[13:54] <Laney> tumbleweed: dmb agenda poke
[17:12] <TJ-> cjwatson: I'm trying to confirm that bug #1271141 is fixed but don't have easy access to the original system that had Win7 installed; do you know what the minimum requirement is for os-prober to believe it has found a Windows bootable file-system (so I can fake one) ?
[17:14] <cjwatson> TJ-: the place to look would be /usr/lib/os-probes/mounted/20microsoft
[17:15] <TJ-> cjohnston: yeah, been there already. Can't find (so far) where WINOSDATA is defined and a batch of tests seem to depend on it being non-empty
[17:15] <TJ-> oops
[17:15] <TJ-> cjwatson:  yeah, been there already. Can't find (so far) where WINOSDATA is defined and a batch of tests seem to depend on it being non-empty
[17:17] <cjwatson> TJ-: you can just pretend it's always empty; that's a hook for migration-assistant IIRC
[17:18] <TJ-> cjwatson: Ahhhh.... explains why I couldn't find it in os-prober or grub sources
[17:19] <cjwatson> Or maybe ubiquity, I'm not sure.  It's not in the current source for anything relevant I can find, so maybe we can drop that patch soon
[17:20] <cjwatson> Apart from a similar test in ubiquity/compat/os-prober
[17:20] <cjwatson> Not going to track down the history just now :)
[17:20] <TJ-> cjwatson: So it looks like it needs "/boot/bcd" with valid Win7 entries
[17:20] <cjwatson> Sounds plausible
[17:20] <cjwatson> Plus appropriate filesystem type
[17:20] <TJ-> cjwatson: that's fine, I think I have enough to fake it now
[17:21] <cjwatson> ok, cool
[17:41] <TJ-> cjwatson: fwiw, fake only requires: "mkdir -p /mnt/tmp && mount /dev/sdXY /mnt/tmp && touch /mnt/tmp/bootmgr && mkdir -p /mnt/tmp/boot && echo 'W.i.n.d.o.w.s. .7' >/mnt/tmp/boot/bcd && umount /mnt/tmp"
[18:04] <TJ-> cjwatson: Do you know of a 'nice' way to co-install grub-pc and grub-efi so that both can be used? The scenario is a USB device being configured to boot on UEFI or BIOS systems, with a GPT
[21:08] <Noskcaj> Does dh_autoreconf support cmake yet?
[21:09] <Noskcaj> if so, gccxml will need a rebuild
[21:09] <juliank> Noskcaj: You can use it with a custom command.
[21:09] <juliank> There's only magic for autoreconf.
[21:10] <juliank> As far as I know, cmake generated files are not shipped in a tarball, so there should not be any need.
[22:10] <infinity> Noskcaj: gccxml needs a *lot* more than autoreconf...
[22:12] <infinity> Noskcaj: It's a fork of an ancient GCC, so doesn't remotely support any arch added to GCC since that point.
[22:12] <jtaylor> hm does that matter?
[22:13] <jtaylor> its only working on the frontend
[22:13] <jtaylor> its problem is that it doesn't support c++11
[22:13] <jtaylor> which is a shame as it allows some neat stuff (e.g. the reflexion of root-system)
[22:47] <xnox> jtaylor: my alternative was to use a fork of gccxml which is just a plugin, rather than embedded copy of gcc.
[22:48] <jtaylor> xnox: link?
[22:48] <xnox> jtaylor: https://github.com/alexleach/gccxml_plugin that one works with gcc-4.8
[22:48] <jtaylor> neat thanks
[22:48] <xnox> jtaylor: it's highly compatible with old gccxml, but not fully.
[23:01] <Noskcaj> infinity, oh. /me runs away from that package
[23:18] <Noskcaj> pitti, PING