Description of problem: Something is suggesting voikko-fi (the Finnish spellcheck dictionaries) How reproducible: Always Steps to Reproduce: 1. install from boot-nonfree.iso minimal system (only with the basic documentation) and configure X window 2. install windowmaker slim and termiantor (with suggestions) 3. install kdiff3 read that something is pulling voikko-fi
Created attachment 1992 [details] rpm -qa
Created attachment 1993 [details] urpmi kdiff3
it's better to debug with urpmi --debug or the report.bug
chosen voikko-fi-1.10-1.mga2.x86_64 for enchant-dictionary selecting voikko-fi-1.10-1.mga2.x86_64 requiring locales-fi for voikko-fi-1.10-1.mga2.x86_64 selecting locales-fi-2.14.1-3.mga2.x86_64
There are other packages providing enchant-dictionary, so selecting 'voikko-fi' automatically seems wrong here. "voikko-fi" does provide enchant-dictionary == 3 while the other ones provide enchant-dictionary == 2 (it was needed when there were other less preferred Finnish dictionaries, like aspell and myspell, so that voikko-fi would be preferred), but I think urpmi should still ask which one to install (and prefer the install the dictionary for installed locale), not sure what is wrong here (maybe the fact it is a suggestion instead of a require affects urpmi somehow?).
CC: (none) => anssi.hannula
How to do it, that spell checking dictionaries will be selected based on the installed locales. I would make it a release-blocker...
It should already be the case, seems like an urpmi bug to me (unless there is something else I'm not seeing).
We have got this relation: each hunspell-XY is requireing corresponding locales-XY http://svnweb.mageia.org/packages/cauldron/meta-task/current/SOURCES/rpmsrate-raw?view=markup 153 CAT_SPELLCHECK 154 5 hunspell-bg hunspell-ca 155 4 voikko-fi http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/distrib/1/x86_64/media/media_info/rpmsrate CAT_SPELLCHECK 5 myspell-af_ZA myspell-am_AM myspell-ar_AR myspell-az_AZ myspell-bg_BG myspell-bn_BN myspell-br myspell-ca_ES myspell-cs_CZ myspell-csb_CSB myspell-cy_GB myspell-da_DK myspell-de_AT myspell-de_CH myspell-de_DE myspell-el_GR myspell-en_AU myspell-en_CA myspell-en_GB myspell-en_NZ myspell-en_US myspell-en_ZA myspell-eo_EO myspell-es_ES myspell-es_MX myspell-et_EE myspell-eu_ES myspell-fa_FA myspell-fa_IR myspell-fi_FI myspell-fo_FO myspell-fr_BE myspell-fr_FR myspell-fur_IT myspell-fy_NL myspell-ga_IE myspell-gd_GB myspell-gl_ES myspell-gsc_FR myspell-he_IL myspell-hi_IN myspell-hr_HR myspell-hu_HU myspell-hy_AM myspell-id_ID myspell-is_IS myspell-it_IT myspell-km_KH myspell-ku_TR myspell-lt_LT myspell-lv_LV myspell-mg_MG myspell-mi_NZ myspell-mn_MN myspell-mr_IN myspell-ms_MY myspell-nb_NO myspell-ne_NP myspell-nl_NL myspell-nn_NO myspell-nr_ZA myspell-ns_ZA myspell-oc_FR myspell-or_OR myspell-pa_PA myspell-pl_PL myspell-pt_BR myspell-pt_PT myspell-ro_RO myspell-ru_RU myspell-rw_RW myspell-sk_SK myspell-sl_SI myspell-ss_ZA myspell-st_ZA myspell-sv_SE myspell-sw_KE myspell-sw_TZ myspell-ta_TA myspell-th_TH myspell-tl_PH myspell-tn_ZA myspell-ts_ZA myspell-uk_UA myspell-uz_UZ myspell-ve_ZA myspell-vi_VI myspell-xh_ZA myspell-zu_ZA 4 aspell-af aspell-am aspell-az aspell-be aspell-bg aspell-bn aspell-br aspell-ca aspell-cs aspell-cy aspell-da aspell-de aspell-el aspell-en aspell-eo aspell-es aspell-et aspell-fa aspell-fi aspell-fo aspell-fr aspell-ga aspell-gd aspell-gl aspell-gv aspell-he aspell-hi aspell-hr aspell-hsb aspell-hu aspell-id aspell-is aspell-it aspell-ku aspell-lt aspell-lv aspell-mg aspell-mi aspell-mk aspell-mn aspell-mr aspell-ms aspell-mt aspell-nb aspell-nds aspell-nl aspell-nn aspell-pa aspell-pl aspell-pt aspell-ro aspell-ru aspell-rw aspell-sc aspell-sk aspell-sl aspell-sv aspell-sw aspell-ta aspell-tl aspell-tr aspell-uk aspell-uz aspell-vi aspell-wa aspell-yi aspell-zu 4 voikko-fi (In reply to comment #7) > It should already be the case, seems like an urpmi bug to me (unless there is > something else I'm not seeing). How it should work?
Urpmi checks if a requires can be satisfied by multiple packages, and if some of the packages require uninstalled locales, those are ignored.
(In reply to comment #9) > Urpmi checks if a requires can be satisfied by multiple packages, and if some > of the packages require uninstalled locales, those are ignored. And what in case of multiple locales installed? locales-en locales-eo locales-pl will it install only one dictionary or them, or three? I will mark this a release blocker as this is annoying bug.
Priority: Normal => release_blocker
It asks the user which one to install.
CC: (none) => thierry.vignaudSource RPM: (none) => urpmi ?
Summary: Something is suggesting voikko-fi (the Finnish spellcheck dictionaries) => Urpmi is not asking which enchant-dictionary mark to install
Lowering to 'normal' priority, as the problem can be fixed after the release.
Priority: release_blocker => NormalCC: (none) => guillomovitch
Assigning to me, I'll have time to take a look in a few weeks.
Status: NEW => ASSIGNEDAssignee: bugsquad => anssi.hannula
Hi, This bug was filed against cauldron, but we do not have cauldron at the moment. Please report whether this bug is still valid for Mageia 2. Thanks :) Cheers, marja
Keywords: (none) => NEEDINFO
(In reply to comment #5) voikko-fi got preferred b/c it has a better arch score from is_arch_compat() as all the others are noarch whereas it's x86_64 or i586
See URPM::Resolve::_find_required_package__sort()
Source RPM: urpmi ? => voikko-fi
BTW, Anssi: there's no reason for the voikko-fi subpackage not to be noarch. Only voikko-fi-tools needs to be arched
And indeed making voikko-fi noarch fixed it.
Status: ASSIGNED => RESOLVEDHardware: i586 => AllResolution: (none) => FIXED
voikko-fi is actually endianness specific, but since all our architectures are little-endian, I guess it's an acceptable workaround.