[Should be merged with #250789 ?] * Jan Kuznik [2004-06-07 22:30:38+0200] > Base config should honour locale settings. The console should be configured > to display the codepage requested by the language selected else the > fresh system displays mangled characters in many programs notably 'dselect'. Indeed, base-config honors the locale setting. By "fresh system", did you mean the newly installed system, that is, the situation after all the second-stage installation finished? If this is the case, please note that there are two bugs filed which report the post-installation console problems: #250376 and #250789. Bug #250376 has just been closed with the new release of languagechooser and the other is still waiting for a proper solution. In the second stage, termwrap runs the base-config, and following is the console setting for your language, which the termwrap uses upon invocation: ISO-8859-2) # Load ISO-8859-2 charset mapping into console try_load_charset "$ENCODING" iso02 iso02.f16 So, there shouldn't be any problem in the second-stage installation (the base-config stage, I mean). But as reported by #250789, termwrap reverts to iso-1 on termination which puts your console in a bad state. On the other hand, you should be fine after rebooting (which IMO, shouln't be need), assuming that the fix of #250376 is applied, which is not case for the iso you used. (We must wait the new languagechooser to include in test isos.) > I suggest Debian Installer installs the package 'fonty' and configures it > accordingly if the selected language implies coding other than 8859-1. No need for fonty. It just does the same thing quoted above. But when I look at the Czech entry in languagelist, I see that the console font for your language is not 'iso02.f16' which would be set by fonty: Czech;cs_CZ;cs_CZ;cs;CZ;cs_CZ:cs:en_GB:en;kbd=lat2-sun16(iso02) So I suggest to apply the attached patch for Czech, which will do the same setting as fonty . By the way, Christian, what about to substitute all those 'lat*-sun16' fonts with the corresponding 'iso*-f16' fonts (just like the 'iso08.f16' for Hebrew). I can prepare a cumulative patch which adresses all the related languages. -- roktas
--- languagelist.orig 2004-06-08 00:59:42.000000000 +0300 +++ languagelist 2004-06-08 01:00:11.000000000 +0300 @@ -52,7 +52,7 @@ Chinese (Simplified);zh;zh_CN;zh;;zh_CN:zh:en_US:en Chinese (Traditional);zh;zh_TW;zh;;zh_TW:zh:en_US:en #X Croatian;hr_HR;hr;HR;;kbd=lat2-sun16(iso02) -Czech;cs_CZ;cs_CZ;cs;CZ;cs_CZ:cs:en_GB:en;kbd=lat2-sun16(iso02) +Czech;cs_CZ;cs_CZ;cs;CZ;cs_CZ:cs:en_GB:en;kbd=iso02.f16(iso02) Danish;da_DK;da_DK;da;DK;da_DK:da:en_GB:en;kbd=lat0-sun16(iso15) Dutch;nl;nl_NL;nl;;nl:en_GB:en;kbd=lat0-sun16(iso15) English;en;en_US;en;;en_GB:en_US:en;kbd=lat0-sun16(iso15)
Attachment:
signature.asc
Description: Digital signature