Re: What's cooking in git.git (Nov 2009, #02; Sun, 08)

From: Tarmigan <tarmigan+git@gmail.com>
Date: Nov 10, 2009, 17:10 UTC
Message-ID: <905315640911100910r5116779eh24796fa5788f4aef@mail.gmail.com>
In-Reply-To: <7vzl6wz36r.fsf@alter.siamese.dyndns.org>

Sorry to bug you so soon again, Shawn.  Thanks for your help.

On Sun, Nov 8, 2009 at 9:18 PM, Junio C Hamano <gitster@pobox.com> wrote:

> * sp/smart-http (2009-11-04) 30 commits
>  + http-backend: Test configuration options
>  + http-backend: Use http.getanyfile to disable dumb HTTP serving
>  + test smart http fetch and push

OK now I'm testing on OSX with Apache/2.2.11, and curl 7.19.4.

Testing on 7f640b778f8cf87159890157a815f1d728573477, I am getting failures starting in the third test of t5560-http-backend.sh, "static file is ok". I also get failures on the test after ("static file if http.getanyfile true is ok") and in the log compare (obviously), but I think those will go away when test 3 is fixed, so I have not looked into them.

Specifically, the GET of info/refs makes something on my machine unhappy. Curl returns 18 (CURLE_PARTIAL_FILE), the test takes a long time to fail, and the "out" file looks OK (compared to a linux machine where the test passes) expect for "Content-Length: 37847251812411". The exact line does not work is curl --include http://127.0.0.1:5560/smart/repo.git/info/refs

Digging into it a bit more with gdb, the call to hdr_int() in http-backend.c looks OK, but then something goes wrong in format_write(). Hmmm it looks like my setup does not like PRIuMAX with size_t, which puts some garbage in the upper bytes of Content-Length. Changing the PRIuMAX to a %zu allows all the tests to pass, but that may not be the preferred solution.

Thanks, Tarmigan