Re: [D-I] Please test Debian Installer with console-setup (4th version)
(Please keep both lists CC'ed) Quoting Yuri Kozlov (yu...@komyakino.ru): On Sat, 11 Jul 2009 14:20:49 +0200 Christian Perrier bubu...@debian.org wrote: As Anton Zinoviev made more changes to console-setup, I reupdated my builds with a 4th version that is using console-setup 1.44: http://people.debian.org/~bubulle/d-i/mini-cs-i386.iso http://people.debian.org/~bubulle/d-i/gtk-mini-cs-i386.iso (QEMU) In Russian: console-setup asks a questions twise -- first at beginning installation and second when installation of 'console-setup' package happened. Yes, this is something that should really be avoided. Anton, any idea why this happens? Not all Russia keyboad layout names is translated: Russia - Legacy; Russia - Ossetian, legacy; Russia - Chuvash; Russia - Udmurt; Russia - Chuvash Latin; Russia - Komi; Russia - Yakut; Russia - Kalmyk; I can not found how they may be translated. This comes from xkeyboard-config. Translations are handled upstream in the Translate Project. If I reset selection all tasks in tasksel I get installed package wfinnish - small finnish dictionary Why? I think this is an unrelated bug. After reboobing: Cyrrilic letters is visible. Keyboard layout switch (I select Winlogo) is not working. What happens if you run dpkg-reconfigure console-setup on the installed system? signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (4th version)
On Sun, Jul 12, 2009 at 10:00:48PM +0200, Christian Perrier wrote: Quoting Yuri Kozlov (yu...@komyakino.ru): (QEMU) In Russian: console-setup asks a questions twise -- first at beginning installation and second when installation of 'console-setup' package happened. Yes, this is something that should really be avoided. Anton, any idea why this happens? No, but two hypotheses: 1. d-i doesn't set the 'seen' flag of the questions. AFAIK console-setup-udeb is the only udeb that asks questions on behalf of a regular package. 2. The udeb didn't ask the 'origin' question, but the regular package is older version so it asks it and triggers the layout question too. After reboobing: Cyrrilic letters is visible. Keyboard layout switch (I select Winlogo) is not working. If console-setup works improperly, please, send the contents of the file /etc/default/console-setup in your reports. All testers, please do the same! Anton Zinoviev -- 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 (3rd version)
Anton Zinoviev, le Fri 10 Jul 2009 17:42:46 +0300, a écrit : On Fri, Jul 10, 2009 at 07:19:43AM +0200, Christian Perrier wrote: That means having a specific way to ask for layouts *in a single question* mor eor less combining the layout/variant (and maybe model) questions. This is exactly what the upstream of xkeyboard-config wants to achieve. Only one question! Not one question that includes all layouts around the world but one question customized for your teritory. Right, but I do agree that in the case of FR for instance, proposing all of nodeadkeys, sundeadkeys, oss, oss_latin9, oss_nodeadkeys, oss_sundeadkeys, latin9, latin9_nodeadkeyts, latin9_sundeadkeys, bepo, bepo_latin9, dvorak, mac, bre, oci and geo variants is insane. Just proposing the basic fr layout and the us layout should be fine. Now the questions would rather be: should the selection be done by the d-i team or (probably better) by upstream xkeyboard-config. I'll see with upstream what they think about it. Samuel -- 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 (3rd version)
Quoting Anton Zinoviev (an...@lml.bas.bg): I have not seen any single Georgian keyboard in my entire computing life in France. Actually fr(geo) is not for Georgian keyboards. It is for typing Georgian on a French keyboard. My point There is no such physical keyboard. Apparently, from what you mention, various keyboard layouts are added, more or less randomly, because someone once worked on one. This is about the same crazyness we had with console layouts in the good old days, where everybody was inventing his|her own layout (fr-latin1, fr-latin9, fr-latin0). If upstream xkb has such policy, I'm afraid it is entirely entering a neverending nightmare of piling up layouts and variants for each and every crazy ideas in the world. Anyway, this is getting off-topic. I'm not in position to change XKB policy, nor do I want it. Here, I'm interested in the consequences we have in D-I layouts, but with time this filtering will become so distant from what the upstream is providing us that it will become a nightmare to support it. I see absolutely no reason for this. I'm proposing to just make that choice each time a language is added to D-I. We're doing this for years (actually since D-I exists). Console-data was maintained by you, i.e. by a member of d-i team. My concern is that Xkeyboard-config changes more rapidly than console-data. As said, the point is not changing xkeyboard-config. The point is in still keeping the idea of the same keyboard management system in D-I and, later, at the system's console, than the one in Xwhile keeping this manageable and scalable for D-I. BTW, it seems to me that there is no other file in XKB with such a diversity of variants as the French file. :) I don't know whether all The German one is also quite interestingwhich is not a surprise: we find there the very same mess we have in console keymaps, inherited from the days where people where thinking it is good to invent their own keyboard layout and publish it everywhere because it was working on their own kitchen sink...:-) BTW, this is what I personnally think about those crazy Dvorak keymaps but don't tell this to the Dvorak and Bépo keymaps addicts..:-) these variants are realy used or not, but if they are not, one method to deal with this mess is to mark some of the variants in the file with the keyword hidden. In this way these variants will be still supported but the configuration programs will not present them to the users. If there is agreement between the French users, the upstream can be convinced to make the changes. I'm not sure that belongs to upstream, again. We should refocus on our point, here: D-I. I'm not interested in changing things upstream, or even changing the way you deal with them in the standard console-setup package. You're certainly handling things the right way. The point is the udeb. I'm currently convinced and still need to be unconvinced that we have to shrink the number of proposed layouts to a list with only 1 or 2 layouts per language, plus some country-specific cases (e.g. the Brazilian layouts). That means having a specific way to ask for layouts *in a single question* mor eor less combining the layout/variant (and maybe model) questions. This only in D-I. Probably with a specific console-setup-udeb.postinst script instead of using the big cs-.postinst and c-s.config scripts. I see this as the only way to preserve one key feature of D-I: keeping things simple for users and minimizing the number of asked questions. And, also, preserve the rationale where size matters. signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
