On 11/11/11 05:44, Karl Koehler wrote:
Hi,
investigating further I have come to the conclusion that the attached
( and previously inlined ) patch is correct, in the sense that regular
expressions of the failing type are now handled correctly.
e.g.
A[\[:\]]
should find an 'A' followed by one of '[',':',']' and it does now (old
behavior: infinite loop).
I don't seem to see any regression tests for reclass ?
Also, and independent of this issue, certain regular expressions don't
work as expected, e.g.
A[\[:]
unexpectedly matches nothing, but
A[:\[]
and
A[\[\:]
do... ( as far as I know ':' should match itself unless one of the
special posix ranges is used. )
Review or comment would be greatly appreciated !
Thanks,
imo the regrex expression is invalid, but... it seems the regrex
parser/compiler thing ( whose code I failed to understand without
spending lots more time on it ) doesn't handle this situation well.
However I agree with you analysis from stepping through the code. I
can't see it can make things worse.
would you be interested in creating some unit tests for this ? that
would be really useful I think
Noel
ps. could you confirm license is LGPLv3+ / MPL.
_______________________________________________
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
