FS#15680 - why do we generate locales on the target system twice?
Attached to Project:
Release Engineering
Opened by Dieter Plaetinck (Dieter_be) - Saturday, 25 July 2009, 11:02 GMT
Last edited by Gerardo Exequiel Pozzi (djgera) - Monday, 26 November 2012, 04:45 GMT
Opened by Dieter Plaetinck (Dieter_be) - Saturday, 25 July 2009, 11:02 GMT
Last edited by Gerardo Exequiel Pozzi (djgera) - Monday, 26 November 2012, 04:45 GMT
|
Details
we run "worker_auto_locale" after install_packages, and we
also run "worker_locales" after configure_system.
worker_auto_locale calls target_configure_initial_locale worker_locales calls target_locale-gen target_configure_initial_locale parses rc.conf and uncomments locales in ${var_TARGET_DIR}/etc/locale.gen, after which it calls target_locale-gen target_locale-gen chroots in the target and runs locale-gen. is it needed that we do this twice? the only things that happen between worker_auto_locale and worker_locales is: - keymap/font updating in ${var_TARGET_DIR}/etc/rc.conf - configure_system step (you know, the interactive thing for editing files, setting root pass etc) - generating mkinitcpio |
This task depends upon
Closed by Gerardo Exequiel Pozzi (djgera)
Monday, 26 November 2012, 04:45 GMT
Reason for closing: Deferred
Monday, 26 November 2012, 04:45 GMT
Reason for closing: Deferred
Otherwise: My opinion in this (and double mkinitcpio) is: So we could be sure that a proper locale file/setting is builded.
The 1,2 minutes this took longer - who cares ?
so it's fairly trivial to implement an md5 check, which is exactly what i will do :)