Little preseed questions
Hello Please, I need some (little) help in some preseeds. I'm searching for this questions in documentation and I havn't found anything. Would you say if exist some preseed (or preseeds) to get: a) Save the installationlog log (/var/log/syslog and partman files) in some location b) Change the default grub entry name Debian GNU/Linux kernel kernel number to a diferent string? (like d-i debian-installer/add-menu-opts string Newpreseeded Debian) c) Enable/Disable show grub menu (the hiddenmenu entry), default entry o timeout? d) The preseed to add kernel parameters (d-i debian-installer/add-kernel-opts string ) is compatible with lilo? e) The name of multipackages selections are? - File Server (file-server) - Mail Server (mail-server) - SQL Server (sql-database) - LAMP (lamp) - DNS (dns-server) - Laptop (laptop) - Print-server (print-server) - Web-server (web-server) f) Don't start X-server (or the login manager gdm/kdm) automatically on boot system? (like xserver-xorg xserver-xorg/enable-server boolean false) Sorry for my english. I'm sorry if are a lot of questions but I can't find on other way this answers. Thanks a lot in advance.
Re: List spam cleanup: half-time scores
Quoting Christian Perrier (bubu...@debian.org): by Holger and myself only these weeks. I still have to process the result for June 28th but I don't expect much more signalled spams. Only 11. Most of them being recent spam from June 2009 (probably signalled on the fly by -boot readers). signature.asc Description: Digital signature
Re: Reducing the complication of choices in console-setup udeb config: first thought
Quoting Samuel Thibault (sthiba...@debian.org): I'd even say, one language/country, one layout. Note for instance french, for which there are a lot of layouts, from BE, CA, CH, FR, ... But there is little chance getting wrong when taking the country into account too. That's why I proposed to first have a list for each country/language pair. Then a list for each language can be used if the user is not happy with the former. My point was *not* about showing a different list to users depending on the language they selected. It was just about building *one* list that would offer users all reasonable choices. Here is what I come with (to be completed...I stopped at Kazakh): No need to do it by hand, it can be automated from xkeyboard-config's data. Any divergence from a hand-written version is just a bug in xkeyboard-config. I don't really see how such a list could be derived from xkeyboard-config. What there does tell me that I need a keymap for fr_FR, another for fr_BE and fr_CH or fr_CAwhile de_DE and de_AT are happy with the same one? probably around 70 for i386|amd64 and much much less for arches where there is a need for a dedicated list Why a dedicated list? Because I see this as the only way to keep things simple. We would then have to just maintain that selection of common keymaps. This is about the same method we're using, now, in console-data: c-d has a big bunch of keymaps for each architectures (admittedly a lot for i386) out of which we select a few to be presented to users in a single question in D-I (while configuring the regular package sticks with a multi-level configuration process). My proposal indeed means creating a specific console-setup-udeb.config that would be *very* simple: asks just one question and set console-setup/* accordingly. After this initial discussion, I think I roughly have the needed scheme in mind and I can probably write down a patch with these ideas. Hopefully, I'll get enough time for this in the next weeks. signature.asc Description: Digital signature
Re: [!!!] D-I SVN repository not available Monday June 29 from 07:00 UTC
On Friday 26 June 2009, Frans Pop wrote: On Monday June 29 committing changes to the Debian Installer subversion repository will not be possible from around 07:00 UTC for maintenance. I expect the repository to be available again around 12:00 UTC at he latest, but please wait for the announcement. The reason for the downtime is a cleanup operation (see [1] for details). During most of that time reading from the repository should be possible, but if possible it is best to avoid using it completely. Attempts to commit changes during the cleanup will result in an error message. Reminder: D-I SVN commit access is now blocked for the next hours. I will mail again when the repository is available again. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reducing the complication of choices in console-setup udeb config: first thought
