Re: [PATCH] gitweb: fix link to parent diff with pathinfo

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, May 24, 2016 at 8:17 PM, Junio C Hamano <gitster@xxxxxxxxx> wrote:
> Richard Braun <rbraun@xxxxxxxxx> writes:
>
>> Gitweb, when used with PATH_INFO, shows a link to parent diff
>> like http://somedomain/somerepo.git/commitdiff/somehash?hp=parenthash.
>> That link reports "400 - Invalid hash parameter".

Richard, at which view is this bad link present? Err... never mind, I see that
gitweb uses link to 'commitdiff' view with 'hash_parent' parameter set only
..in two places (well, perhaps there are some which get hash_parent from
-replay, but I doubt it): the "commit" view (link to each parent commit)
and in "commitdiff" view for octopus merges (e.g. "pu" in git.git).

The problem is not with ?hp=parenthash, but with /somehash part, which
somehow got invalid (from the error message). I have checked using
http://repo.or.cz/git.git, and while it has the bug (i.e. uses '?hp=...' instead
of path_info), there is no "400 - Invalid has parameter" error.

>> As I understand it, it should instead directly point to the parent diff,
>> i.e. turn it into http://somedomain/somerepo.git/commitdiff/parenthash,

Actually, the correct path_info based URI would be

  http://somedomain/somerepo.git/commitdiff/parenthash..somehash

Just like href() does with hash_parent_base and hash_base for blobdiff.
Urgh... href() is a bit of mess, now I see it when I am not current.

>> and delete 'hash_parent' element from the %params hash once we used it,
>> otherwise the '?hp=parenthash' string is appended.

That's correct: the unstated rule of href is that if it went into path_info,
it is deleted (not everything can be expressed with path_info), the rest
goes into query parameters. So without deleting the element, it would
be duplicated.

Note that using query parameter when we can use path_info is a minor
error; URL should work anyway (and I don't see why it doesn't - somewhat
the 'hash' parameter got incorrect...).

>>
>> Signed-off-by: Richard Braun <rbraun@xxxxxxxxx>
>> ---
>
> Pinging...

I'm sorry, I didn't notice it was meant for me. Simple "Jakub,..."
would be enough.

On Tue, May 24, 2016 at 8:39 PM, Junio C Hamano <gitster@xxxxxxxxx> wrote:
> Richard Braun <rbraun@xxxxxxxxx> writes:
>
>> On Tue, May 24, 2016 at 11:17:28AM -0700, Junio C Hamano wrote:
>>> Pinging...
>>
>> Hum, see [1].
>>
>> Tell me if I need to resend.
>
> Sorry, check the "To:" field of the message you are responding to.
> The ping was not meant to (and was not addressed to) you.  It asked
> for comments from an area expert.

Only this made me realize that you are expecting *my* response.

>>  gitweb/gitweb.perl | 7 ++++++-
>>  1 file changed, 6 insertions(+), 1 deletion(-)
>>
>> diff --git a/gitweb/gitweb.perl b/gitweb/gitweb.perl
>> index 05d7910..f7f7936 100755
>> --- a/gitweb/gitweb.perl
>> +++ b/gitweb/gitweb.perl
>> @@ -1423,7 +1423,12 @@ sub href {
>>                       delete $params{'hash'};
>>                       delete $params{'hash_base'};
>>               } elsif (defined $params{'hash'}) {
>> -                     $href .= esc_path_info($params{'hash'});
>> +                     if (defined $params{'hash_parent'}) {
>> +                             $href .= esc_path_info($params{'hash_parent'});
>> +                             delete $params{'hash_parent'};
>> +                     } else {
>> +                             $href .= esc_path_info($params{'hash'});
>> +                     }

This should read _something_ like this

+                     if (defined $params{'hash_parent'}) {
+                             $href .=
esc_path_info($params{'hash_parent'}) . '..';
+                             delete $params{'hash_parent'};
+                     }
+                     $href .= esc_path_info($params{'hash'});
+                      delete $params{'hash'};

Otherwise you would get correct link in your situation with
bad 'hash' parameter, but not the view that was requested;
it would not be diff between current and given parent, but
commitdiff for parent (to grandparent(s)).

Richard, thanks for finding a problematic thing, but you
need to search more for a true fix.

Regards
-- 
Jakub Narebski
--
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



[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]