From: John Tapsell <johnflux@gmail.com>
Date: Thu, 05 Mar 2009 08:00:37 GMT
Subject: Re: [PATCH] Clarify the "cannot lock existing info/refs" error
Message-ID: <43d8ce650903050000g4466ab2fne5fb8ed009808346@mail.gmail.com>
In-Reply-To: <20090304192853.GA10567@localhost>

2009/3/4 Clemens Buchacher <drizzd@aon.at>:
> On Wed, Mar 04, 2009 at 03:37:06PM +0000, John Tapsell wrote:
>> -fprintf(stderr, "Error: cannot lock existing info/refs\n");
>> +error("cannot lock existing info/refs on remote server\nPerhaps the
>> server is currently busy, or your ~/.netrc file is not configured
>> correctly.");
>
> In my experience this is usually caused by http-push crashing and leaving
> stale locks behind until it times out after 10 minutes. I don't think we
> should speculate here unless we can narrow down the error condition.

Yeah, I was thinking of trying to narrow it down as well.

I personally get that error 3 out of 4 times, roughly, that I try to
push.  But that's probably because I'm using a very busy git server.
So in my case, I just keep retrying until it succeeds.

On google I found that people had been getting that error if they have
the wrong password.

Can we at least change it to:

-fprintf(stderr, "Error: cannot lock existing info/refs\n");
+error("cannot lock existing info/refs on remote server\n");

It's currently confusing as to whether it's a local error or a remote error.

John