Hi Ludo’ ! On 12/23/20 5:29 PM, Ludovic Courtès wrote: > Hi! > > paul skribis: > >> From 17a476dd234fa4a128eb90ccc8ac9a565045bfa1 Mon Sep 17 00:00:00 2001 >> From: Giacomo Leidi >> Date: Thu, 5 Nov 2020 02:04:20 +0100 >> Subject: [PATCH] gnu: Add hunspell-dict-it-it. >> >> * gnu/packages/hunspell.scm: New file; >> (hunspell-dictionary): New variable; >> (hunspell-dict-it-it): New variable. >> * gnu/local.mk (GNU_SYSTEM_MODULES): Add it. > I’m picking up this patch as it seems to have fallen through the > cracks. :-) Definitely not a problem but still thanks a lot :) > > Overall it LGTM; a couple of comments below. > > Apparently this is different from the Hunspell dictionaries made from > Aspell word lists that appear in (gnu packages aspell), right? > > Should we name the packages differently to avoid collisions with the > existing ‘hunspell-dict-*’ packages? We definitely can but if I understood correctly they are generated from STOW wordlists [0] which seems to contain only variants of the English language. Since also the Libreoffice repository contains some of those variants, I was wondering, to achieve better consistency with future dictionaries, if it's worth either - to migrate the duplicates to use (hunspell-dictionary) from (gnu packages hunspell) - or just to use STOW for the English language and its variants and Libreoffice dictionaries for all the supported languages I believe either way we should be able to avoid collisions, but maybe I should just rename this one dictionary and, after we reach consensus, we can always decide to change it back. What do you think? Should I move this discussion to guix-devel? >> + (synopsis >> + (string-append "Hunspell dictionary for " full-name " (" dict-name ")")) > This is recommended against as this prevents i18n: > > https://guix.gnu.org/manual/en/html_node/Synopses-and-Descriptions.html I apologize about this, I attached an updated patch. Thank you for your time, Cheers Giacomo [0]: http://wordlist.aspell.net/scowl-readme/