On 9/28/15 10:38 PM, [email protected] wrote:
On Monday, September 14, 2015 at 6:37:06 AM UTC-7, Axel Hecht wrote:
We're calling malformed an error/warning right now. Missing is an error,
obsolete is a warning.
I'm not sure that's ultimately the right answer, in particular in the
light of locales which want to re-use en-US accesskeys.
I'd like to make that easy.
Doesn't make l10n-merge easier, though :-)
I'm wondering, can we restrict ourselves in the model here to DOM nodes.
In which case an obsolete attribute might actually be an error in the
sense that the l10n-merge code should remove it,
Sounds good to me.
while a missing attribute (or value) might be just a warning?
I have more trouble with that.
Missing attribute is not very different from missing a value and I'm concerned
if we try to make it work.
Example:
A `placeholder` attribute of an entity is de-facto its value. Missing it should
be an error and we should use fallback language entity then.
Accesskeys are interesting, in that I can see how you might want to skip them in your
partial locale and reuse the 'masters' one, but in that case, I'd expect some sort of
metainformation stating "missing accesskeys are ok, reuse the parent locale
ones" whitelist.
zb.
The question boils down to:
Is there a DOM element where we wouldn't want to take the text content
and some attributes because one attribute is missing?
Is there a DOM element where we wouldn't want to take the ..... because
the text content is missing?
Do we need to sanitize obsolete attributes? Or is the code in
https://github.com/l20n/l20n.js/blob/master/src/bindings/html/overlay.js
strong enough?
Axel
_______________________________________________
tools-l10n mailing list
[email protected]
https://lists.mozilla.org/listinfo/tools-l10n
