On Sunday, June 2, 2013 3:58:11 AM UTC+2, Josh wrote:
> On Jun 1, 2013, at 5:07 PM, Marc Weber <[email protected]> wrote:
> 
> 
> 
> > I've been thinking about what's causing trouble to me.
> 
> > Lack of workflows - lack of documentation.
> 
> > 
> 
> > http://vim.sourceforge.net/ -> download -> patch
> 
> > No hint about how patches should be submitted.
> 
> > No hinting about release cycles
> 
> > NO hinting about how to manage vim source and patches.
> 
> > 
> 
> > I'm used to git. So let me explain how I do it for git(hub) based
> 
> > projects:
> 
> > 
> 
> > 1) submitting patches
> 
> > 
> 
> >  1) clone
> 
> >  2) create topgit based branch which basically means having a command
> 
> >  to merge upstream changes automatically. I can export this branch as a
> 
> >    single patch when I feel ready. mercurial has the [pbranch] extension
> 
> >    which also works perfectly fine
> 
> >  3) push my patch to github, create pull request
> 
> > 
> 
> >  If I have an updated patch I just "push -f" overriding history,
> 
> >  and github will tell show to the rest of world "this patch has an
> 
> >  outdated version people already commented on. Click here to read all
> 
> >  about the older patch
> 
> > 
> 
> >  This way its easy to react to comments (which on github can be placed
> 
> >  nearby lines) as well as below the pull request description
> 
> > 
> 
> > 2) reviewing patches, applying them
> 
> >  way 1): curl 'patch' | git am
> 
> >  way 2): add a remote location
> 
> >    git add remote contributor github://the-url
> 
> >    git fetch
> 
> > 
> 
> >  This way works like "hg pull other-repo-url"
> 
> >  However there is one big difference:
> 
> >  All branches coming from that repo are prefixed by that repo name,
> 
> >  in this example contributor/master contributor/feature-X
> 
> > 
> 
> >  Thus its not me having to worry about importing foreign commits into
> 
> >  my repository - because I can always can identify and drop them again,
> 
> >  and distinguish them from my own branches
> 
> > 
> 
> >  People can vote for patches by commenting by "+1" (yes, it sucks, but at
> 
> >  least you can vote). And for release managers it is important to see
> 
> >  how many people want a patch/feature because resourcese are altways
> 
> >  limited. This way it can be made transparent.
> 
> > 
> 
> >  Patches even can be labeled such as
> 
> >  "bram_wants_this_patch_be_patient_till_he_has_reviewed_it"
> 
> >  or
> 
> >  "idea_nice_code_quality_should_be_improved"
> 
> > 
> 
> >  Then people know what to do.
> 
> > 
> 
> > This all can be documented nicely, eg that's what we did for nixos:
> 
> > https://nixos.org/wiki/Contributing
> 
> > 
> 
> > So let me compare this to Vim:
> 
> > Even the [developing] page stops at "how to get latest version of Vim"
> 
> > not talking about how to
> 
> >  - design new proposals
> 
> >  - submitting them
> 
> > 
> 
> > Summary: Almost all problems would be gone by providing an official
> 
> > github based mirror using git. Patches which are ready could still be
> 
> > applied to mercurial.
> 
> > 
> 
> > I know there are many mercurial users on this list, please helsp my mind
> 
> > change, and tell me how the I can setup a mercurial based workflow which
> 
> > doesn't get into my way - which is as easy as the git based is I
> 
> > described above.
> 
> > 
> 
> > If there is no alternative, would other people join and help maintaining
> 
> > a git based mirror? Bram, would you even eventually consider looking
> 
> > there for most up to date patches? We could still announce them on the
> 
> > mailinglist.
> 
> > 
> 
> > People like ZyX should not spend their time on repating lists of patches
> 
> > which should be included. Such behaviour might be a sign of lack of
> 
> > infrastructure - and that infrastructure is there, we just have to use
> 
> > it.
> 
> > 
> 
> > Yes, I know that you can clone repositories on google code, too.
> 
> > However there is no way for people telling "I'm ready with a patch and I
> 
> > want that other people discuss and comment on it", thus something like
> 
> > pull requests.
> 
> > 
> 
> > Thus how many people would be interested in a github based workflow,
> 
> > because it just solves quite a lot of problem getting out of the way of
> 
> > devs letting them do what they should do: write code, comment code and
> 
> > get things done?
> 
> > 
> 
> > So what about using an github based mirror? Who else would be interested
> 
> > in it?
> 
> > 
> 
> > The mailinglist is not appropriate, because its hard to keep track of
> 
> > patches/ideas
> 
> > - to be discussed
> 
> > - to be reviewed
> 
> > - closed (merged, rejected?)
> 
> > 
> 
> > Marc Weber
> 
> > 
> 
> > links:
> 
> > [pbranch]: http://mercurial.selenic.com/wiki/PatchBranchExtension
> 
> > [developing]: http://vim.sourceforge.net/develop.php
> 
> > 
> 
> > -- 
> 
> > -- 
> 
> > You received this message from the "vim_dev" maillist.
> 
> > Do not top-post! Type your reply below the text you are replying to.
> 
> > For more information, visit http://www.vim.org/maillist.php
> 
> > 
> 
> > --- 
> 
> > You received this message because you are subscribed to the Google Groups 
> > "vim_dev" group.
> 
> > To unsubscribe from this group and stop receiving emails from it, send an 
> > email to [email protected].
> 
> > For more options, visit https://groups.google.com/groups/opt_out.
> 
> > 
> 
> > 
> 
> 
> 
> +1 maybe we can get vim on GH and travis-ci!
> 
> 
> 
> (Ps please let me know if I'm posting incorrectly to the ml)

+1 for GH

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Raspunde prin e-mail lui