| ogra_ | cjwatson, seems i found a better way than mangling diversions for flash-kernel ... http://paste.ubuntu.com/1137623/ | 10:54 |
|---|---|---|
| ogra_ | (just struck me that we re-introduced this in ubuntu for live-build) | 10:55 |
| cjwatson | I'm afraid that won't help, because it only affects direct children of live-installer.postinst | 11:07 |
| cjwatson | We need to suppress any update-initramfs calls until flash-kernel-installer.postinst gets round to running | 11:07 |
| cjwatson | If you do it that way, and pkgsel ends up installing anything that updates the initramfs, then it will fail | 11:08 |
| ogra_ | hmm, where can i set it globally ? | 11:08 |
| ogra_ | flash-kernel-installer can forcefully unset it then | 11:09 |
| xnox | ogra_: divert update-initramfs to /bin/true in preinst and finally remove that diversion in the flash-kernel-installer.postinst and run it again.... ugly but.... | 11:10 |
| ogra_ | and i dont mind update-initramfs calls ... i only mind the kernel and initramfs hooks (which that var specifically supresses) | 11:10 |
| xnox | oh. ok. | 11:11 |
| cjwatson | I did think about this. I believe the diversion is the best answer. | 11:11 |
| ogra_ | but then the diversion should also happen globally instead of having it done by single udebs imho | 11:12 |
| cjwatson | No, because the reason for it is specific to the needs of flash-kernel. | 11:12 |
| cjwatson | It's entirely reasonable for that to be done in flash-kernel-installer. | 11:12 |
| cjwatson | In fact, it is best for it to be entirely contained within flash-kernel-installer. | 11:12 |
| cjwatson | (Modulo the pkgsel tweak, which is really - IMO - a fix for something pkgsel was doing slightly wrong.) | 11:13 |
| ogra_ | that still means i need to hack up pkgsel to mangle DIVERTS= if the f-k diversion exists | 11:13 |
| cjwatson | You said you were going to do that. Are you having problems? | 11:13 |
| cjwatson | It should take about five minutes. | 11:13 |
| cjwatson | I am happy to help if you want. | 11:14 |
| ogra_ | no, but its a lot more code overall than the one export line :) | 11:14 |
| cjwatson | But the one export line is wrong. | 11:14 |
| ogra_ | cjwatson, http://paste.ubuntu.com/1137676/ | 11:33 |
| cjwatson | I'd suggest doing the check up top when setting DIVERTS instead. It should be easier and cleaner. | 11:36 |
| ogra_ | well, lazyness, there was a loop already ;) | 11:36 |
| * ogra_ puts it at the top | 11:36 | |
| cjwatson | i.e. DIVERTS=; for divert in /usr/bin/fc-cache /usr/sbin/update-initramfs; do if chroot /target blah; then DIVERTS="$DIVERTS $divert"; fi; done | 11:37 |
| cjwatson | only with spacing and such | 11:37 |
| ogra_ | ah, saves the sed calls, ok | 11:37 |
| cjwatson | Anything where I have to count sed backslashes probably means there's a better way to write it :) | 11:37 |
| ogra_ | haha | 11:37 |
| ogra_ | k, so it boils down to http://paste.ubuntu.com/1137684/ | 11:43 |
| ogra_ | cjwatson, ^^ if you dont shout, i'll upload now :) | 12:13 |
| cjwatson | looks fine | 12:18 |
| ogra_ | uploaded :) | 12:19 |
| === cjwatson_ is now known as cjwatson | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!