Hi Amos,
At 06.03 10/07/2008, Amos Jeffries wrote:
Guido Serassio wrote:
> HI Amos,
>
> I'm trying to maintain up-to-date the Windows support even if it doesn't
> run, so I have found a little build problem.
On that train of thought...
You have mentioned, bzr (still waiting) and IPC weirdness.
--> We recently found a major bug in the configure code which prevented
the IPv6 --with-* options working. Since windows depends heavily on
those, I think that may be a reason why 3-HEAD was failing to run earlier.
Anything else fully blocking?
The situation is very bad:
On Windows Squid3 STABLE crashes after few requests, or, without
requests, during shutdown. I'm not sure about the reason of that,
while I'm still not able to debug the code because I cannot
understand it at all.
May be a problem related to the new run-time (Squid3 uses MSVCRT8,
while Squid 2 is still using the old MSVCRT), or may be an
incomplete/incorrect C++ translation of the Windows specific code, or
may be a very little stupid thing. But without a true Windows C++
developer the answer cannot be found.
For the same reason, I have never tried to run Squid3 HEAD on Windows
after IPV6 merge, there are still too much "excitements" in the
STABLE code .... :-(
Regards
Guido
Amos
>
> This change broke fake_auth on Windows:
> http://www.squid-cache.org/Versions/v3/HEAD/changesets/b8894.patch
>
> strtok_r() is not available on Windows. In the past there was an own
> implementation of strtok_r(), but it was removed in favour of our
> strwordtok():
> http://www.squid-cache.org/Versions/v3/HEAD/changesets/11355.patch
>
> Any suggestion on how to replace the strtok_r() usage ?
>
> Regards
>
> Guido
-
========================================================
Guido Serassio
Acme Consulting S.r.l. - Microsoft Certified Partner
Via Lucia Savarino, 1 10098 - Rivoli (TO) - ITALY
Tel. : +39.011.9530135 Fax. : +39.011.9781115
Email: [EMAIL PROTECTED]
WWW: http://www.acmeconsulting.it/