On Fri, Jul 10, 2009 at 07:19:43AM +0200, Christian Perrier wrote: That means having a specific way to ask for layouts *in a single question* mor eor less combining the layout/variant (and maybe model) questions. This is exactly what the upstream of xkeyboard-config wants to achieve. Only one question! Not one question that includes all layouts around the world but one question customized for your teritory. I see this as the only way to preserve one key feature of D-I: keeping things simple for users and minimizing the number of asked questions. And, also, preserve the rationale where size matters. Only one question provided the US QWERTY layout is added to the list of France and the other countries that need it. ;-) No, I am not joking. :-) Yes, there are problems in the current state of xkeyboard-config but I think the proper way to deal with these problems is to report them to the upstream rather than to do something Debian-specific. Anton Zinoviev -- 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 (3rd version)
On Wednesday 08 July 2009, Christian Perrier wrote: A *third version* of test D-I images with console-setup replacing kbdchooser has been built. There is something weird in this image (not related to c-s). Please compare the two images below: - http://people.debian.org/~fjp/tmp/d-i/multiselect_joeyh.png - http://people.debian.org/~fjp/tmp/d-i/multiselect_bubulle.png In Christan's image the cursor (current line indicator) is wrong. The images were built the same day. I checked the dpkg status files in the image, but they show no relevant differences. It still could be libs that get reduced and thus don't have a udeb, but I checked the most likely candidates and did not see any recent uploads. Suggest people watch out for this as it could still be a regression somewhere, especially if it also turns up in Joey's builds. -- 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 (3rd version)
Quoting Frans Pop (elen...@planet.nl): Suggest people watch out for this as it could still be a regression somewhere, especially if it also turns up in Joey's builds. I have to mention that my build environment is not necessarily guaranteed as clean. I don't use my own machine anymore to build packages (I only use pbuilder/cowbuilder on my home server, most of the time) and it's quite some time since I built test images for some D-I features. In that specific case, it of course doesn't matter that much but what you noticed is really weird, indeed... signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
[not subscribed, please cc on replies] On Fri, Jul 10, 2009 at 07:19:43 +0200, Christian Perrier wrote: I'm not sure that belongs to upstream, again. We should refocus on our point, here: D-I. I'm not interested in changing things upstream, or even changing the way you deal with them in the standard console-setup package. You're certainly handling things the right way. I think this is the absolute wrong way to go, fwiw. Making things significantly different in the installer means more maintenance overhead, instead of improving things for everyone and sharing the maintenance burden, if this was happening upstream. The way to present this information in c-s can be improved, some xk-c strings might need to be changed to make them easier to understand, some variants might need to be hidden because they're mostly unused, etc, and sure, that's a lot of work, but there's no reason this should be specific to d-i IMO. Cheers, Julien -- 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 (3rd version)
Quoting Julien Cristau (jcris...@debian.org): I'm not sure that belongs to upstream, again. We should refocus on our point, here: D-I. I'm not interested in changing things upstream, or even changing the way you deal with them in the standard console-setup package. You're certainly handling things the right way. I think this is the absolute wrong way to go, fwiw. Making things significantly different in the installer means more maintenance overhead, instead of improving things for everyone and sharing the There seem to be some misunderstanding here. I don't necessarily want to make things different. The point is trying to keep the use of c-s by D-I sustainable, particulrly in terms of size and memory impact. The very big number of variants brings in a lot of stringsand, potentially, a lot of translated material. That will mathematically make the size impact of D-I grow up exponentially over time. That is a concern. Of course, efforts were made to minimize this and this is really welcomed, for sure. But I still need to be convinced that this will be enough. Again, my proposal is not necessarily changing the way c-s worksbut finding a way to have it present much much less variants to users so that it's less confusing for them (usability reason) and has less size impact on D-I (technical reason). I am not specifically favouring a given option, here. I imagined a way to do this (that involves manual maintenance, which I thinks is sutainable) but that may not be the only option. Anyway, I haven't proposed any implementation and I don't know if I'll be able to do so..:). Moreover, I haven't described very precisely what I imagine. In short, we're basically shaking ideas here... signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
Quoting Luca Favatella (slacky...@gmail.com): I tested mini-cs-i386.iso in italian on qemu. One of the first steps in d-i was this (look at the comments with --): --8 (1) Please select the layout matching the keyboard for this machine. -- english string Keyboard layout: -- english string Same for French. Weird, given that these strings are translated. Apparently, for some reason, the console-setup/variant template is not translated at all, except its choices (which do not follow the standard debconf way) Italia-- default, chosen by me Italia - Georgiana -- I never heard about this... Italia - Macintosh -- perhaps this is for Intel Apple Italia - Tasti muti esclusi -- (translated: Italy - except mute buttons) what is this?!? No dead keys. Dead keys are keys that do not produce a character immediately but combine with the next typed key to produce a character. Example on the French keymap: ^ right of the P key. If you hit hit, you get nothing, but if you hit an a after, you get an â. Apparently, though, these No dead keys and Georgian variants appear for all layouts! I also get a Georgian variant for the France layout. Sounds like a bug somewhere. After downloading something (perhaps the base system?), this popped up: --8 (2) [!!] Configurazione in corso di console-setup -- translated: Running translation of console-setup Origine della tastiera: -- translated: Keyboard origin: ... -- a lot of choices Yeah...:-( It means that console-setup is reconfigured on the base system without keeping choices picked in the udeb. That's a bug. At the _first_ boot, before the Setting the system clock. message, I got --8 (3) WARNING: undefined kernel key code for 153 WARNING: undefined kernel key code for 160 WARNING: undefined kernel key code for 162 ... - other messages with other numbers WARNING: undefined kernel key code for 245 I'm not sure about this one At the console (at every boot), all keys of the keyboard are working, except for characters ì and ò, that are printed as rhombus (I don't know if I chose the wrong keyboard layout for qemu...). Where are these keys located on a standard italian keyboard layout? signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
Quoting Luca Favatella (slacky...@gmail.com): Italia - Georgiana -- I never heard about this... Well, apparently, there *is* a geo variant for Italian layout as well as for French. Don't ask me why I'm not really aware of such a huge need for typing Georgian when living in France or Italy whatever good feeling we might have with the Georgian people..:-) All this is a good illustration of reasons that should make us shrink the number of choices in c-s. We don't want to carry a translation for each and every such exotic keymap in all languages signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
On 09/07/2009, Christian Perrier bubu...@debian.org wrote: Quoting Luca Favatella (slacky...@gmail.com): [...] At the console (at every boot), all keys of the keyboard are working, except for characters ì and ò, that are printed as rhombus (I don't know if I chose the wrong keyboard layout for qemu...). Where are these keys located on a standard italian keyboard layout? Please see this: http://en.wikipedia.org/wiki/File:KB_Italian.svg -- 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 (3rd version)
Christian Perrier, le Thu 09 Jul 2009 06:47:00 +0200, a écrit : Apparently, though, these No dead keys and Georgian variants appear for all layouts! I also get a Georgian variant for the France layout. Sounds like a bug somewhere. No, there really are Georgian variants for fr and it, and a lot of layouts have a nodeadkeys variant. At the _first_ boot, before the Setting the system clock. message, I got --8 (3) WARNING: undefined kernel key code for 153 WARNING: undefined kernel key code for 160 WARNING: undefined kernel key code for 162 ... - other messages with other numbers WARNING: undefined kernel key code for 245 I'm not sure about this one That's odd, it's supposed to be fixed in the current release. Samuel -- 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 (3rd version)
On Thu, Jul 09, 2009 at 06:57:34AM +0200, Christian Perrier wrote: Quoting Luca Favatella (slacky...@gmail.com): Italia - Georgiana -- I never heard about this... Well, apparently, there *is* a geo variant for Italian layout as well as for French. Don't ask me why Actualy the answer is simple. The keymaps of XKB are organized in such a way that the French file should support all people living in France - no matter how few or how many are these people. You can expect that in the future the lists of the layouts for France and the layouts for Italy will become even bigger. Don't think of these lists as lists of Italian of French keyboards - they are not this. These lists are sort of filters that allow the keyboard configuration programs to present the user with a much shorter list than the list of all existing layouts - a list that includes only the layouts that are used in the teritory of the user. When some layout is used in many teritories then this layout will be includes in the lists for all these teritories. Now consider for example the Georgian layout based on AZERTY - this layout can be used only on teritories where the keyboards are imprinted according to the AZERTY layout. So fr(geo) is a layout for France and it has nothing to do with Georgia. The layouts in XKB are not language-based and there are many good reasons why this is so. You can organize some filtering for useful layouts, but with time this filtering will become so distant from what the upstream is providing us that it will become a nightmare to support it. Anton Zinoviev -- 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 (3rd version)
On Thu, Jul 09, 2009 at 06:47:00AM +0200, Christian Perrier wrote: One of the first steps in d-i was this (look at the comments with --): --8 (1) Please select the layout matching the keyboard for this machine. -- english string Keyboard layout: -- english string Same for French. Weird, given that these strings are translated. Actualy they are not. I changed the string of console-setup/varian the last moment. The previous text was There is more than one possible keyboard layout with the origin you selected. Please select the layout matching the keyboard for this machine. and now it is simply Please select the layout matching the keyboard for this machine. After downloading something (perhaps the base system?), this popped up: --8 (2) [!!] Configurazione in corso di console-setup -- translated: Running translation of console-setup Origine della tastiera: -- translated: Keyboard origin: ... -- a lot of choices Yeah...:-( It means that console-setup is reconfigured on the base system without keeping choices picked in the udeb. That's a bug. At the _first_ boot, before the Setting the system clock. message, I got --8 (3) WARNING: undefined kernel key code for 153 WARNING: undefined kernel key code for 160 WARNING: undefined kernel key code for 162 ... - other messages with other numbers WARNING: undefined kernel key code for 245 I'm not sure about this one These both are not really bugs. They are consequences from tesging having an old version of console-setup. At the console (at every boot), all keys of the keyboard are working, except for characters ì and ò, that are printed as rhombus (I don't know if I chose the wrong keyboard layout for qemu...). Where are these keys located on a standard italian keyboard layout? Ahh! I thought that 1.40 was at last a good version of console-setup but there were too many changes so new bugs were inevitable... :) There is a bug in the preseeding code of the udeb that causes /etc/default/console-setup to include the following setting: CODESET=Arabic Anton Zinoviev -- 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 (3rd version)
Quoting Anton Zinoviev (an...@lml.bas.bg): Actualy the answer is simple. The keymaps of XKB are organized in such a way that the French file should support all people living in France - no matter how few or how many are these people. You can expect that in This is very surprising, really. There are probably, living in France, people from each and every country in the world...or nearly so. So, should we then carry hundreds of keymaps because there are Bulgarian people living here? Either these people use the crappy french keyboard we put in front of our computers to trick out foreigners...or they buy a US keyboard at the next computer store...or they bring a Bulgarian keyboard from home. In the latter case, they will happily choose Other, then Bulgaria, then the right variant. I have not seen any single Georgian keyboard in my entire computing life in France. I may understand some reasons to sort things by country. Fine. But, in such case, only list layouts representing physical keyboards (or sometimes virtualbepo and dvorak layouts come to mind) that are indeed really *used* in the said country, according to the local people. Now consider for example the Georgian layout based on AZERTY - this layout can be used only on teritories where the keyboards are imprinted according to the AZERTY layout. So fr(geo) is a layout for France and it has nothing to do with Georgia. In that case, you're telling me that there is a layout basically meant to handle Georgian language that is based on the AZERTY layout (which some people call the French layout). So, what' you're telling me is that there should be some family layouts named Azerty, Qwerty, Qwertz, etc. *that* would make sense. Putting a Georgian keymap under