Re: squid2 ipv6 questions

From: Amos Jeffries <squid3@dont-contact.us>
Date: Tue, 2 Oct 2007 11:20:18 +1300 (NZDT)

> I have several questions on patching squid2:
>
> * what are the procedures to include my ipv6 patch into HEAD?
>

There is an old 'ipv6' branch in CVS under the 'squid' module. I have
already cleaned out the old dead experiments and brought it up to match
the current HEAD for you.

cvs -d....@devel.cvs.squid-cache.org:/cvsroot/squid checkout -ripv6 -d
squid2-ipv6 squid
Where ... is your sourceforge login as a developer.

Also added details to http://devel.squid-cache.org/projects.html (you may
wish to check after the next site rebuild).
That 2.x branch is now yours to do with as you need.

Your patch code needs to:
 - apply cleanly to the branch
 - go through any alpha testing there
 - match http://wiki.squid-cache.org/Squid2CodingGuidelines
 - have a manual audit before approval in merging up to HEAD

Also a good idea just before it goes up for approval to do
"cvsmerge HEAD" to make it as close to HEAD as possible before the patch
release.

Any problems just ask in squid-dev.

> * are there any squid test units/programs to check my code?
> So far I'm just running my STABLE13 patch without any problem.
> If the patch is to be included in HEAD, I think it should need to
> pass some tests.

hmm, good point. I'm not too sure about 2.x testing. That would be a
Henrik question.

> * how to set entry->ptr_flag used in cf_gen.c:gen_dump() ?
> I changed in_addr into sock_storage and I want to pass sock_addr pointer
> to
> dump_address()
> I understand that this function is used for debugging purposes only.

In 3.0 I had no problems changing the config types in src/cf.data.pre and
coding parse_*, free_*, dump_* functions in src/cache_cf.c to match them.

> * strict casting?
> my patch produces a lot of warning since I'm not doing any casting
> --mainly sockaddr_storage* into sockaddr*-- when passing parameters

yes type-strictness is needed. If you MUST do casting at all, use a
proper visible code casting "(type*)X".
In the public releases all warnings will be seen as compile errors.

>
> That's all for now.
>
> best regards,
> -husni
>

Thanks for this husni.

HTH
Amos
Received on Mon Oct 01 2007 - 16:20:47 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Oct 30 2007 - 13:00:03 MDT