libtool-patches
[Top][All Lists]
Advanced

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

Re: weird tags in libtool's experimental git repo


From: Ralf Wildenhues
Subject: Re: weird tags in libtool's experimental git repo
Date: Sun, 16 Mar 2008 12:24:06 +0100
User-agent: Mutt/1.5.17+20080114 (2008-01-14)

* Jim Meyering wrote on Sun, Mar 16, 2008 at 12:03:12PM CET:
> 
> How to clean-up:
>  - remote bogus tags and prune
>  - simply regenerate libtool's git repo from CVS
> 
> What would you like?
> I tend to prefer the latter, but people will have to clone again.

While we're at cleaning the repo, let's bring up two issues on this list
that I noted before.  I think the repo should be redone with these two
changes, but would like the opinion of the others:


First, the conversion to git routinely maps the login of the cvs
committer to the author of the git commit (helped with a map login to
name plus email address).  It would be nicer if the actual author of the
commit could be taken instead, if possible, to give due credit.  A good
approximation to that is to use the first person listed in the ChangeLog
entry.  (Note that git carries the notion of an author of a patch as
well as a committer; the latter should IMHO be inferred from the cvs
login.)

I've had quite good results with this conversion, when done after your
initial conversion (thanks BTW for help from Jim and the git list):

git filter-branch -d /dev/shm/t --tree-filter '
  line=`sed -n "s,^[12][90][0-9][0-9]-[0-1][0-9]-[0-3][0-9]  *\([A-Za-z].*\),\1,
                s,.*  \([A-Za-z].*\),\1,
                /./{
                        s/(tiny change)//
                        s/ *$//
                        p
                        q
                }" ChangeLog`
  author=`echo "$line" | sed "s, *[<(].*,,"`
  email=`echo "$line" | sed "s,[^,(]*[<(],<,; s/[)>].*/>/"`
  test -n "$author" && test -n "$email" && {
        GIT_AUTHOR_NAME="$author"
        GIT_AUTHOR_EMAIL="$email"
        export GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL
  }'


This leaves a couple of commits with bogus email addresses, and a couple
with typos in the address.  Not sure whether I want to address that.
However, some people then get multiple different attributions.  For
example, Gary committed using several different email addresses, which
would then show up in the Author: field.


The second issue is that some of the commit messages use a different
encoding and thus end up being displayed wrongly.  I have yet to
experiment with iconv or some other convertor for an automatic way to
fix this, but I'm sure it's possible.

Comments, opinions?

If you all agree on this, then I don't have a problem with fixing all
three issues in one go, as each of them will require the users to
re-clone the repo.

Thanks,
Ralf




reply via email to

[Prev in Thread] Current Thread [Next in Thread]