France does not. Really not. The layouts in XKB are not language-based and there are many good reasons why this is so. You can organize some filtering for useful And there about about as many bad reasons for doing this. One reason I find for sorting layouts by countries is because some *countries* begin to normalize the keymaps. *Here*, I agree that sorting by country could make some sense. But I don't think there is a French standard for a Georgian keyboard layout. Another reason I find is because this is easy to preseed. If ppl answered they live in France, they would expect getting keyboard layouts used in the country. So, in short, I am not against sorting by countryand anyway, I don't have much influence on this. Moreover, the recent work you did makes things quite straightforward for the standard configuration of c-s. My concern is more about we we have in D-I layouts, but with time this filtering will become so distant from what the upstream is providing us that it will become a nightmare to support it. I see absolutely no reason for this. I'm proposing to just make that choice each time a language is added to D-I. We're doing this for years (actually since D-I exists). This is not filtering, this is selecting the most common model/layout/variant combinations and propose only these ones in order to follow the guidelines that has prevailed all time long during the development of D-I: keep things simple and easy to cope with by random users. signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
On Thursday 09 July 2009, Anton Zinoviev wrote: If you want you can submit a wishlist report to allow preseeding of the list of layouts to show. console-setup/displayed_layouts = fr(basic),fr(nodeadkeys),fr(dvorak),ch(fr) Please NO. That is absolutely not what preseeding is for. Preseeding is to *automate* installations, not to tune manual installations and certainly not if the preseeding is more complex than just using the dialogs. I'll do a another test myself probably this weekend and will probably include a rough proposal for a user interface for D-I. -- 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 (3rd version)
On Thu, Jul 09, 2009 at 07:54:49PM +0200, Christian Perrier wrote: Actualy the answer is simple. The keymaps of XKB are organized in such a way that the French file should support all people living in France - no matter how few or how many are these people. You can expect that in This is very surprising, really. There are probably, living in France, people from each and every country in the world...or nearly so. So, should we then carry hundreds of keymaps because there are Bulgarian people living here? My opinion was no but recently it changed to I don't know. I suppose if xkeyboard-config is maintained in the current way, some future version of it will include AZERTY layouts for all major scripts in the world. This is not necessarily bad - a French student who is studying Greek or Russian is not going to use Greek or Russian keyboards because with such a keyboard the student will have difficulties typing French or because with AZERTY layout the student won't have to learn the positions of the letters of the foreign script. I have not seen any single Georgian keyboard in my entire computing life in France. Actually fr(geo) is not for Georgian keyboards. It is for typing Georgian on a French keyboard. So, what' you're telling me is that there should be some family layouts named Azerty, Qwerty, Qwertz, etc. *that* would make sense. Putting a Georgian keymap under France does not. Really not. OK, I agree that it makes sense. :) And about a month ago this was exactly my opinion on the matter. Now I am in the category I don't know. Another reason I find is because this is easy to preseed. If ppl answered they live in France, they would expect getting keyboard layouts used in the country. If you want you can submit a wishlist report to allow preseeding of the list of layouts to show. console-setup/displayed_layouts = fr(basic),fr(nodeadkeys),fr(dvorak),ch(fr) === Veuillez choisir celle qui correspond au clavier de cette machine. Disposition du clavier : France France - Éliminer les touches mortes France - Dvorak Suisse - Français === layouts, but with time this filtering will become so distant from what the upstream is providing us that it will become a nightmare to support it. I see absolutely no reason for this. I'm proposing to just make that choice each time a language is added to D-I. We're doing this for years (actually since D-I exists). Console-data was maintained by you, i.e. by a member of d-i team. My concern is that Xkeyboard-config changes more rapidly than console-data. BTW, it seems to me that there is no other file in XKB with such a diversity of variants as the French file. :) I don't know whether all these variants are realy used or not, but if they are not, one method to deal with this mess is to mark some of the variants in the file with the keyword hidden. In this way these variants will be still supported but the configuration programs will not present them to the users. If there is agreement between the French users, the upstream can be convinced to make the changes. Anton Zinoviev -- 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 (3rd version)
On Thu, Jul 09, 2009 at 09:36:18PM +0200, Frans Pop wrote: On Thursday 09 July 2009, Anton Zinoviev wrote: If you want you can submit a wishlist report to allow preseeding of the list of layouts to show. console-setup/displayed_layouts = fr(basic),fr(nodeadkeys),fr(dvorak),ch(fr) Please NO. That is absolutely not what preseeding is for. Preseeding is to *automate* installations, not to tune manual installations and certainly not if the preseeding is more complex than just using the dialogs. There is no harm in using preseeding to tune manual installations. Except maybe for those who write the installation manual. ;-) Nobody is obliged to use such a feature. The point is whether such a feature will be wanted by the users or not. I don't know. I'll do a another test myself probably this weekend and will probably include a rough proposal for a user interface for D-I. Unfortunately there were some bugs discovered in 1.40 and I don't know how to build an image with 1.41 to test it. But on the other hand in 1.41 I removed almost all differences between console-setup-udeb and console-setup-mini and the last package seems to work, so... Anton Zinoviev -- 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 (3rd version)
On Thursday 09 July 2009, Anton Zinoviev wrote: On Thu, Jul 09, 2009 at 09:36:18PM +0200, Frans Pop wrote: On Thursday 09 July 2009, Anton Zinoviev wrote: If you want you can submit a wishlist report to allow preseeding of the list of layouts to show. console-setup/displayed_layouts = fr(basic),fr(nodeadkeys),fr(dvorak),ch(fr) Please NO. That is absolutely not what preseeding is for. Preseeding is to *automate* installations, not to tune manual installations and certainly not if the preseeding is more complex than just using the dialogs. There is no harm in using preseeding to tune manual installations. Except maybe for those who write the installation manual. ;-) Nobody is obliged to use such a feature. The point is whether such a feature will be wanted by the users or not. I don't know. No sane user is going to type such a preseed command at the boot prompt (which would be the only realistic use case). If he's going to type anything at all at the boot prompt, it is going to be the exact keymap that he actually wants to select (and thus skip the question). Please accept from me that the feature does not make any sense. At all. -- 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 (3rd version)
