I am not able to clone via https because gerrit doesn't propose this way
of cloning.
When I clone via http, I see that git is starting the download of objects:
remote: Counting objects: 256, done
remote: Finding sources: 100% (256/256)
Receiving objects: 46% ... (this part always fails at 46 %).
So, I think that the proxy part is not a problem (there is no proxy set
here).
If I set GIT_CURL_VERBOSE=1 before cloning, I've got the following error:
* Problem (2) in the Chunked-Encoded data
* Closing connection 1
YC
Le 14/11/2013 00:40, brian m. carlson a écrit :
On Wed, Nov 13, 2013 at 08:43:28AM +0100, Yann COLLETTE wrote:
Hello,
When I perform the git clone via git, it works. The problem is only
happening via http.
I tried to play with http.postBuffer and I set this parameter to
it's maximum (a little bit before a git clone triggered a memory
alloc problem) and I see that something big is trying to be
downloaded. But I don't see such a big file in my history of
commits. The maximum one seems to be around 50 Mo ...
Please keep the list in CC.
http.postBuffer doesn't affect clones, only pushes, so that isn't
relevant here. You're experiencing something that is dropping the
connection over HTTP. So you either have a bad proxy, or something else
is causing the connection to be dropped. Since it's only over HTTP, I
suspect it's the former.
Do you have HTTPS, and if so, does it work if you try to clone over
that?
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html