Christian Perrier, le Mon 29 Jun 2009 06:58:21 +0200, a écrit : My point was *not* about showing a different list to users depending on the language they selected. It was just about building *one* list that would offer users all reasonable choices. Ah. I'm still afraid that's not very scalable (you already announced ~70 choices, and I expect to see the list still growing, as there are something like a hundred scripts in Unicode, latin being just one while accounting for a lot of already existing keymaps...). When the default choice is correct, that should be ok, but for borderline cases, it will not be convenient for the user to browse it. Here is what I come with (to be completed...I stopped at Kazakh): No need to do it by hand, it can be automated from xkeyboard-config's data. Any divergence from a hand-written version is just a bug in xkeyboard-config. I don't really see how such a list could be derived from xkeyboard-config. With your point of view, that may be different indeed. What there does tell me that I need a keymap for fr_FR, another for fr_BE and fr_CH or fr_CAwhile de_DE and de_AT are happy with the same one? Thanks to the language/country tags attached to layouts and variants. The FR layout has a fra language tag, the BE layout has ger and a fra tags, the CH layout has ger and gsw tags, while its fr variant has a fra tag, the CA layout has a fra tag, the DE layout has a ger tag. There is no AT layout, but upstream will be happy to include one that just includes the DE one (and not the CH one). Yes, it would wear a different name than the DE one, but that's probably an opportunity to make the name clearer for the user. Yes, thus the list would get very long, and that's why I proposed to show several size-increasing versions. My point is: why should Debian maintain that information while it should already be in the xkb database? probably around 70 for i386|amd64 and much much less for arches where there is a need for a dedicated list Why a dedicated list? Because I see this as the only way to keep things simple. We would then have to just maintain that selection of common keymaps. Why only the common keymaps? To mimic Franc's way of asking: aren't people using uncommon keymaps allowed to install Debian? I do agree that we should blacklist exotic keymaps like my keymap with a lot of stuff to type unicode arrows in d-i, but there are keymaps for a lot of languages, and you are proposing to handle additions by hand, that's a pain. Samuel -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Little preseed questions
On Mon, Jun 29, 2009 at 08:13:41AM +0200, MaTa wrote: Hello Please, I need some (little) help in some preseeds. I'm searching for this questions in documentation and I havn't found anything. Would you say if exist some preseed (or preseeds) to get: a) Save the installationlog log (/var/log/syslog and partman files) in some location The package that does this is save-logs; I think what you want is save-logs/menu = mounted file system, and then save-logs/directory = /target/root or something. b) Change the default grub entry name Debian GNU/Linux kernel kernel number to a diferent string? (like d-i debian-installer/add-menu-opts string Newpreseeded Debian) I don't think this can be done; it's trivial to do in the late_command. d) The preseed to add kernel parameters (d-i debian-installer/add-kernel-opts string ) is compatible with lilo? Doesn't look like it. e) The name of multipackages selections are? That's tasksel/first; the actual set of options isn't static, unfortunately. I don't like installing packages during the install, so I don't know how to get the full list of tasks. f) Don't start X-server (or the login manager gdm/kdm) automatically on boot system? (like xserver-xorg xserver-xorg/enable-server boolean false) Don't install a display manager. - Matt -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [!!!] D-I SVN repository not available Monday June 29 from 07:00 UTC