On Thu, Jul 09, 2009 at 10:03:10PM +0200, Frans Pop wrote: No sane user is going to type such a preseed command at the boot prompt My understanding of what Christian wanted it this: he wanted to be able to preseed (not at the boot prompt) console-setup/layout but not console-setup/variant. As a result he would obtain an installation media for the people in France and he didn't want the Georgian keyboard to be in the list of the displayed keyboard layouts. Please accept from me that the feature does not make any sense. At all. OK, OK. Anton Zinoviev -- 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 (3rd version)
On Thursday 09 July 2009, Anton Zinoviev wrote: On Thu, Jul 09, 2009 at 10:03:10PM +0200, Frans Pop wrote: No sane user is going to type such a preseed command at the boot prompt My understanding of what Christian wanted it this: he wanted to be able to preseed (not at the boot prompt) console-setup/layout but not console-setup/variant. As a result he would obtain an installation media for the people in France and he didn't want the Georgian keyboard to be in the list of the displayed keyboard layouts. I think that what he wants (and I tend to agree) is to limit *at package build time* what keymaps are shown *during an install*. Personally I don't have any problem with c-s offering the full 200.000 keymap variants when a user reconfigures c-s on an installed system. But during installation you have to somehow limit the choices, at least what is displayed by default for a specific country selected in localechooser, to a sane reduced list. I suggest you wait for my next test results and user interface proposal. I would appreciate having a new test image with the latest uploaded version before that. BTW, did you already fix the bug in Christian's last posted image where go back in the first dialog and the dialog after other results in an error? I also saw that the default for the encoding is no longer set correctly in that image. -- 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 (2nd version)
Quoting Frans Pop (elen...@planet.nl): Christian Perrier wrote: That's probably a weird consequence of /me forgetting to explicitely add console-setup-pc-ekmap to the list of included packages..:-( Please also don't forget to drop console-keymaps-at this time. Hmm, unless I am very mistaken, console-keymaps-at was not included in that second version of the test images. signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (3rd version)
On 08/07/2009, Christian Perrier bubu...@debian.org wrote: A *third version* of test D-I images with console-setup replacing kbdchooser has been built. [...] I would appreciate if these images could be tested, particularly in various languages. I tested mini-cs-i386.iso in italian on qemu. One of the first steps in d-i was this (look at the comments with --): --8 (1) Please select the layout matching the keyboard for this machine. -- english string Keyboard layout: -- english string Italia-- default, chosen by me Italia - Georgiana -- I never heard about this... Italia - Macintosh -- perhaps this is for Intel Apple Italia - Tasti muti esclusi -- (translated: Italy - except mute buttons) what is this?!? Other -- english string --8 (1) After downloading something (perhaps the base system?), this popped up: --8 (2) [!!] Configurazione in corso di console-setup -- translated: Running translation of console-setup Origine della tastiera: -- translated: Keyboard origin: ... -- a lot of choices --8 (2) and then the same of (1) but translated in italian. It is suggested to test localized installations of a standard system, not necessarily with X, and check whether things work as expected on the installed system (particularly display at the console as well as the keymap used in the console). At the _first_ boot, before the Setting the system clock. message, I got --8 (3) WARNING: undefined kernel key code for 153 WARNING: undefined kernel key code for 160 WARNING: undefined kernel key code for 162 ... - other messages with other numbers WARNING: undefined kernel key code for 245 --8 (3) And then (always at the first boot), I got a very similar message (I didn't check if the numbers were the same) before the INIT: Entering runlevel: 2 message. At the console (at every boot), all keys of the keyboard are working, except for characters ì and ò, that are printed as rhombus (I don't know if I chose the wrong keyboard layout for qemu...). [...] If you provide feeback, please do so by answering to this mail, keeping both lists CC'ed (but *don't* CC me as I read both lists!). Done. Thanks for all your work, Luca Favatella -- 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 (2nd version)
On Mon, Jul 06, 2009 at 11:29:53PM +0200, Christian Perrier wrote: A *second version* of test D-I images with console-setup replacing kbdchooser has been built. Please do not test this image yet. There are two seriour bugs in it: 1. the config script of console-setup is not executed because I forgot that it is renamed 2. For some reason the image includes amiga keymaps instead of pc keymaps I will prepare a new version of console-setup in the next few hours. Additionaly, this new version will simplify the keymap question (as requested by d-i team). Anton Zinoviev -- 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 (2nd version)
Quoting Anton Zinoviev (an...@lml.bas.bg): Please do not test this image yet. There are two seriour bugs in it: 1. the config script of console-setup is not executed because I forgot that it is renamed OK, I'll try to rebuild the image with the package you just uploaded. Please stay tuned. 2. For some reason the image includes amiga keymaps instead of pc keymaps That's probably a weird consequence of /me forgetting to explicitely add console-setup-pc-ekmap to the list of included packages..:-( signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup (2nd version)
Christian Perrier wrote: That's probably a weird consequence of /me forgetting to explicitely add console-setup-pc-ekmap to the list of included packages..:-( Please also don't forget to drop console-keymaps-at this time. -- 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: [D-I] Please test Debian Installer with console-setup
On Fri, Jun 26, 2009 at 09:28:05PM +0300, Anton Zinoviev wrote: On Thu, Jun 18, 2009 at 02:53:42PM +0100, Colin Watson wrote: This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). Really? kbd-udeb contains setfont. What has the d-i team done to stop kbd-udeb from being present? console-setup-udeb depends on kbd-udeb so I don't see what the problem could be here. Hmm. I have forgotten this. Then I don't know why the font was not loaded on the text consoles. 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. Anton Zinoviev -- 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
