| Age | Commit message (Collapse) | Author |
|---|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
In the log interface, there is a toggle link at the top, but it isn't clear
what's being toggled. I've changed it to "Expand" and "Collapse" to make it
clear that you are getting more and less information, respectively.
Signed-off-by: Justin Waters <[email protected]>
|
|
|
|
|
|
|
|
|
|
Commits are now decorated with a clickable 'label' for each ref pointing
at it, similar to how gitweb and gitk displays commit decorations.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
We don't need support for fetching/pushing in libgit.a, hence we don't need
to link with libcurl.
Noticed-by: Robin Redeker <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Robin Redeker <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
|
|
Also, fix a related bug in the test-suite.
Noticed-by: Jim Meyering <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
|
|
Add a space between the committer name and email, and remove
superfluous spaces in the date header.
This makes cgit-generated patches match the output from git-format-patch
almost exactly, at least as far as the email headers go.
Signed-off-by: Tomas Carnecky <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Since cgit now returns a 404 error when the snapshot filename cannot be
resolved to a commit object, the testscript needs to request a valid
snapshot name.
Also, the script assumed that the toplevel directory in the snapshot
would get the name of the repository but it's now named similar to
the requested snapshot.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
The updated mime-types seems to work better than the old ones with both
safari and firefox.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
SCRIPT_NAME is used as fallback for virtual-root when virtual-root is
unspecified in cgitrc and PATH_INFO is used as fallback for the query-
string parameter 'url' when the latter is unspecified. But until now,
the use of PATH_INFO depended on virtual-root also being unspecified,
i.e. it was impossible to use PATH_INFO when virtual-root was specified.
This commit makes the fallback on SCRIPT_NAME and PATH_INFO independent
code paths, i.e. it is now possible to specify virtual-root in cgitrc
while still using PATH_INFO (instead of rewrite rules) to get 'pretty
urls'.
Noticed-by: Jack Moffitt <[email protected]>
Noticed-by: LiKai Liu <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
|
|
The "((expr))" construct is not implemented by e.g. dash, so this commit
replaces the construct with a more portable one.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Noticed-by: Ramsay Jones <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Conflicts:
cgit.c
cgit.h
|
|
|
|
|
|
|
|
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
The non-standard function strcasestr is only defined if _GNU_SOURCE has
also been defined.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
If the makefile doesn't automatically define the correct build variables
it is nice to be able to define them explicitly.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Ramsay Jones <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
The 'h' querystring parameter in cgit is normally used to specify
a branch (i.e. a ref below refs/heads/), but if a repository contains
a tag with the same name as a branch the output from ui-log would use
the tag as start-revision.
This patch tries to fix the issue by checking if the specified ref is
valid as a branch name; if so, the full refname is used in the call
to setup_revisions().
Noticed-by: Takamori Yamaguchi <[email protected]>
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Such links was printed as the object type followed by the objects complete
sha1. We still use the complete sha1 in the link but we no longer show it
in all its glory; only the first 10 hex chars are printed.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
The snapshot function has only been linked to from the commit page while
users often would want to download a certain release. With this patch,
direct download links will now be printed for each tagged release on the
repo summary page.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Since we know that ui-snapshot.c is able to extract the revision from the
filename, there's no longer necessary to specify the revision with a 'id'
querystring argument.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
The modified get_ref_from_filename() supports the following snapshot
formats:
* $REV.$EXT
* $REPO[-_]*v?$REV.$EXT
This implies that the following urls will retrieve the expected revision:
* http://hjemli.net/git/cgit/snapshot/v0.8.1.tar.gz
* http://hjemli.net/git/cgit/snapshot/0.8.1.tar.gz
* http://hjemli.net/git/cgit/snapshot/cgit-0.8.1.tar.gz
* http://hjemli.net/git/cgit/snapshot/cgit-140012d7a8.tar.gz
Signed-off-by: Lars Hjemli <[email protected]>
|
|
This patch sets the directory prefix in archives to be the filename,
excluding the suffix (.tar.gz, .tar.bz2 etc).
The patch also removes the prefix parameter in cgit_print_snapshot()
as the prefix might differ.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
If the commit message has extra long lines it's better to use the author/
files/lines columns to show those lines than to push the columns of screen
and force the users to hscroll.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Some users prefer to see the full message, so to make these users happy
the new querystring parameter "showmsg" can be used to print the full
commit message per log entry.
A link is provided in the log heading to make this function accessible,
and all links and forms tries to preserve the users preference.
Note: the new link is not displayed on the summary page since the point
of the summary page is to be a summary, but it is still obeyed if specified
manually.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
When sorting the list of repositories by their last modification time,
cgit would (in the worst case) invoke fstat(3) four times and open(3)
twice for each callback from qsort(3). This obviously scales very badly.
Now, the calculated modtime for each repo is saved in repo->mtime, thus
keeping the number of stat/open invocations identical for sorted and
unsorted repo-listings.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
When sorting on e.g. owner, it's not interesting to get all repos
without owner at the top of the list.
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
The new function is then used by both print_modtime() and
cgit_reposort_modtime().
Signed-off-by: Lars Hjemli <[email protected]>
|
|
Signed-off-by: Lars Hjemli <[email protected]>
|
|
|
|
When the agefile was empty the old code would happily reuse the static
buffer filled by a previous call to read_agefile().
Signed-off-by: Lars Hjemli <[email protected]>
|
|
|
