Re: [PATCH] Add ERR support to smart HTTP

From: Shawn O. Pearce <spearce@spearce.org>
Date: Sep 6, 2010, 18:19 UTC
Message-ID: <20100906181921.GN32601@spearce.org>
In-Reply-To: <AANLkTikmU9_Vg2+=73yjPyaaDSqk73Bvs1HyNjFDWqNY@mail.gmail.com>

Sitaram Chamarty <sitaramc@gmail.com> wrote:

Show 14 quoted lines

> On Mon, Sep 6, 2010 at 8:26 PM, Shawn O. Pearce <spearce@spearce.org> wrote:
> > That is, the following will trigger a correct error on the client:
> >
> >  200 OK
> >  Content-Type: application/x-git-upload-pack-advertisement
> >
> >  001e# service=git-upload-pack
> >  0022ERR You shall not do this
> 
> are those counts accurate for the specific example you show or just made up?
> 
> It seems the first line has a count in hex that includes the newline
> at the end, and the second one has a count in decimal that does not
> include the newline nor even the 4-digits plus "ERR"

Feh.  I can't count.  The first count is correct.  The second count
should also be 001e.  I guess that should be obvious by just looking
at the two lines, they are equal in length.  :-)

Show 5 quoted lines

> > Likewise if you wanted to do this with receive-pack, replace upload
> > with receive above and adjust the pkt-line lengths.
> 
> ok... what about all the other service commands?  like /info/refs?
> What should I put there?

The only other command that matters is info/refs.

For smart clients, its what I said above.

For dumb clients, you have to use some sort of HTTP error status that isn't 404. Dumb clients pre-1.6.6 use a curl error message buffer to print out an error. But they don't check the format of info/refs at all, and skip over garbage and/or interpret garbage as valid input. So we can't use a hack like "ERR blah" to even trigger a parsing failure.

-- 
Shawn.