Makes a tagged git repository out of a series of input directories representing significant states of a source tree.
For a script that does the opposite, see unmkrepo.
Usage:
mkrepo.sh input-dir output-dir [include-dir]
Where:
input-diris the directory containing the input directories.output-diris the directory where the repository should be created. It must not yet exist.include-diris an optional directory containing a set of files to be included in each commit.
When you run this script, the directories within input-dir will be
visited, and a new commit will be created for each. Each commit will be
tagged using the directory name. The order of processing is alphanumerical
by default, but may be customized.
Prior to each commit, a scan will be performed for empty directories.
For each found, an empty .gitignore file will be placed within,
ensuring that the directory's existence will be recorded in git.
By default, the commit message for each directory will be identical to
the directory/tag name. But if a file exists in input-dir with a
.txt extension and the same base name of the directory, the
content of that file will be used as the commit message instead.
By default, each comm
5426
it will be created as a child of the previous commit
on the master branch. But if a file exists in input-dir with a
.branch extension and the same base name of the directory, a new
branch will be created instead, and the commit will be created and tagged
on it, after which the new branch ref will be deleted. The parent of this
commit will be the ref whose name is found in the .branch file.
This functionality is useful, for example, when a bugfix release is made for an older version of the software which has already been used as the baseline for a newer version. See the demo for an example.
If you want directories to be processed in non-alphanumerical order,
you may create a file in the input directory called mkrepo.order
containing a list of the directories/tags, one per line, in the order you
want them processed.