First of all, I would like to politely ask everybody on the list to
change subject if, well, subject of the discussion changes. I was
totally under impression that this topic still discusses libidn2
extension in PECL and might miss discussion about intl IDNA patch if
David didn't point it to me (thanks!). Let's use subjects for their
intended purpose :)
Now about the patch:
0. I consider it a bugfix, so I am OK with getting it in anytime it's
ready. Thanks for making the quick patch, Gustavo!
1. I'm not sure I understand why we need two options fields. We already
have one options field, won't that be enough? We can combine options and
space them in a way that old ones work fine with new ones, can't we?
And have default variant so one of them is always true (probably 2003).
2. I think optional by-ref parameter to receive IDNA-specific error
codes is right. Generic intl error reporting is about reporting, well,
generic errors and is not meant to have granularity enough to store
whole UIDNAInfo stuff.
3. I'm not sure I understand the code in php_intl_bad_args() - it
composes error message into the buff but then only sets it if msg !=
NULL. It doesn't check msg before using it in sprintf. Looks like
something's missing there: I don't see how msg can be NULL there at all
but if we check it, let's check it right :) Another place has the same
4. Also we'd want some tests with that ;)
Stanislav Malyshev, Software Architect
(408)454-6900 ext. 227