On Monday 29 June 2009, Frans Pop wrote: On Friday 26 June 2009, Frans Pop wrote: On Monday June 29 committing changes to the Debian Installer subversion repository will not be possible from around 07:00 UTC for maintenance. I expect the repository to be available again around 12:00 UTC at he latest, but please wait for the announcement. The reason for the downtime is a cleanup operation (see [1] for details). During most of that time reading from the repository should be possible, but if possible it is best to avoid using it completely. Attempts to commit changes during the cleanup will result in an error message. Reminder: D-I SVN commit access is now blocked for the next hours. I will mail again when the repository is available again. I've had to abandon the cleanup for now because alioth was unacceptably slow (the cleanup would have taken a few days instead of a few hours...). I may try again this afternoon. The repository is now available again. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Little preseed questions
Thanks again 2009/6/29 Matthew Palmer mpal...@debian.org On Mon, Jun 29, 2009 at 08:13:41AM +0200, MaTa wrote: Hello Please, I need some (little) help in some preseeds. I'm searching for this questions in documentation and I havn't found anything. Would you say if exist some preseed (or preseeds) to get: a) Save the installationlog log (/var/log/syslog and partman files) in some location The package that does this is save-logs; I think what you want is save-logs/menu = mounted file system, and then save-logs/directory = /target/root or something. Yes, I think that can be in a late_command like cp /var/log/syslog /target/directory. I would known if there are some standard `reseed to do this like save-log/copy /target/directory b) Change the default grub entry name Debian GNU/Linux kernel kernel number to a diferent string? (like d-i debian-installer/add-menu-opts string Newpreseeded Debian) I don't think this can be done; it's trivial to do in the late_command. d) The preseed to add kernel parameters (d-i debian-installer/add-kernel-opts string ) is compatible with lilo? Doesn't look like it. e) The name of multipackages selections are? That's tasksel/first; the actual set of options isn't static, unfortunately. I don't like installing packages during the install, so I don't know how to get the full list of tasks. I will try with diferences of dpkg -l diferences with a standard instalation installed system and a laptop installation. f) Don't start X-server (or the login manager gdm/kdm) automatically on boot system? (like xserver-xorg xserver-xorg/enable-server boolean false) Don't install a display manager. Thanks, but I would install the x-server and display manager, but I don't want that start at boot. - Matt Thanks again
Re: Review of console-setup wrt D-I [very long]
On Sun, Jun 28, 2009 at 03:05:22PM +0200, Frans Pop wrote: On Sunday 28 June 2009, Colin Watson wrote: I think we should stop asking this question entirely. Hardly anyone really needs anything other than pc105, except for the Brazilian and Japanese cases that can be derived automatically from the keyboard layout anyway. How appropriate is pc105 as a default? None of the 7 keyboards [1] I have here has the extra LSGT key ( + ), so they are all pc104... As Steve says, the distinction is not one that is important to present to users. You don't care if you have an extra symbol on a key you don't have. You *would* care if you end up with an important symbol you can't type because you don't have the relevant key (this happens with Brazilian, where / goes on the 106th key ...), but if that happens you're using the wrong layout anyway. -- Colin Watson [cjwat...@debian.org] -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [D-I] Please test Debian Installer with console-setup
On Sat, Jun 27, 2009 at 10:12:19PM +0300, Anton Zinoviev wrote: On Sat, Jun 27, 2009 at 08:35:44PM +0200, Samuel Thibault wrote: Anton Zinoviev, le Sat 27 Jun 2009 18:52:34 +0300, a écrit : This was because setfont can not load compressed fonts without gzip. I changed console-setup-fonts-udeb to include uncompressed fonts. Console-setup-fonts-udeb is not required so if the space is tight it can be dropped. Err, isn't it less costly to include gzip than to include uncompressed fonts? Only if gzip is smaller than 90K. busybox already has gunzip. Just fix setfont and all will be well ... -- Colin Watson [cjwat...@debian.org] -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Review of console-setup wrt D-I [very long]
