[Bug] Query string not being phrased correctly when question marks present in config URL.

[Bug] Query string not being phrased correctly when question marks present in config URL.

[Steven Lawler]

Cause:
[remote "repo"]
  url = http://example.com/git/example.com?foo=bar
There is a question mark in the URL of the repo URL.

Effect (Taken from Apache logs):
[22/Sep/2014:14:12:07 -0400] "GET
/git/example.com?foo=bar/info/refs&service=git-receive-pack HTTP/1.1"
403 207 "-" "git/1.9.4.msysgit.1"

Git attempts to correct the issue by making the query string continue
using ampersands where it would have started the original query
string.

Expected outcome:
The git client should move the ?foo=bar onto the beginning (or end) of
the query string that it is creating. For example:
[22/Sep/2014:14:12:07 -0400] "GET
/git/example.com/info/refs?foo=bar&service=git-receive-pack HTTP/1.1"
403 207 "-" "git/1.9.4.msysgit.1"

This is caused on the git client, double checked with Wireshark. I
have not tested this on HTTPS, but imagine the output is the same.

Re: [Bug] Query string not being phrased correctly when question marks present in config URL.

[Jeff King]

On Tue, Sep 23, 2014 at 04:20:55AM +1000, Steven Lawler wrote:

> Cause:
> [remote "repo"]
>   url = http://example.com/git/example.com?foo=bar
> There is a question mark in the URL of the repo URL.

Is the question mark there because it is separating query parameters
from the path, or is "?foo=bar" part of the literal path? If the latter,
it needs to be percent-escaped.

> Effect (Taken from Apache logs):
> [22/Sep/2014:14:12:07 -0400] "GET
> /git/example.com?foo=bar/info/refs&service=git-receive-pack HTTP/1.1"
> 403 207 "-" "git/1.9.4.msysgit.1"
> 
> Git attempts to correct the issue by making the query string continue
> using ampersands where it would have started the original query
> string.

It has to, because:

  http://example.com/git/example.com?foo=bar?service=git-receive-pack

would not mark the "service" parameter as a separate parameter of the
query-string.  However, from your expected output below, I do not think
that part is the thing you are complaining about:

> Expected outcome:
> The git client should move the ?foo=bar onto the beginning (or end) of
> the query string that it is creating. For example:
> [22/Sep/2014:14:12:07 -0400] "GET
> /git/example.com/info/refs?foo=bar&service=git-receive-pack HTTP/1.1"
> 403 207 "-" "git/1.9.4.msysgit.1"

The difference here is not the ampersand but the placement of
"info/refs".  Let's forget about the "service" parameter for a moment.
Git needs to access the "info/refs" directory of the repo specified by
the URL it is passed. Given:

  http://host/path?key=value

which part of the URL specifies the path to the repository (and that we
could munge to access info/refs)? I do not think there is a well-defined
answer. If "path" is the path to the repo and "key=value" is an
unrelated parameter, you would want:

  http://host/path/info/refs?key=value

which matches your example. But if the key is specifying the repository
path in its value, you would want to append it directly there. For
example,

  http://host/git?path=my-repo.git

should become:

  http://host/git?path=my-repo.git/info/refs

So I think you would need a config option or other mechanism to specify
which form your URL requires.

-Peff