Re: [PATCH] fixed unicode problem and added POSIX expression

["Daniel F. Savarese" <df...@savarese.org>]

>I wrote new patch and post again. It has following features:

The patch looks largely sound, but I'm going to change a few things before
applying it.  First, thanks a lot for the time you've put in.  Also,
thanks for updating Perl5Debug, which people tend to overlook when
adding code.

I have the following nits though.  I'd rather use
Collections.synchronizedMap(new HashMap()) over Hashtable, but since
it's read only there's no need for the synchronizedMap part and 
it should be private static final.  Perl's posix expressions support the
following negation syntax which wasn't implemented in the patch: [:^digit:]
The [:graph:] class actually refers to any alphanumeric or punctuation
character, not symbols like Character.MATH_SYMBOL (it's just a poorly
named character class).  Printable characters aren't the complement of
control characters, they are defined in Perl 5.6 as any alphanumeric,
punctuation, or space.  There are other subtle deviations.  My mentioning
of it is to encourage code contributors to be thorough and check the
Perl documentation and/or source to verify expected behavior.  There
are a couple more involved observations I was going to make, but I don't
want to sound like I'm criticizing excessively.  We can make
improvements/changes later.

Takashi, thank you for your work.  I'm going to apply a slightly modified
version of the patch and tag the release as release-2.0.2-dev-2 and put
a binary/source distribution in the download area.

daniel

Re: [PATCH] fixed unicode problem and added POSIX expression

[Takashi Okamoto <to...@rd.nttdata.co.jp>]

daniel:
> Perl's posix expressions support the
> following negation syntax which wasn't implemented in the patch:
[:^digit:]

[^[:digit:]] is available instead of [:^digit:].
However I can implement this feature to add negative OpCode.  For example:

    // NDIGIT means negative digit
      __hashPOSIX.put("^digit",     OpCode._NDIGIT);

Should I implement it for Perl compatibility?

> The [:graph:] class actually refers to any alphanumeric or punctuation
> character, not symbols like Character.MATH_SYMBOL (it's just a poorly
> named character class).  Printable characters aren't the complement of
> control characters, they are defined in Perl 5.6 as any alphanumeric,
> punctuation, or space.  There are other subtle deviations.  My mentioning
> of it is to encourage code contributors to be thorough and check the
> Perl documentation and/or source to verify expected behavior.

Thanks for your advices. I didn't read Perl 5.6 document and source so
carefully.
I'll refer it more cafefully next time.
--------------------------------------
Takashi Okamoto