[PATCH] Location header recovering to absoluteURI

Martin Conte Mac Donell reflejo at gmail.com
Fri Jan 23 21:10:26 CET 2009


On Fri, Jan 23, 2009 at 3:15 PM, Stipe Tolj <st at tolj.org> wrote:
> Hi list,
>
> we refuse to follow party defined absoluteURI Location header values. Which is
> "correct" when being pedantic with the specs. But some mayor HTTP services, i.e.
> Yahoo do send URI /xxx as Location header values. This patch ensures we are able
> to recover from this circumstance, and continue the following.
>
> Please review, test and vote for committing to CVS HEAD.
>
> Stipe

I'm just visiting here since i've got an insanely amount of work, but
i'v two comments.

1) Eclipse?, really?

2) I'm not pretty sure about "if (trans->port != 80)", it'll work fine
in most cases but there is a border case when it'll not (Yeah, i know.
I'm anal), the case is when you are going through ssl by port 80 (Yeah
ok, shouldn't happen but you never know). I think the safest way is
append port number always, it won't hurt anyway.

M.



More information about the devel mailing list