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? Samuel -- 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 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. But I think that the first version of d-i with c-s should go without console-setup-fonts-udeb, so these 90k are irrelevant. Anton Zinoviev -- 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
Anton Zinoviev, le Sat 27 Jun 2009 22:12:19 +0300, a écrit : 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. € ls /bin/gzip -l -rwxr-xr-x 1 root root 63K fév 28 03:59 /bin/gzip* And it's not even a lightweight version :) Samuel -- 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 Saturday 27 June 2009, Anton Zinoviev wrote: On Fri, Jun 26, 2009 at 09:28:05PM +0300, Anton Zinoviev wrote: On Thu, Jun 18, 2009 at 02:53:42PM +0100, Colin Watson wrote: This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). Really? kbd-udeb contains setfont. What has the d-i team done to stop kbd-udeb from being present? console-setup-udeb depends on kbd-udeb so I don't see what the problem could be here. Hmm. I have forgotten this. Then I don't know why the font was not loaded on the text consoles. This was because setfont can not load compressed fonts without gzip. I changed console-setup-fonts-udeb to include uncompressed fonts. However, as it only uses 'gzip -d' that could be fixed by replacing that with 'gunzip' (which _is_ supported in busybox-udeb), right? -- 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 10:10:29PM +0200, Frans Pop wrote: However, as it only uses 'gzip -d' that could be fixed by replacing that with 'gunzip' (which _is_ supported in busybox-udeb), right? I created a script /bin/gzip with the following contents #!/bin/sh gunzip $@ and this works - setupcon and setfont load font on the consoles. Anton Zinoviev -- 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 Saturday 27 June 2009, Anton Zinoviev wrote: On Sat, Jun 27, 2009 at 10:10:29PM +0200, Frans Pop wrote: However, as it only uses 'gzip -d' that could be fixed by replacing that with 'gunzip' (which _is_ supported in busybox-udeb), right? I created a script /bin/gzip with the following contents #!/bin/sh gunzip $@ and this works - setupcon and setfont load font on the consoles. I don't think that's an acceptable solution as it suggests it's possible to also zip files in D-I, which is very simply not the case. Also, won't setupcon now not think that gzip is available and start trying to gzip keyboard files? Why not just change setupcon and setfont to use gunzip? Is there any technical reason to prefer 'gzip -d' over 'gunzip'? Please try and keep things clean! This is a kludge. And an ugly one at that. Cheers, FJP P.S. I had not seen that you had an internal function for which. -- 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 Sun, Jun 28, 2009 at 12:44:06AM +0200, Frans Pop wrote: Why not just change setupcon and setfont to use gunzip? Is there any technical reason to prefer 'gzip -d' over 'gunzip'? Please try and keep things clean! This is a kludge. And an ugly one at that. :) But ofcourse this was only a test! I am not proposing this as an actual solution. Setfont can be changed but I am not sure that this is realy needed. To remove the fonts is the simplest solution. Anton Zinoviev -- 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 Thu, Jun 18, 2009 at 02:53:42PM +0100, Colin Watson wrote: This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). Really? kbd-udeb contains setfont. What has the d-i team done to stop kbd-udeb from being present? console-setup-udeb depends on kbd-udeb so I don't see what the problem could be here. Hmm. I have forgotten this. Then I don't know why the font was not loaded on the text consoles. Anton Zinoviev -- 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 Fri, Jun 12, 2009 at 05:49:15PM +0300, Anton Zinoviev wrote: On Thu, Jun 11, 2009 at 01:35:57PM +0200, Miroslav Kure wrote: * During the instalation, we have correct font only on the first tty (the installer itself). If I switch to the shell on the second console, most of the accented letters are prited as boxes instead. After the install, all six ttys are correctly initialized with the proper font. This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). Really? kbd-udeb contains setfont. What has the d-i team done to stop kbd-udeb from being present? console-setup-udeb depends on kbd-udeb so I don't see what the problem could be here. -- 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
Quoting Anton Zinoviev (an...@lml.bas.bg): On Thu, Jun 11, 2009 at 01:35:57PM +0200, Miroslav Kure wrote: * During the instalation, we have correct font only on the first tty (the installer itself). If I switch to the shell on the second console, most of the accented letters are prited as boxes instead. After the install, all six ttys are correctly initialized with the proper font. This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). Yes, indeed. Actually, the virtual consoles in D-I have never well supported the display of non ASCII characters (I should recheck this but anyway this is something that is not *that* useful anyway..:-)) signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup
Anton Zinoviev, le Fri 12 Jun 2009 17:49:15 +0300, a écrit : the console layout used circumflex instead of caron because dead_caron is not properly supported by the kernel. That used to be the case: in k_dead() there is only a handful of diacriticals: `, ', ^, ~, and ,. But nowadays there is dead2 (key type 13, right between slock and brl), which can handle 256 diacriticals. The mapping is up to userspace through the compose table. That means that to make dead_caron working, we can make console-setup use a 'v' diacritical for instance: add compose entries like v + c - č, and make dead_caron emit K(13,'v'). Samuel -- 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 Fri, Jun 12, 2009 at 05:49:15PM +0300, Anton Zinoviev wrote: On Thu, Jun 11, 2009 at 01:35:57PM +0200, Miroslav Kure wrote: * During the instalation, we have correct font only on the first tty (the installer itself). If I switch to the shell on the second console, most of the accented letters are prited as boxes instead. After the install, all six ttys are correctly initialized with the proper font. This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). Well, to be honest, I do not see much point in supporting nonascii characters on the other consoles *during installation*. Important is that the installer screens itself work as expected. * With the previous setup, when selecting Czech keyboard, it still defaulted to English layout and only by pressing some magic key (IIRC Scroll Lock), it switched to the Czech one. Now we have Czech layout by default (this can probably be seen as an improvement by most of the users), but I found no way to switch back to the English layout (and eventualy back to Czech). I know how to do that after the installation (by editing the XKB* options in /etc/default/console-setup), but it would be nice if we were able to preset these options even for the installation. It is possible to make this, however the same setup will be used by X too and this will be unexpected by the users. On the contrary, it will be perfectly expected by the users. When you install any of the Microsoft operating systems and choose to install Czech keyboard layout, it will by default