Re: [PATCH] mktag.c: improve verification of tagger field and tests
- From
Brandon Casey <casey@nrlssc.navy.mil>
- Date
- Mar 26, 2008, 16:26 UTC
- Message-ID
- <47EA7927.9060201@nrlssc.navy.mil>
- In-Reply-To
- <1b46aba20803260421t4db4987gc6fc8b2e556032e0@mail.gmail.com>
Carlos Rica wrote:
Show 8 quoted lines
> On Wed, Mar 26, 2008 at 1:40 AM, Brandon Casey <casey@nrlssc.navy.mil> wrote: >> Well, since I looked at this code and used it in filter-branch, I figured I >> should fix the verification code for the tagger field (even though it's >> probably dieing soon). >> >> I'm thinking this utility should be fairly strict about the format it accepts. > > Why not using git-tag to make tags in filter-branch?
How?
With cat-file and mktag I don't have to parse the tag, I can ignore nearly the entire thing and only change the parts I'm interested in.
With git-tag, wouldn't I have to parse the tag, splitting the header from the body and redirect the body to a tmpfile, parse the tagger field, and set GIT_COMMITTER_NAME, GIT_COMMITTER_EMAIL, GIT_COMMITTER_DATE environment variables?
> git-mktag was used in git-tag.sh before convert it into > builtin-tag.c, and I didn't know that anyone was using it.
I don't think anything other than filter-branch is (and the usage there is only tentative).
Show 8 quoted lines
> I agree that, if this program exists and it is used, > we should double-check the accepted format and data, > so this patch is a good addition. > > However, I think that we should progressively deprecate its > use to avoid mantaining two different ways for creating tags, > so you must have a very good reason to keep using > this tool in a script...
Eventually filter-branch along with the rest of git will be rewritten in c and the need for many low-level git tools will vanish.
-brandon