Re: [PATCH] Docs: web server must setenv GIT_PROTOCOL for v2

Jeff King [off-list ref] wrote:

On Wed, Sep 08, 2021 at 06:48:47AM -0400, Jeff King wrote:

quoted

Both of the included examples here have been tested to work. The one for
lighttpd is a little less direct than I'd like, but I couldn't find a
way to directly set an environment variable to the value of a request
header. From my reading of the documentation, lighttpd will set
HTTP_GIT_PROTOCOL automatically, but git-http-backend looks only at
GIT_PROTOCOL. Arguably http-backend should do this translation itself.

So having discovered this, I kind of wonder if these documentation
patches are barking up the wrong tree. There is no reason we would not
want v2 to work out of the box (after all, it does for git://).

Agreed.

The patch below does that (and could replace both my and Konstantin's
documentation patches).

<snip>

-- >8 --
Subject: [PATCH] http-backend: handle HTTP_GIT_PROTOCOL CGI variable

When a client requests the v2 protocol over HTTP, they set the
Git-Protocol header. Webservers will generaly make that available to our

"generally"

CGI as HTTP_GIT_PROTOCOL in the environment. However, that's not
sufficient for upload-pack, etc, to respect it; they look in
GIT_PROTOCOL (without the HTTP_ prefix).

Either the webserver or the CGI is responsible for relaying that HTTP
header into the GIT_PROTOCOL variable. Traditionally, our tests have
configured the webserver to do so, but that's a burden on the server
admin. We can make this work out of the box by having the http-backend
CGI copy the contents.

Agreed.  I've completely overlooked GIT_PROTOCOL support, so far...

This seems to be the right thing to do; I think I'll add support
for it when I spawn git-http-backend in something I work on.
(I also don't currently pass all HTTP headers in env when
spawning CGI, maybe I should *shrug*)