View mode: basic / threaded / horizontal-split · Log in · Help
October 08, 2006
Re: DMD 0.169 release
"Dave" <Dave_member@pathlink.com> wrote in message 
news:egbuqt$1qq6$1@digitaldaemon.com...
> nobody_ wrote:
>> ~ Wondering as to how this would effect the shootout ~
>> ?????
>>
>>> Walter, I'm really beginning to suspect that you're clairvoyant ! How 
>>> did you find out about this gamedev thread ? ( 
>>> http://www.gamedev.net/community/forums/topic.asp?whichpage=1&pagesize=25&topic_id=418236 )
>>>
>>> Thanks for the fixes ! :-D
>>
>>
>
> It won't, but getting the tail-recursion optimization back will ;)

Mkay, I thought it might effect pidigits. :(
(btw. those questionmarks were supposed to be a smiley (weaboo style). :)
October 13, 2006
Re: DMD 0.169 release
> Fixed Bugzilla 395, but there are probably more UTF bugs in
> std.regexp.

Hmm. Fixing Phobos whenever something is brought up, is probably a good 
tack.

OTOH, quite some work may be saved if we study enough to get a feeling 
for what _not_ to even try to implement. Time savings should be substantial.

An example:

UTF-bugs in STD-Regexp may be one particularly prominent case.

One might want to develop a Robust library in D. One might instead be 
business oriented, which means, get something that works "somewhat" like 
you need, and then callously copy that.

---

The opposite tack is to adopt the PCRE library as such. Then we'd of 
course submit to the whims of the PCRC guys, but in the decades past us, 
we've seen that this guy really is at it for its own sake. (As 
especially opposed by "for the money".)

Another problem is, the UTF definition keeps changing every once in a 
while. Why not let Professionals take care of the whole shebang?
October 13, 2006
Re: DMD 0.169 release
std.regexp has been around for 6+ years. It comes from one I did in C++ 
that was very intensively tested. I think it has held up very well. The 
only thing it lacks is being thoroughly tested for UTF. I don't think 
that's justification for starting over with something else.
October 13, 2006
Re: DMD 0.169 release
Walter Bright wrote:
> std.regexp has been around for 6+ years. It comes from one I did in C++ 
> that was very intensively tested. I think it has held up very well. The 
> only thing it lacks is being thoroughly tested for UTF. I don't think 
> that's justification for starting over with something else.

Can you clear up a mystery about it? From looking at the code, it looks 
as though it can do lazy matching (references to REnmq, "minimal munch", 
and parsing of *?, +?, etc), and it's passed the simple tests I've tried 
on it. But it's not documented! So is the lazy matching:
(a) working, but not documented, or
(b) unfinished and buggy?
October 13, 2006
Re: DMD 0.169 release
Don Clugston wrote:
> Walter Bright wrote:
>> std.regexp has been around for 6+ years. It comes from one I did in 
>> C++ that was very intensively tested. I think it has held up very 
>> well. The only thing it lacks is being thoroughly tested for UTF. I 
>> don't think that's justification for starting over with something else.
> 
> Can you clear up a mystery about it? From looking at the code, it looks 
> as though it can do lazy matching (references to REnmq, "minimal munch", 
> and parsing of *?, +?, etc), and it's passed the simple tests I've tried 
> on it. But it's not documented! So is the lazy matching:
> (a) working, but not documented, or
> (b) unfinished and buggy?

(a) lazy documentation <g>.
October 13, 2006
Re: DMD 0.169 release
Walter Bright wrote:
> Don Clugston wrote:
>> Walter Bright wrote:
>>> std.regexp has been around for 6+ years. It comes from one I did in 
>>> C++ that was very intensively tested. I think it has held up very 
>>> well. The only thing it lacks is being thoroughly tested for UTF. I 
>>> don't think that's justification for starting over with something else.
>>
>> Can you clear up a mystery about it? From looking at the code, it 
>> looks as though it can do lazy matching (references to REnmq, "minimal 
>> munch", and parsing of *?, +?, etc), and it's passed the simple tests 
>> I've tried on it. But it's not documented! So is the lazy matching:
>> (a) working, but not documented, or
>> (b) unfinished and buggy?
> 
> (a) lazy documentation <g>.

Awesome! I hoped that was it. We could have an Easter Egg competition -- 
find the coolest thing in D, that isn't documented. <g>
A memorable previous entry was the simplified function template syntax.
October 14, 2006
Re: DMD 0.169 release
Don Clugston schrieb:
> Awesome! I hoped that was it. We could have an Easter Egg competition -- 
> find the coolest thing in D, that isn't documented. <g>
> A memorable previous entry was the simplified function template syntax.

I want credit for finding it if there will ever be such a competition *g*
Next ›   Last »
1 2
Top | Discussion index | About this forum | D home