>So, how could we check whether "Portuguese" wikis would break by doing
>this change?

According to Wikipedia, Portuguese sorting is as follows: "In addition
[to letters that are used in english], the following characters with
diacritics are used: Áá, Ââ, Ãã, Àà, Çç, Éé, Êê, Íí, Óó, Ôô, Õõ, Úú.
These are not, however, treated as independent letters in collation,
nor do they have entries of their own in Portuguese dictionaries. When
two words differ only in the presence or absence of a diacritic, the
one without it is collated first"

I just tested on my personal wiki, and can confirm that the ordering
when using this setting is as wikipedia describes they should be. I
didn't test super-exhaustively, but I feel very confident that this
setting would work fine for Portuguese without any further tailorings
needed.

Here's a screenshot of how uca-default sorts various letters used in
Portuguese: http://imgbin.org/images/7280.png .
The sort order used is consistent with
http://www.evertype.com/alphabets/portuguese.pdf (As far as I can
tell, assuming i'm reading that pdf correctly)


[from a different email]
>> So, how could we check whether "Portuguese" wikis would break by doing
>> this change?
>
>As Tim said:
>* Set one of the test wikis (Testwiki, Testwiki2) to Portuguese
>* change to said collation
>* do some editing
>* see if it breaks

Note, setting the language to Portuguese is unnecessary as we
currently do not support per-language tailoring of the collation. All
languages get sorted the same at the moment (there are bugs in
bugzilla to change this, and really it should be changed, but such
per-language support has yet to be implemented. However even if it was
fixed, its unclear if such a setting would be based on wiki content
language or not)

-bawolff

_______________________________________________
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Reply via email to