On Monday 29 June 2009, Colin Watson wrote: On Sun, Jun 28, 2009 at 03:05:22PM +0200, Frans Pop wrote: On Sunday 28 June 2009, Colin Watson wrote: I think we should stop asking this question entirely. Hardly anyone really needs anything other than pc105, except for the Brazilian and Japanese cases that can be derived automatically from the keyboard layout anyway. How appropriate is pc105 as a default? None of the 7 keyboards [1] I have here has the extra LSGT key ( + ), so they are all pc104... As Steve says, the distinction is not one that is important to present to users. You don't care if you have an extra symbol on a key you don't have. You *would* care if you end up with an important symbol you can't type because you don't have the relevant key (this happens with Brazilian, where / goes on the 106th key ...), but if that happens you're using the wrong layout anyway. OK. I wasn't sure about that, so I asked :-) Thanks both. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Online help
On Thu, Jun 25, 2009 at 12:09:44PM +0100, Colin Watson wrote: Here's an initial draft of a patch for this, tested interactively using the changes to src/test/help.templates included in the patch. It was surprisingly easy! I've only done the newt interface so far, not the GTK interface; please have a look over this, since I'd like to have general consensus that I'm on the right track before going any further. Here's a second version with added support for the text and GTK frontends, and a fix for a reference leak in the newt frontend implementation. Please review. I'd like to commit this towards the end of this week if there are no objections. Index: debian/changelog === --- debian/changelog(revision 59150) +++ debian/changelog(working copy) @@ -1,3 +1,12 @@ +cdebconf (0.143) UNRELEASED; urgency=low + + * Add online help support. The Help field in a template may refer to +another template whose description contains help text; frontends may use +this to display online help on request. Currently implemented for the +text, newt, and gtk frontends. + + -- Colin Watson cjwat...@debian.org Mon, 29 Jun 2009 13:04:55 +0100 + cdebconf (0.142) unstable; urgency=low [ Frans Pop ] Index: debian/cdebconf-newt-udeb.templates === --- debian/cdebconf-newt-udeb.templates (revision 59150) +++ debian/cdebconf-newt-udeb.templates (working copy) @@ -33,4 +33,11 @@ # Help line displayed at the bottom of the cdebconf newt interface. # Translators: must fit within 80 characters. # :sl1: -_Description: Tab moves between items; Space selects; Enter activates buttons +_Description: Tab moves; Space selects; Enter activates buttons + +Template: debconf/help-line-f1 +Type: text +# Help line displayed at the bottom of the cdebconf newt interface. +# Translators: must fit within 80 characters. +# :sl1: +_Description: F1 for help; Tab moves; Space selects; Enter activates buttons Index: debian/cdebconf-gtk-udeb.templates === --- debian/cdebconf-gtk-udeb.templates (revision 59150) +++ debian/cdebconf-gtk-udeb.templates (working copy) @@ -22,6 +22,12 @@ # :sl1: _Description: No +Template: debconf/button-help +Type: text +# Translators, this text will appear on a button, so KEEP IT SHORT +# :sl1: +_Description: Help + Template: debconf/text-direction Type: text # TRANSLATORS: This should be translated to RTL or LTR depending of Index: debian/copyright === --- debian/copyright(revision 59150) +++ debian/copyright(working copy) @@ -15,8 +15,8 @@ (c) Jason Gunthorpe j...@debian.org (derived portions are public domain) -CDebConf is copyrighted (c) 2000-2007 by Randolph Chung ta...@debian.org -and the d-i team (see above) under the following license: +CDebConf is copyrighted (c) 2000-2009 by Randolph Chung ta...@debian.org, +the d-i team (see above), and Canonical Ltd. under the following license: Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions Index: src/template.c === --- src/template.c (revision 59150) +++ src/template.c (working copy) @@ -21,6 +21,7 @@ indices, description, extended_description, +help, NULL }; @@ -134,6 +135,7 @@ DELETE(t-tag); DELETE(t-type); + DELETE(t-help); p = t-fields; DELETE(t); while (p != NULL) @@ -175,6 +177,7 @@ struct template_l10n_fields *from, *to; ret-type = STRDUP(t-type); +ret-help = STRDUP(t-help); if (t-fields == NULL) return ret; @@ -349,7 +352,7 @@ * Input: a field name * Output: the value of the given field in the given language, field * name may be any of type, default, choices, indices, - * description and extended_description + * description, extended_description and help * Description: get field value * Assumptions: */ @@ -367,6 +370,8 @@ return t-tag; else if (strcasecmp(field, type) == 0) return t-type; +else if (strcasecmp(field, help) == 0) +return t-help; /* If field is Foo-xx.UTF-8 then call template_lget(t, xx, Foo) */ if (strchr(field, '-') != NULL) @@ -428,7 +433,7 @@ * Input: a field name * Output: the value of the given field in the given language, field * name may be any of type, default, choices, indices, - * description and extended_description + * description, extended_description and help * Description: get field value * Assumptions: Arguments have been previously checked, lang and field * are not
