[TYPO3-dev] The future of Bug Days

[Philipp Gampe]

Mathias Schreiber [wmdb] wrote:

> Am 22.07.11 15:10, schrieb JoH asenau:
>> Hi Francois.
>>
>> Just a wild guess, but could it be that the decrease of particpiants is
>> not due to routine, but just the contrary: Missing routine with the new
>> git/gerrit-concept, which is at least overly complicated to set up
>> compared to the "old school" SVN, people are used to?
Might be, but those who have the routine already are much more productive.

A review for me is
$ git co <branch> (alias for git checkout <branch>)
$ git r 1234 [*]
$ git diff HEAD^ (I aliased this to git dh)
... testing ...
$ git ra [*]
$ git rr [*]

Can you beat this with svn?

If you still have problems, maybe we need to update/extend the docs?
http://wiki.typo3.org/Git

>> There are lots of people - especially the windows users - including
>> myself, who didn't manage to enable their system yet to do a simple push
>> from PHP-storm onto the review system. So for these people it doesn't
>> make much sense to participate in a bug day.
Use the command line. There is no way to put git into a GUI, unless you 
either clutter the interface or leave out all feature.
I don't think the push problem is a Windows+PhpStorm issue, but a PhpStorm 
issue in general. AFAIR, PhpStorm does not support the push hook which is 
need for gerrit.

> Plus I dislike the voting scheme in gerrit.
> If I want to I can block out every changeset by -1ing it to death.
> Or I do -1 because I have no idea about the intentions of a changeset
> are... how to remove this from gerrit afterwards? I haven't figured out
> how to do this yet :)
The author can abandon a patch.
You could have send -1 to the core list too. Nothing you can do about this. 
If somebody wants to bitch vote a patch, he can do it. There is no way to 
avoid this unless you want to limit the vote to trusted person.

> I do see the benefits of gerrit without question but I kinda miss the
> "last word" feature the old core list had.
It is still there. Any core member can just +2 and push the patch. I think 
he can also set -2 to block a patch.

> but all in all... I find GIT utterly horrible to use.
Start using it more. If I would need to work with a large svn project now, I 
would use git-svn. I don't want to miss the features of git any more:
* topic branches
* easy branch switching
* speed
* local history with all changes
* git bisec
* git rebase
* git stash
... 


[*] I use a shell script which does some thinks for me
$ git r 1234 (alias for git review 1234)
--> cherry-pick the latest patchset from gerrit issue 1234
--> you can copy paste an equivalent at the gerrit issue website
$ git ra (alias for git review approve)
--> vote +1 verified and +1 review
--> you can vote directly on the website too
$ git rr (alias for git review reset)
--> delete changes, clean up and return to previous branch

This review script is not ready for public, but already in "good" shape. You 
may get it here (git again):
https://github.com/pgampe/gerrit-tools

Install according to instruction and run git review help to see what you can 
do.

Best regards
-- 
Philipp Gampe