e particular bug that you are talking about?
Boyer Moore overflow patch:
<http://lists.gnu.org/archive/html/bug-grep/2005-06/msg00039.html>
<http://lists.gnu.org/archive/html/bug-grep/2005-06/msg00040.html>
<http://lists.gnu.org/archive/html/bug-grep/2005-06/msg00048.html>
Please see the revised bug-fix patch, attached.
Charles Levert wrote:
Given that, it's pointless and even ambiguous
(if LC_ALL isn't already exported) to use
LC_ALL=... function-call ...
LC_ALL=... function-call ...
instead of
LC_ALL=...; export LC_ALL
function-call ...
functi
Charles Levert wrote:
The following patch should be equivalent and
much more readable.
+b17='b'
+b85="$b17$b17$b17$b17$b17"
+b255="$b85$b85$b85"
+x16=''
+x64="$x16$x16$x16$x16"
+x256="$x64$x64$x64$x64"
+bm="cbbba${b255}c"
+grep_test "a${b255}a" "" "a${b255}b"
+g
Paul Eggert wrote:
Julian Foad <[EMAIL PROTECTED]> writes:
+Moo that's not _that_ easy; you have to avoid an overflow. */
"Moo"?
Heh! I also wondered what that was, and assumed it was some expression like
the "Boy" in the previous sentence. I'll write my own version of that commen
> >for (i = 0; i < kwset->mind; ++i)
> > - delta[(unsigned char) kwset->target[i]] = kwset->mind - (i + 1);
> > + delta[(unsigned char) kwset->target[i]] =
> > + MIN(kwset->mind - (i + 1), UCHAR_MAX);
>
> There should be a space after the "MIN". (This is the style
> recommended by
Julian Foad <[EMAIL PROTECTED]> writes:
> + Moo that's not _that_ easy; you have to avoid an overflow. */
"Moo"?
>for (i = 0; i < kwset->mind; ++i)
> - delta[(unsigned char) kwset->target[i]] = kwset->mind - (i + 1);
> + delta[(unsigned char) kwset->target[i]] =
> + MI
* On Tuesday 2005-06-14 at 22:19:26 -0400, Charles Levert wrote:
> * On Wednesday 2005-06-15 at 01:45:10 +0100, Julian Foad wrote:
>
> I noticed there's a problem with LC_ALL still
> being set to $u from above, which it shouldn't
> be. I'll have to investigate that separately.
It turns out that,
* On Wednesday 2005-06-15 at 01:45:10 +0100, Julian Foad wrote:
>
> diff -u -3 -p -d -r1.8 foad1.sh
The following patch should be equivalent and
much more readable.
I haven't looked at the rest yet, so I can't
vouch for the validity of the tests themselves,
regardless of how they're scripted.
I
I'm looking back at this post from 22 November 2004:
Stepan Kasal wrote:
some time ago I have created a patch to kwset.c, which fixes an
overflow in the Boyer-Moore algorithm. According to our rules, I'm
posting it here for review before I check it in.
In a follow-up post, Stepan provided t
