[ANNOUNCE] Cogito-0.16.2

[ANNOUNCE] Cogito-0.16.2

[Petr Baudis]

  Hello,

  this is Cogito version 0.16.2, the next stable release of the
human-friendly version control UI for the Linus' GIT tool. Share
and enjoy at:

	http://www.kernel.org/pub/software/scm/cogito/

  Very few changes since the last release - one testcase, trivial fix,
documentation improvements. And one important bugfix! If you are a
cg-clean user, certainly upgrade - cg-clean in 0.16.1 had a serious bug
which would cause it remove all subdirectories when running from the
source root with the -d option - which is rather nasty if you have any
uncommitted local changes there.

  Now we finally have a testcase for cg-clean, and the cg-clean code was
further tweaked to reduce potential for any more code mistakes. The
major reason for both 0.16.1 and 0.16.2 were cg-clean bugs, but
hopefully no more problems are waiting for us in this realm, and the
rest of Cogito seems to be thankfully pretty stable throughout the 0.16
line, so I expect things to finally calm down a bit here.

  I spent the last two hours making a large patch fixing all potentially
whitespace-unsafe Cogito code, but I've decided to keep it for 0.17 only
- Cogito meeting with whitespaces is a very rare occasion, and such
a large patch is bound to contain some new bugs coming from trivial
mistakes and such.

  Also, Cogito is now very ineffective when cloning big-packed
repositories over the git protocol or over ssh, since git-fetch-pack
unpacks the objects on the local side, which bogs things down a lot and
makes the repository grow into enormous proportions - I'm not yet sure
if I will backport the fix to 0.16 since I want to be maximally careful
not to gravely break anything again and keep 0.16 as stable as possible.
(Hm. Call this "the Debian dilemma". ;-)

  The new stuff since 0.16.1 is:

Pavel Roskin:
      cg-clean test
      cg-clean: be safer by running everything from relpath

Petr Baudis:
      cg-clean testcase: Minor tidyup
      cg-clean testcase: Clarify description
      Fixed cg-clean removing ALL subdirs when invoked from root
            (Discovered and originally fixed by Pavel Roskin.)
      Fix cg-status reporting every merge as a squash-merge
      Improve cg-admin-setuprepo documentation
      In the cg-commit text, give the user hint on how to abort
      cogito-0.16.2

  Happy hacking,

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
VI has two modes: the one in which it beeps and the one in which
it doesn't.

Re: [ANNOUNCE] Cogito-0.16.2

[Alan Chandler]

On Wednesday 14 December 2005 23:10, Petr Baudis wrote:
...
>   Also, Cogito is now very ineffective when cloning big-packed
> repositories over the git protocol or over ssh, since git-fetch-pack
> unpacks the objects on the local side, which bogs things down a lot and
> makes the repository grow into enormous proportions - I'm not yet sure
> if I will backport the fix to 0.16 since I want to be maximally careful
> not to gravely break anything again and keep 0.16 as stable as possible.
> (Hm. Call this "the Debian dilemma". ;-)

Note sure what your Debian dilemma is.  Right now, Debian (Unstable) seems 
broken, because cogito is still at 0.16.0 and hickups on the bash 3.1 issue 
discussed in an earlier thread.  Having updated to the debian version from my 
earlier self built version earlier this week, I immediately found I couldn't 
commit.

After waiting a day for the next release to filter through and it not doing 
so, I had to rebuild my own version again - and as far as I am aware, despite 
a new version of git-core arriving on my "update" this  morning, there has 
been no new cogito package.


-- 
Alan Chandler
http://www.chandlerfamily.org.uk
Open Source. It's the difference between trust and antitrust.

Re: [ANNOUNCE] Cogito-0.16.2

[Petr Baudis]

Dear diary, on Thu, Dec 15, 2005 at 07:57:52AM CET, I got a letter
where Alan Chandler <alan@chandlerfamily.org.uk> said that...
> On Wednesday 14 December 2005 23:10, Petr Baudis wrote:
> ...
> >   Also, Cogito is now very ineffective when cloning big-packed
> > repositories over the git protocol or over ssh, since git-fetch-pack
> > unpacks the objects on the local side, which bogs things down a lot and
> > makes the repository grow into enormous proportions - I'm not yet sure
> > if I will backport the fix to 0.16 since I want to be maximally careful
> > not to gravely break anything again and keep 0.16 as stable as possible.
> > (Hm. Call this "the Debian dilemma". ;-)
> 
> Note sure what your Debian dilemma is.

Well, I meant this in a general sense - in the stable Debian branch,
there's also this dilemma about how grave bugfixes to still merge,
without risking another bug induced by those.

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
VI has two modes: the one in which it beeps and the one in which
it doesn't.