On 2024-06-30 at 21:33:33, Mikko Johannes Koivunalho wrote:

I want to create part or all of the `git commit -m *` command's
message with a script. I work for one Jira ticket at a time and I
need to create many commits which all start with "TKT-123: "
(any Jira ticket number/identifier). To make the commits faster,
I run `git commit --message="TKT-123: add new file"` on the command
line.

I want the Bash completion mechanism to propose the message for
me. I would fetch the newest ticket number and place it on the command
line when I type `git commit --message=<TAB>`.

Example:
# Doing commit:
git commit --message=<TAB>
# you would get (also without the closing double quote):
git commit --message="ABC-1234

This is easy to do with a prepare-commit-msg or commit-msg hook, and
those are the intended tools for this purpose.  I've used these hooks to
generate a message for a ticket based on the branch name at a past
company (so a branch called tkt-123 would result in the TKT-123: entry
in the appropriate place in the commit message).

While you certainly can commit on the command line, it's not encouraged
because you're supposed to write a commit message that explains the
commit in detail.  Only very rarely is a single line commit message
useful, and even in the case you've cited, I'd want to know why you
added a new file.  What does the file do?  What problem is it supposed
to solve?  Why are we adding a new file when we could add the changes to
an existing file?  Why is this change valuable at all?

That being said, all of this is possible with a prepare-commit-msg hook,
and that's a better and more generic way than using the bash completion,
which is specific to bash.
-- 
brian m. carlson (they/them or he/him)
Toronto, Ontario, CA