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
Console-setup changed a lot since the long mail of Frans Pop. I will make an upload as soon as the current package migrates to testing. In the following I am summarizing some of the changes. On Sun, Jun 28, 2009 at 05:59:02PM +0200, Christian Perrier wrote: Size impact --- 1. The model and the ttys questions are removed. 2. The translations of the origin and layout questions are compressed so they take less space (143K) than if they were uncompressed in the Debconf database (794K). 3. The comments and the indentation are removed from the scripts. 4. The dependency on console-setup-fonts-udeb is removed so this package will not be included in the image. Special handling of translations Translations of keyboard layout, model, variant are handled in /usr/share/c-s-mini/c-s.config. That brings a significant impact over handling them in debconf as it doesn't allow dropping translations to save memory as it's donne in low memory situations. Now the translations are in /usr/share/console-setup-mini/kbdnames.gz. If you remove this file or filter the information inside it (simple grep can be used), console-setup-udeb will still work properly. Offering keymaps for all architectures -- In expert mode, keymap models for all arches are proposed. I removed this question from the udeb. The regular package should not ask it if its installation comes after the udeb. Too many choices The c-s udeb proposes everything proposed by xkeyboard-config. That makes way too many possible combinations, therefore confusing choices and a size impact (because of i18n). No changes here. I feel that reducing the number of supported layouts is not right. Wording problems for some choices - Do we want to fork the translations in xkeyboard-config? Useless questions - Some questions were not formerly asked and seem generally useless *even in expert mode*. Questioned templates: altgr, compose, encoding, charset, console font, font size, virtual consoles I removed the questions about the encoding and the virtual consoles. For the rest it is not clear that they are useless in expert mode. For example from a past discussion in this mailinglist I concluded that some people are using AltGr and others - Compose. Moreover, those who use Compose sometimes think that they are using AltGr (because they map Compose on the AltGr key) and those who use AltGr sometimes think they are using Compose. Preseeding broken - Is it really broken? I mean I didn't write the preseeding code in console-setup and I've never tested it but I can't remember reports that it is broken. You can set the following templates: console-setup/modelcode - pc105 console-setup/layoutcode - fr,gb console-setup/variantcode - nodeadkeys,dvorak console-setup/optionscode - grp:ctrl_shift_toggle,compose:rwin,lv3:ralt_switch After this, Control+Shift combination will toggle between French keyboard without dead keys and British Dvorak keyboard. The right Logo-key will be Compose-key and the right Alt-key will be AltGr. That issue seems related to not able to configure the keyboard twice. I hope this problem is solved. Allow skipping keyboard configuration - There's a regression here. I think this can be useful in only one case - when console-setup (and console-data) do not support the keyboard model OR when console-setup is unable to detect automatically the keyboard model. In the new console-setup if you want to leave the console keyboard unchanged you can use the following preseeding: console-setup/modelcode - unknown Skip configuration for Serial and UML installs --- Another regression. If necessary you can use the same preseeding console-setup/modelcode - unknown But I think it is preferable to configure the keyboard of the host even if its keyboard is not used during the installation. Compose key does not work in D-I already reported: find bug number If this is so, then the next logical question is: does Compose work with console-data? Anton Zinoviev -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Mon, Jun 29, 2009 at 04:55:28PM +0300, Anton Zinoviev wrote: But I think it is preferable to configure the keyboard of the host even if its keyboard is not used during the installation. I wrote this about installs on serial console, not about UML installs. Anton Zinoviev -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Monday 29 June 2009, Anton Zinoviev wrote: On Mon, Jun 29, 2009 at 04:55:28PM +0300, Anton Zinoviev wrote: But I think it is preferable to configure the keyboard of the host even if its keyboard is not used during the installation. I wrote this about installs on serial console, not about UML installs. In a lot of cases the system does not *have* a keyboard... I have a completely headless HPPA box. And we do need c-s for HPPA because there are also HPPA systems that _do_ have keyboards. