git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] Update setup-git-server-over-http.txt based on my experience.

From
Johannes Schindelin <johannes.schindelin@gmx.de>
Date
Apr 3, 2008, 20:28 UTC
Message-ID
<alpine.LSU.1.00.0804032218550.4008@racer.site>
In-Reply-To
<vpqtziimzcp.fsf@bauges.imag.fr>
Hi,
On Thu, 3 Apr 2008, Matthieu Moy wrote:
Show 9 quoted lines
> Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:
> 
> > I don't like the commit subject.  At all.
> 
> Well, I can't think of a better one. The patch introduces several minor 
> changes, and I don't think it deserves a real patch serie for each 
> unrelated change.
> 
> Suggestions welcome.

Since your changes are mostly about using WebDAV to copy an empty git repository to the server, you could allude to that.

Show 18 quoted lines
> > On Thu, 3 Apr 2008, Matthieu Moy wrote:
> >
> >> -- have git installed at the server _and_ client
> >> +- have git installed on the client, and preferably on the server
> >
> > How do you want to initialise the repository on the server, then?  
> > (IOW you should mention here that you need a WebDAV client on the 
> > client if you do not have Git on the server, and that it is way more 
> > fiddly.)
> 
> What part of
> 
>  Another
>  option is to generate an empty repository at the client and copy it to
>  the server with a WebDAV client (which is the only option if Git is
>  not installed on the server).
> 
> is unclear ?
It is too long.
-- have git installed on the client
-- have git installed on the server (or if you cannot, get a WebDAV 
   client such as Konqueror, Internet Exploder, etc.)
Show 7 quoted lines
> >> -In effect, this probably means you're going to be root.
> >> +In effect, this probably means you're going to be root, or that you're
> >> +using a preconfigured WebDAV server.
> >
> > Either you strike "probably" or you skip what you added.
> 
> Why?
Because you do not need to be "root" if WebDAV is preconfigured?
Show 13 quoted lines
> >> @@ -169,7 +170,8 @@ On Debian:
> >>  
> >>     Most tests should pass.
> >>  
> >> -A command line tool to test WebDAV is cadaver.
> >> +A command line tool to test WebDAV is cadaver. If you prefer GUIs,
> >> +konqueror can open WebDAV URLs as "webdav://..." or "webdavs://...".
> >
> > What do you do if you have MacOSX, or Windows?
> 
> You read messages instead of truncating them if you're on windows.
> 
> I have no idea about MacOSX, but that's not the point of my patch.
Right, but the text you modify is not supposed to be Linux-specific.
Show 17 quoted lines
> >> -Make sure that you have HTTP support, i.e. your git was built with curl.
> >> -The easiest way to check is to look for the executable 'git-http-push'.
> >> +Make sure that you have HTTP support, i.e. your git was built with
> >> +curl (and a recent enough version).
> >
> > Say what version.  Otherwise this comment will soon be very, very stale.
> >
> >> +The easiest way to check is to +look for the executable 
> >> 'git-http-push'. The command "git http-push" +with no argument should 
> >> display a usage message.
> >
> > My search revealed that http-push was in Git since tags/v0.99.9e^2~9^2~4.  
> > Which is not recent at all.
> 
> Before my patch, the explanation says that you just need to check
> whether you have git-http-push, which is insufficient. With my patch,
> it gives an accurate check. I believe this is an improvement.

Oh? So a usage message qualifies for "recent enough"? In that case, v0.99.9f is "recent enough". I find that hardly helpful.

Show 10 quoted lines
> >> +Also note that the URL should point to the git repository itself, that
> >> +is, to the '.git/' directory and not the working tree in case the
> >> +repository is non-bare.
> >
> > It makes no sense to describe the case of a non-bare repository.
> 
> Actually, it makes no real sense to have a non-bare repository. But I've 
> been bitten by this (I just typed "git init" without --bare, and 
> uploaded it). Since git-http-push gives _very_ bad error messages, it's 
> good to point the user to potential mistakes to help troubleshooting.

If it makes no real sense to have a non-bare repository, why do you even _suggest_ what to do in that case? You should rather say that it makes no sense to have a non-bare repository. And probably add "because a http-push will not update the working directory anyway".

Show 11 quoted lines
> > >> +Using a proxy:
> >> +--------------
> >> +
> >> +If you have to access the WebDAV server from behind an HTTP(S) proxy,
> >> +set the variable 'all_proxy' to 'http://proxy-host.com:port', or
> >> +'http://login-on-proxy:passwd-on-proxy@proxy-host.com:port'. See 'man
> >> +curl' for details.
> >
> > You only need http_proxy.
> 
> What makes you think that?
I use it in msysgit.

(Okay, not for pushing, but for fetching, which technically uses the same protocol.)

> (hint: I've been bitten by this too).

The only way I see this could bite you is if curl does not heed http_proxy for https:// protocol.

There's another thing. My curl manpage insists that ALL_PROXY is all capital. Can you clarify?

Ciao, Dscho

Previous: Matthieu Moy
Message 4 of 4 in “Update setup-git-server-over-http.txt based on my experience.”
  1. Update setup-git-server-over-http.txt based on my experience.Matthieu Moy, Apr 3, 2008
  2. Johannes SchindelinApr 3, 2008
  3. Matthieu MoyApr 3, 2008
  4. Johannes SchindelinApr 3, 2008

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.