install also the English one as an option. Indeed, we had similar setup in Debian when we used XFree86 and localization-config packages: for installations in Czech it used the following setup for X: 'cs_CZ' = { LAYOUT = 'us,cz_qwerty', XKBOPTIONS = 'grp:shift_toggle,grp_led:scroll' }, The us layout came first, as it was consistent with the console - these days I would use cz first and drop the qwerty variant, but the point is the same. It is possible to preseed such configuration by giving the value of XKBLAYOUT to console-setup/layoutcode and the value of XKBVARIANT to console-setup/variantcode. Thanks, I will preseed my installations with this! -- Miroslav Kure -- 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 Thu, Jun 11, 2009 at 01:35:57PM +0200, Miroslav Kure wrote: * During the instalation, we have correct font only on the first tty (the installer itself). If I switch to the shell on the second console, most of the accented letters are prited as boxes instead. After the install, all six ttys are correctly initialized with the proper font. This is expected because the installer doesn't have setfont utility (but if the d-i team decides to put there setfont, then no change in console-setup is required in order to make all consoles configured). * With the previous setup, when selecting Czech keyboard, it still defaulted to English layout and only by pressing some magic key (IIRC Scroll Lock), it switched to the Czech one. Now we have Czech layout by default (this can probably be seen as an improvement by most of the users), but I found no way to switch back to the English layout (and eventualy back to Czech). I know how to do that after the installation (by editing the XKB* options in /etc/default/console-setup), but it would be nice if we were able to preset these options even for the installation. It is possible to make this, however the same setup will be used by X too and this will be unexpected by the users. It is possible to preseed such configuration by giving the value of XKBLAYOUT to console-setup/layoutcode and the value of XKBVARIANT to console-setup/variantcode. some letters need to be written using so called 'dead key' (for Czech it is '=' key on the English keyboards) and these do not work. E.g. 'ť' does not have its own key, so we need to write that as 'Shift+=' (this does not produce any character yet) followed by 't' - 'ť' should appear. Instead of the expected result, two-character sequence appears:'^t'. There is a difference between the layout in console-data and XKB. In the console keymap of console-data Shift+= produces dead_circumflex. In /usr/share/X11/xkb/symbols/cz (and consequently in console-setup too) Shift+= produces dead_caron. As far as I know the Chezch alphabet contains letters with caron so the X layout must be correct and the console layout used circumflex instead of caron because dead_caron is not properly supported by the kernel. If this is so, you can fill a bug report agains the kernel. I hope this bug will be easy to fix - only an addition to some table. Anton Zinoviev -- 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 Tue, Jun 02, 2009 at 08:19:03AM +0200, Christian Perrier wrote: [...] Packages have now reached the state where it is possible to test things out and provide some feedback. For this to happen, I built two netboot i386 ISO images (one[2] is console-based and another one[3] is using the graphical installer), where kbd-config has been replaced by console-setup-udeb and localechooser activates the use of console-setup on the installed system by default. I would appreciate if these images could be tested, particularly in various languages. It is suggested to test localized installations of a standard system, not necessarily with X, and check whether things work as expected on the installed system (particularly display at the console as well as the keymap used in the console). Hi Christian, I made a quick test in Czech language, here are my findings so far: * During the instalation, we have correct font only on the first tty (the installer itself). If I switch to the shell on the second console, most of the accented letters are prited as boxes instead. After the install, all six ttys are correctly initialized with the proper font. * With the previous setup, when selecting Czech keyboard, it still defaulted to English layout and only by pressing some magic key (IIRC Scroll Lock), it switched to the Czech one. Now we have Czech layout by default (this can probably be seen as an improvement by most of the users), but I found no way to switch back to the English layout (and eventualy back to Czech). I know how to do that after the installation (by editing the XKB* options in /etc/default/console-setup), but it would be nice if we were able to preset these options even for the installation. * Some accented letters can't be written on the console (under X everything is fine). The accented letters that have their own key (e.g. pressing '2' will write 'ě') can be used without any problems, but some letters need to be written using so called 'dead key' (for Czech it is '=' key on the English keyboards) and these do not work. E.g. 'ť' does not have its own key, so we need to write that as 'Shift+=' (this does not produce any character yet) followed by 't' - 'ť' should appear. Instead of the expected result, two-character sequence appears:'^t'. Cheers -- Miroslav Kure -- 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
Miroslav Kure, le Thu 11 Jun 2009 13:35:57 +0200, a écrit : * Some accented letters can't be written on the console (under X everything is fine). The accented letters that have their own key (e.g. pressing '2' will write 'ě') can be used without any problems, but some letters need to be written using so called 'dead key' (for Czech it is '=' key on the English keyboards) and these do not work. E.g. 'ť' does not have its own key, so we need to write that as 'Shift+=' (this does not produce any character yet) followed by 't' - 'ť' should appear. Instead of the expected result, two-character sequence appears:'^t'. Mmm, this works for French's ê for instance, so I guess what could be missing is the composition table for non-latin1 fonts. Samuel -- 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
Samuel Thibault, le Thu 11 Jun 2009 14:24:34 +0200, a écrit : so I guess what could be missing is the composition table for non-latin1 fonts. Maybe console-setup should parse libx11's Compose file to get them. Samuel -- 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
Quoting Hideki Yamane (henr...@debian.or.jp): But they stop installation with debootstrap failure message Warning: Failure trying to run: chroot /target mount -t proc proc /proc in my environment. So I test keyboard layout with virtual console (good) but cannot test with X. I suspect this has of course nothing to do with console-setup. I'm just testing a rebuilt image right now, to see if I see this problem. You mention you wanted to test X and that things work in virtual consoles: did you actually choose the desktop install? That one failed, but an install without X didn't fail. Do I understand properly? signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup