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Kenshi Muto: i386/amd64 d-i images for Lenny, kernel version 2.6.30 + firmware
For those who would like to use Linux kernel 2.6.30 for installing Debian Lenny, [here it is][2]. Enjoy. Very interesting. I added support to a number of new ARM devices to squeeze and it would be great to provide a 2.6.30 based lenny installer. Can you describe what's needed to create such 2.6.30 based lenny images? More generally, it seems to me that it would be great to offer such images on backports.org or similar. I'm copying debian-boot in the hope to find out how much interest there would be in this. -- Martin Michlmayr http://www.cyrius.com/ -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Mon, Jun 29, 2009 at 06:03:21PM +0200, Frans Pop wrote: On Monday 29 June 2009, Anton Zinoviev wrote: On Mon, Jun 29, 2009 at 04:55:28PM +0300, Anton Zinoviev wrote: But I think it is preferable to configure the keyboard of the host even if its keyboard is not used during the installation. I wrote this about installs on serial console, not about UML installs. In a lot of cases the system does not *have* a keyboard... I have a completely headless HPPA box. And we do need c-s for HPPA because there are also HPPA systems that _do_ have keyboards. Should I add one additional question or preseeding is enough? Anton Zinoviev -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Monday 29 June 2009, Anton Zinoviev wrote: On Mon, Jun 29, 2009 at 06:03:21PM +0200, Frans Pop wrote: On Monday 29 June 2009, Anton Zinoviev wrote: On Mon, Jun 29, 2009 at 04:55:28PM +0300, Anton Zinoviev wrote: But I think it is preferable to configure the keyboard of the host even if its keyboard is not used during the installation. I wrote this about installs on serial console, not about UML installs. In a lot of cases the system does not *have* a keyboard... I have a completely headless HPPA box. And we do need c-s for HPPA because there are also HPPA systems that _do_ have keyboards. Should I add one additional question or preseeding is enough? No, you need to implement automatic detection for this case. Please see kbd-chooser.c. It's not all that hard to find the code, nor should it be all that hard to implement it in C. See also the 90console script in finish-install. It would also be a good idea for you to run kbd-chooser in expert mode some time so you can see how the no keyboard question is implemented. It's not a separate question... -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Monday 29 June 2009, Frans Pop wrote: On Monday 29 June 2009, Anton Zinoviev wrote: On Mon, Jun 29, 2009 at 06:03:21PM +0200, Frans Pop wrote: On Monday 29 June 2009, Anton Zinoviev wrote: On Mon, Jun 29, 2009 at 04:55:28PM +0300, Anton Zinoviev wrote: But I think it is preferable to configure the keyboard of the host even if its keyboard is not used during the installation. I wrote this about installs on serial console, not about UML installs. In a lot of cases the system does not *have* a keyboard... I have a completely headless HPPA box. And we do need c-s for HPPA because there are also HPPA systems that _do_ have keyboards. Should I add one additional question or preseeding is enough? No, you need to implement automatic detection for this case. Better: c-s needs to support automatic detection for this case. Please see kbd-chooser.c. It's not all that hard to find the code, nor should it be all that hard to implement it in C. See also the 90console script in finish-install. It would also be a good idea for you to run kbd-chooser in expert mode some time so you can see how the no keyboard question is implemented. It's not a separate question... Not that c-s needs to implement it in exactly the same way, but it should give you a better understanding of what exactly the current functionality is that c-s is aiming to replace. