Re: Bug#759290: debian-installer: 'Automated install' does not delay the locale and keyboard questions
Control: reopen -1 Samuel Thibault writes: > Control: tags -1 + pending > > Hello, > > Andreas B. Mundt, le Tue 26 Aug 2014 08:16:01 +0200, a écrit : >> Before, I was using the method "B.2.5. Using a DHCP server to specify >> preconfiguration files" described in [1]. Is there a reason not to >> delay the questions with that method? > > This apparently takes back March 2010: locale & kbd selection was > disabled only on explicit URL on boot parameters. I've made it always > done whenever auto is enabled, as I don't see why we shouldn't do it. > Fixing fully-automated-from-dhcp seems more important to me than the > inconvenience of a qwerty layout for typing the URL if really none is > available. Regarding this commit: http://anonscm.debian.org/cgit/d-i/preseed.git/commit/?id=f2e7bae02caec5a66e25d5af7aab90eb7648deca Please would you explain what you were trying to do here, as I don't think you understood the intent behind the original code or perhaps what the bug was really about. The reason I say that is that your attempted fix completely mangles the logic of the if statement. The &&'s after the db_get's are about checking that the setting is set, you cannot invert those and get a useful result. I'm now unsure how to fix this without reintroducing this bug. One could restore the if statement, with an empty then, but I think that the real bug is that we don't have a test for the DHCP provided preseed, so the real fix would be to add such a test to the if as it was before. Cheers, Phil. -- |)| Philip Hands [+44 (0)20 8530 9560] HANDS.COM Ltd. |-| http://www.hands.com/http://ftp.uk.debian.org/ |(| Hugo-Klemm-Strasse 34, 21075 Hamburg,GERMANY signature.asc Description: PGP signature
Re: Bug#759290: debian-installer: 'Automated install' does not delay the locale and keyboard questions
Philip Hands, le Fri 18 Sep 2015 09:46:43 +0100, a écrit : > The &&'s after the db_get's are about checking that the setting is set, > you cannot invert those and get a useful result. Why not? AIUI, db_get is a shell function which returns true when it gets a value, which it puts in RET. Why shouldn't inverting the tests work? We still preserve the sequentiality of db_get call and testing for RET. Did you notice an actual bug? > One could restore the if statement, with an empty then, Yes, that's a possibility, but I still don't see why it's needed. > but I think that the real bug is that we don't have a test for the > DHCP provided preseed, so the real fix would be to add such a test to > the if as it was before. Yes, except that at that point of the d-i boot, we have *not* had the DHCP preseed, so it's impossible to test it. So we don't have any other choice than disabling locale & kbd anyway, so DHCP can be tried without asking any question. Samuel
Bug#796591: debian-installer: beep on x86 UEFI for accessibility
Cyril Brulebois, le Mon 14 Sep 2015 10:20:38 +0200, a écrit : > I've pushed two commits to the pu/jessie branch. Thanks! > Double-checking/testing is more than welcome. I tested it, they seem to just work as appropriate. Samuel
