2014-01-15 00:24:27 |
Thomas Ward |
description |
In trusty-proposed, wireshark 1.10.5-1 fails to build from source.
The build log for this is at https://launchpadlibrarian.net/162346177/buildlog_ubuntu-trusty-amd64.wireshark_1.10.5-1_FAILEDTOBUILD.txt.gz
----
I looked into this, as I use the latest wireshark as part of the networking courses I am taking, and figured out that this is handled upstream. The upstream bug is here: https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=9340
The FTBFS in the version in trusty-proposed is introduced by the configure.ac file not permitting the use of deprecated features.
Upstream revisions to wireshark have permitted the software to build by allowing the use of deprecated features. Upstream has not said that they would fix the issue of deprecated features because they are moving to Qt in some future release according to that upstream bug.
The upstream patch, when applied, correctly fixes the FTBFS that is listed in the aforementioned build log. That patch is here: http://anonsvn.wireshark.org/viewvc/trunk/configure.ac?r1=54337&r2=54336&pathrev=54337&view=patch
However, that patch is insufficient to fix the FTBFS. Through testing with the upstream patch, there are additional FTBFS problems, and according to upstream, this is due to the C99 `g_memmove()` function not existing in recent glib versions. Therefore, Wireshark upstream has fallen back to using the C90 `memmove()` function.
----
The attached debdiff contains the following changes that should resolve the FTBFS:
* Update maintainer field with `update-maintainer` command.
* Add new patch debian/patches/allow-deprecated-gtk-functions.patch, which contains code derived from the upstream patch to allow the use of deprecated commands if Gtk is 3.10 or newer
* Add new patch debian/patches/fix-g_memmove-ftbfs-issues.patch, authored by Thomas Ward (teward), to replace instances of `g_memmove()` with `memmove()` to fix additional FTBFS problems.
* New debian/changelog entry describing these changes. |
In trusty-proposed, wireshark 1.10.5-1 fails to build from source.
The build log for this is at https://launchpadlibrarian.net/162346177/buildlog_ubuntu-trusty-amd64.wireshark_1.10.5-1_FAILEDTOBUILD.txt.gz
----
I looked into this, as I use the latest wireshark as part of the networking courses I am taking, and figured out that this is handled upstream. The upstream bug is here: https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=9340
The FTBFS in the version in trusty-proposed is introduced by the configure.ac file not permitting the use of deprecated features.
Upstream revisions to wireshark have permitted the software to build by allowing the use of deprecated features. Upstream has not said that they would fix the issue of deprecated features because they are moving to Qt in some future release according to that upstream bug.
The upstream patch, when applied, correctly fixes the FTBFS that is listed in the aforementioned build log. That patch is here: http://anonsvn.wireshark.org/viewvc/trunk/configure.ac?r1=54337&r2=54336&pathrev=54337&view=patch
However, that patch is insufficient to fix the FTBFS. Through testing with the upstream patch, there are additional FTBFS problems, and according to upstream, this is due to the C99 `g_memmove()` function not existing in recent glib versions. Therefore, Wireshark upstream has fallen back to using the C90 `memmove()` function.
----
The attached debdiff contains the following changes that should resolve the FTBFS:
* Update maintainer field with `update-maintainer` command, which may not be necessary. (This change was done to satisfy `debuilder -S` so that the package could build for testing inside of sbuild schroots.)
* Add new patch debian/patches/allow-deprecated-gtk-functions.patch, which contains code derived from the upstream patch to allow the use of deprecated commands if Gtk is 3.10 or newer
* Add new patch debian/patches/fix-g_memmove-ftbfs-issues.patch, authored by Thomas Ward (teward), to replace instances of `g_memmove()` with `memmove()` to fix additional FTBFS problems.
* New debian/changelog entry describing these changes. |
|