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Mon, Jun 29, 2009 at 07:32:39PM +0200, Frans Pop wrote: Should I add one additional question or preseeding is enough? No, you need to implement automatic detection for this case. Please see kbd-chooser.c. It's not all that hard to find the code, nor should it be all that hard to implement it in C. See also the 90console script in finish-install. I didn't known that such detection is possible. No C is required in order to detect the presence of a physical keyboard - the following shell code does the trick: tests=.. if grep $tests /proc/bus/input/devices /proc/bus/usb/devices /dev/null; then # there is a keyboard else # there is no keyboard fi The detection of the serial console requires an ioctl but I think console-setup-udeb doesn't need to know whether the console is serial or not. I have some questions: 1. Can console-setup-udeb expect that /proc is mounted? Is there some dependency about this? 2. Can console-setup-udeb expect that /proc/bus/usb is mounted? Is there some dependency about this? 3. If the answer of 1 or 2 is no, then what can be done? 4. Are there architectures where kernel =2.4 is used? The detection doesn't work for such kernels. 5. What about the following scenario: the system is installed without a keyboard, console-setup-udeb leaves the keyboard unconfigured and then the user attaches a keyboard. Anton Zinoviev -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Monday 29 June 2009, Anton Zinoviev wrote: I have some questions: 1. Can console-setup-udeb expect that /proc is mounted? Is there some dependency about this? /proc is guaranteed to be mounted in D-I. 2. Can console-setup-udeb expect that /proc/bus/usb is mounted? Is there some dependency about this? For 2.6.29 and 2.6.30 you can't depend on it being mounted. See usb-kbd.c in kbd-chooser; again the code is fairly simple (I hacked on it ages ago so it must be as I knew even less C then than I know now). But note that /proc/bus/usb (or rather usbfs) is going away in 2.6.31! See #534412, it's very relevant. 3. If the answer of 1 or 2 is no, then what can be done? Maybe you can just use /proc/input/devices (which _will_ be available) as I don't think the distinction between USB and PS/2 is really relevant anymore? 4. Are there architectures where kernel =2.4 is used? The detection doesn't work for such kernels. No. 2.4 is dead for Debian. You really only need to worry about Squeeze and that is 2.6.31+. 5. What about the following scenario: the system is installed without a keyboard, console-setup-udeb leaves the keyboard unconfigured and then the user attaches a keyboard. Not our problem. If the user installs without keyboard, it's up to him to configure userland for it. Just as he may have to install software for other hardware that is added after the installation. D-I's job is to install the system as we find it, which means that there is no reason to install c-s onto a headless system. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
nano migration
Hi all, Would it be acceptable to allow nano into testing finally? It has been waiting for 84 days for the maintainer to request migration to testing. BTW, has there been any progress towards automatic udeb migration? Even just an automatic ping, nano has been waiting for 10 days to -boot, -release and the maintainer might be good. -- bye, pabs http://wiki.debian.org/PaulWise signature.asc Description: This is a digitally signed message part
Re: Reformulation of criticisms about console-setup switch (was: Re: Review of console-setup wrt D-I [very long])
On Monday 29 June 2009, Frans Pop wrote: On Monday 29 June 2009, Anton Zinoviev wrote: 2. Can console-setup-udeb expect that /proc/bus/usb is mounted? Is there some dependency about this? For 2.6.29 and 2.6.30 you can't depend on it being mounted. See usb-kbd.c in kbd-chooser; again the code is fairly simple (I hacked on it ages ago so it must be as I knew even less C then than I know now). But note that /proc/bus/usb (or rather usbfs) is going away in 2.6.31! See #534412, it's very relevant. If you do want to use USB information, then the best solution is to take it directly from sysfs (better than mounting debugfs). Almost all information in /proc/bus/usb/devices can also be found in sysfs. See the usb-list script I recently added for installation-report for example code how to walk through sysfs correctly. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org