Quoting Christian Perrier (bubu...@debian.org): Quoting Hideki Yamane (henr...@debian.or.jp): But they stop installation with debootstrap failure message Warning: Failure trying to run: chroot /target mount -t proc proc /proc in my environment. So I test keyboard layout with virtual console (good) but cannot test with X. I suspect this has of course nothing to do with console-setup. This is #532050. signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup
On Sat, 6 Jun 2009 11:54:59 +0200 Christian Perrier bubu...@debian.org wrote: I suspect this has of course nothing to do with console-setup. This is #532050. Yes, this one, thanks. console-setup with virtual console is fine, so X also is okay, I guess. -- Regards, Hideki Yamane henrich @ debian.or.jp/iijmio-mail.jp http://wiki.debian.org/HidekiYamane -- 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
Hi, On Tue, 2 Jun 2009 08:19:03 +0200 Christian Perrier bubu...@debian.org wrote: Packages have now reached the state where it is possible to test things out and provide some feedback. For this to happen, I built two netboot i386 ISO images (one[2] is console-based and another one[3] is using the graphical installer), where kbd-config has been replaced by console-setup-udeb and localechooser activates the use of console-setup on the installed system by default. Keyboard select question is different from previous one, good. But they stop installation with debootstrap failure message Warning: Failure trying to run: chroot /target mount -t proc proc /proc in my environment. So I test keyboard layout with virtual console (good) but cannot test with X. -- Regards, Hideki Yamane henrich @ debian.or.jp/iijmio-mail.jp http://wiki.debian.org/HidekiYamane -- 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
Quoting Frans Pop (elen...@planet.nl): On Thursday 04 June 2009, Samuel Thibault wrote: Frans Pop, le Thu 04 Jun 2009 00:36:55 +0200, a ?crit : So I was wondering if there are issues that people who've worked on c-s are aware of but that have not been listed. I don't think there are. You'll be surprised then :-P Expect my comments within the week. I can promise you a long mail. That will be much welcomed. To answer the question you didn't ask, it's very likely that all this has been very loosely tested. I personnally did put my free time more in working on the integration than really testing the result. Particularly because I think we are far enough from a moment where we want to have a very stable and reliable relese and I thus think the most important is to expose this new feature as widely as possible. So, you big list of concerns/issues is certainly not unexpectedand welcomed as it will give all of us a good ground to work on. signature.asc Description: Digital signature
Re: [D-I] Please test Debian Installer with console-setup
On Tuesday 02 June 2009, Christian Perrier wrote: Packages have now reached the state where it is possible to test things out and provide some feedback. For this to happen, I built two netboot i386 ISO images (one[2] is console-based and another one[3] is using the graphical installer), where kbd-config has been replaced by console-setup-udeb and localechooser activates the use of console-setup on the installed system by default. I have today tested your image and found a number of issues, various of which are IMO important or even RC. I found these issues after about 15 minutes of testing, but I expect checking them and writing them up is going to take me at least a couple of hours. I'm just wondering if any of the issues I've found are already known. So I'd appreciate if the people who've worked on console-setup up till now could list the issues or TODO items they are aware of. It would be great to have such a list as a reference point for my mail and future discussion. I'm only talking about issues related to the use of console-setup in D-I, so general issues (that also or mainly affect its use in installed systems) can be ignored. TIA, FJP signature.asc Description: This is a digitally signed message part.
Re: [D-I] Please test Debian Installer with console-setup
Frans Pop, le Wed 03 Jun 2009 19:23:01 +0200, a écrit : So I'd appreciate if the people who've worked on console-setup up till now could list the issues or TODO items they are aware of. The TODO list is on http://wiki.debian.org/DebianInstaller/ConsoleSetupSwitch Samuel -- 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 Wednesday 03 June 2009, Samuel Thibault wrote: Frans Pop, le Wed 03 Jun 2009 19:23:01 +0200, a ?crit : So I'd appreciate if the people who've worked on console-setup up till now could list the issues or TODO items they are aware of. The TODO list is on http://wiki.debian.org/DebianInstaller/ConsoleSetupSwitch Yes, I'm aware of that. Problem is that after 15 minutes of testing I have a list of 16 issues, most of which I consider important to critical, and I only recognize two of them on that wiki page. So I was wondering if there are issues that people who've worked on c-s are aware of but that have not been listed. And I haven't done anything special: only the first few steps of a regular and an expert install, a few comparisons with a current daily image and a quick look at the postinst. I've not done a code review and have not even tried looking at how things end up on the installed system. Cheers, FJP -- 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
Frans Pop, le Thu 04 Jun 2009 00:36:55 +0200, a écrit : So I was wondering if there are issues that people who've worked on c-s are aware of but that have not been listed. I don't think there are. Samuel -- 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 Thursday 04 June 2009, Samuel Thibault wrote: Frans Pop, le Thu 04 Jun 2009 00:36:55 +0200, a ?crit : So I was wondering if there are issues that people who've worked on c-s are aware of but that have not been listed. I don't think there are. You'll be surprised then :-P Expect my comments within the week. I can promise you a long mail. -- 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
Il giorno lun, 01/06/2009 alle 15.35 +0200, Christian Perrier ha scritto: One of the release goals of Debian Installer for squeeze is dropping the use of console-data keyboard mappings, to replace them by console-setup [1]. ... Please just don't pay too much attention to possibly untranslated questions at the keyboard config step during installation. That is expected as keymap-related strings, that belong to xkeyboard-config, will need a big translation effort in order to support all languages supported in D-I. That will happen later..:-) ... If you provide feeback, please do so by answering to this mail, keeping both lists CC'ed (but *don't* CC me as I read both lists!). Many thanks in advance for your help, [1] http://wiki.debian.org/DebianInstaller/ConsoleSetupSwitch [2] http://people.debian.org/~bubulle/d-i/mini-cs-i386.iso I tested the mini-cs-i386.iso into a VirtualBox virtual machine selecting, as you can imagine, Italian. Everything worked ok except the italian mirrors that did not work. I'll test the gtk iso ASAP. Bye Stefano -- Stefano Canepa aka sc: s...@linux.it - http://www.stefanocanepa.it Three great virtues of a programmer: laziness, impatience and hubris. Le tre grandi virtù di un programmatore: pigrizia, impazienza e arroganza. (Larry Wall) signature.asc Description: Questa è una parte del messaggio firmata digitalmente