Qt-interest Archive, March 2007
QRegExp minimal match
Message 1 in thread
Hi,
with setMinimal(true) and the string "abc" I expected the QRegExp
"^(.*)(.)?(.)?$" to set the submatches $1=a, $2=b and $3=c.
But it results in $1=abc. Why ?
Using Qt 4.2.2.
cu, Bernd
--
[ signature omitted ]
Message 2 in thread
Hi Bernd,
> with setMinimal(true) and the string "abc" I expected the QRegExp
> "^(.*)(.)?(.)?$" to set the submatches $1=a, $2=b and $3=c.
>
> But it results in $1=abc. Why ?
The minimal matching is "global", i.e. it tries to find the shortest
match possible for the entire string. The individual matches within the
string are still performed as usual.
I guess this begs the question: Why does't QRegExp support minimal
quantifiers, e.g., *? and +?. The short answer is that is very difficult
to accomplish with the current regexp engine. In the long term, I'd like
us to rewrite the engine (or use a 3rd party engine) so that Qt's regexp
support this kind of features and follow some standard (whether it be
Perl, JavaScript, POSIX, ...).
Regards,
Jasmin
--
[ signature omitted ]
Message 3 in thread
Jasmin Blanchette schrieb:
> so that Qt's regexp
> support this kind of features and follow some standard (whether it be
> Perl, JavaScript, POSIX, ...).
vim! ;-)
Martin, not explicitely looking for another editor war...
--
[ signature omitted ]
Message 4 in thread
Hi Jasmin,
thanks, I understand now.
Minimal quantifiers would be appreciated ;-)
Regards,
Bernd
--
[ signature omitted ]
Message 5 in thread
Bernd Strohhaecker schrieb:
> with setMinimal(true) and the string "abc" I expected the QRegExp
> "^(.*)(.)?(.)?$" to set the submatches $1=a, $2=b and $3=c.
Jasmins competent and authorative answer aside ;-) you could try if
"^(.)*(.)?(.)?$" or "^(.)?(.)?(.)?$" accomplishes what you need. But
that's just a "if you don't know how it works, try to find out" approach...
Martin
--
[ signature omitted ]
Message 6 in thread
Hi Martin,
also thanks. But this was only an example, the real regexp was more
complicated and I only wondered why my first approach didn't work.
Especially because it workes fine in Perl - and also vbscript.regexp ;-)
FWIW, to realize this with greedy quantifier and leave exactly two chars
for the later subexpressions you could also use a negative lookahead:
^((?:(?!..$).)*)(.)?(.)?$
cu, Bernd
--
[ signature omitted ]
