[12:09] morning [13:17] hm, so I'm doing a build with `export DEB_BUILD_MAINT_OPTIONS=optimize=-lto` [13:17] but I see it kreeping back in via cflags from other libraries, like [13:17] -- Found GSSAPI: -L/usr/lib/x86_64-linux-gnu/mit-krb5 -Wl,-Bsymbolic-functions -flto=auto -ffat-lto-objects -flto=auto -Wl,-z,relro -lgssapi_krb5 -lkrb5 -lk5crypto -lcom_err [13:18] so when the time comes during the build to link with gssap (in this case), -lto is back [13:18] this can't be good, having just parts of the build using lto, and other parts not? [13:19] doko: any tip/insight? ^ [13:28] I have vague memories of the gssapi pkg-config dumping way too much stuff into its Cflags field at some point [13:52] I remember that too, it was another flag we were disabling iirc [13:53] symbolic-functions [13:53] https://lists.ubuntu.com/archives/ubuntu-devel/2018-June/040371.html [13:59] changing all usage of krb5-config to pkg-config would be a good starting point [14:01] isn't it a policy item to not ship pkg-config files in dev packages? [14:01] or lintian [14:01] I vaguely remember something about it [14:02] oh, lintian is very unhappy with krb5: https://lintian.debian.org/sources/krb5 [14:14] ahasenack: I would consider it a mistake to record these flags in any *-config file/package [14:14] doko: so bug in src:krb5? [14:14] doko: that being said, is it an issue if krb5 was built with lto, and then another package which links with krb5 is NOT built with lto? [14:38] ahasenack: it's *possible*, but I've mentioned a couple of times that I'm concerned about increasing divergence between the buildd environment and the distro (which could lead to things being built and passing tests that won't work in normal environments), and if you've addressed that concern then I missed it [14:38] ahasenack: launchpad-buildd comes from https://git.launchpad.net/launchpad-buildd [14:38] ok [14:38] I understand the concern [14:38] and new info came to light today, we won't be needing that change [14:39] I mean, the test will still not use uring, because memlock is too low, but at least it won't crash [14:39] ahasenack: OK. To be clear I'm not necessarily vetoing, I just have/had a concern that I want to make sure is taken into account and addressed as far as possible before doing this kind of thing [14:39] there is still a valid point in debating whether we should bump that limit distro-wide for focal and others [14:39] so I added a systemd task to the bug [14:40] might email ubuntu-devel@ to create a thread about it [15:11] Trevinho: hi, I'm the dev of Ubuntu Unity and was interested in removing the transparent border around the dash and hud for a flatter look, but couldn't find any code related to that (tried setting the border height and width to 0 in DashStyle.cpp, but that didn't remove the border) === rs20097 is now known as rs2009 [17:12] rs2009: IIRC that can be themed via files [17:13] rs2009: check if I wrote that in the past in https://wiki.ubuntu.com/Unity/Theming [17:13] nah, it doesn't seem lik ethat === chrisccoulson_ is now known as chrisccoulson [17:22] hey Trevinho: oh, I replaced those files with empty transparent 1x1 images and it seems to work great [17:23] I was trying to change the version number to 7.6 as I've fixed quite a few bugs and added features, but changing it in CMakeLists and dch doesn't change the version to 7.6 (although 7.59 works) [17:24] Khurshid suggested changing libunity9 too, but I wasn't sure where to change the version in it (the configure.ac file doesn't contain anything related to Unity 7.5) [17:51] Trevinho: ^ [19:20] Hi bdmurray, can you please remove im-config from the jammy unapproved queue. (I have changed my mind and will address the issue via a gnome-settings-daemon patch instead.)