 2014-04-25 20:13 UTC by Gilles Dartiguelongue
Modified: 2015-03-31 00:28 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
0001-Add-missing-files-to-webextension-library.patch (1.40 KB, patch)
2014-04-25 20:13 UTC, Gilles Dartiguelongue
none Details | Review

Description Gilles Dartiguelongue 2014-04-25 20:13:19 UTC
Created attachment 275164 [details] [review]
0001-Add-missing-files-to-webextension-library.patch

When starting epiphany from terminal or checking journalctl:

avril 22 21:09:36 kanae epiphany.desktop[32013]: Error loading module '/usr/lib64/epiphany/3.12/web-extensions/libephywebextension.so': /usr/lib64/epiphany/3.12/web-extensions/libephywebextension.so: undefined symbol: ephy_profile_utils_set_migration_version

This appears to be due to missing sources found in lib.

Attached patch solves the issue.
Comment 1 Carlos Garcia Campos 2014-07-21 11:23:41 UTC
hmm, the thing is that the dot_dir should already exist when the web extension is started, because ephy_file_helpers_init() is called by ephy-main.c with EPHY_FILE_HELPERS_ENSURE_EXISTS flag and the profile dot dir. And that profile dot dir is passed to the web extension, that calls ephy_file_helpers_init() without the EPHY_FILE_HELPERS_ENSURE_EXISTS flag. So we shouldn't really need that symbol in the web extension.
Comment 2 Michael Catanzaro 2014-08-10 21:19:02 UTC
My standard response to bugs from Gentoo is "please post your CFLAGS and LDFLAGS".
Comment 3 Michael Catanzaro 2015-03-29 20:57:49 UTC
Marking INCOMPLETE since the requested information was not provided
Comment 4 Alexandre Rostovtsev 2015-03-30 14:37:33 UTC
(In reply to Michael Catanzaro from comment #2)
> My standard response to bugs from Gentoo is "please post your CFLAGS and
> LDFLAGS".

You need to change your standard response. Gentoo *devs* - meaning people with "@gentoo.org" emails - are perfectly aware of the difference between crazy and sane CFLAGS/LDFLAGS, and generally stick with "-march=native -O2" or equivalent. The same is true for experienced Gentoo users.

What you should have asked in this case is what *toolchain* was being used by the Gentoo user.

A probable reason why you couldn't reproduce the problem because you were testing with an older version of gcc and the ld.bfd linker, while the reporter was using the newest gcc release and ld.gold.
Comment 5 Michael Catanzaro 2015-03-31 00:28:53 UTC
OK, thanks, let's try that question then: what toolchain are you using? It works fine for me when build with ld.gold and GCC 4.9.2....