Re: Question about critical_factorization() in the Two-Way algorithm

[Pádraig Brady]

On 22/06/10 15:09, Eric Blake wrote:
> On 06/22/2010 03:18 AM, Pádraig Brady wrote:
>>
>> Note the docs for "memmem" say it fixes
>>
>>   "This function returns incorrect values in some cases, such as when
>>    given an empty needle: glibc <= 2.0, Cygwin 1.5.x."
>>
>> Could that functionality be rolled into memmem-simple so that
>> memmem is just the "fast/fat" version? If not could you briefly
>> expand on "some cases" above as I don't know from the docs if
>> I can use memmem-simple.
> 
> The idea is that memmem-simple gives you the replacement only if the
> system memmem is buggy (gives wrong answers or reads beyond bounds[1])
> but without regards to whether it is quadratic, while memmem also gives
> you the replacement if the system memmem is slow.

Cool, that's what I expected from the names.
So do we just need to move a docs around a little like:

diff --git a/doc/glibc-functions/memmem.texi b/doc/glibc-functions/memmem.texi
index c7e3d73..7d34817 100644
--- a/doc/glibc-functions/memmem.texi
+++ b/doc/glibc-functions/memmem.texi
@@ -13,14 +13,14 @@ MacOS X 10.3, FreeBSD 5.2.1, OpenBSD 4.0, AIX 4.3.2, HP-UX 
11, IRIX 6.5, OSF/1 5
 @item
 This function has reversed arguments on some older platforms:
 Linux libc 5.0.9
address@hidden itemize
-
-Portability problems fixed by Gnulib module @code{memmem}:
address@hidden
 @item
 This function returns incorrect values in some cases, such as when
 given an empty needle:
 glibc <= 2.0, Cygwin 1.5.x.
address@hidden itemize
+
+Portability problems fixed by Gnulib module @code{memmem}:
address@hidden
 @item
 This function has quadratic instead of linear worst-case complexity on some
 platforms: