ROX Desktop

2009/4/22 Bernard Jungen <bjung50169@...>:
> Hi list,
>
> I've been playing with rox source code lately and came up with 3 patches to be
> found as attachments. These have been generated from my local git repo, having
> failed to generate a useable fork at repo.or.cz.

You might just need to push to it. I can't remember, but they might
start off empty. Did you get an error?

> Included in one patch is the ability to make SendTo items for regular files
> only, as is done for "grouped" files. There doesn't seem to be any other way
> to do it.

I've put them here for people to play with:

http://repo.or.cz/w/rox-filer.git?a=shortlog;h=refs/heads/bj-sendto

I'm not sure the compile option on the menu is a good idea, though.
For most users, the directory will be read-only and this option will
just generate errors. There's a Makefile in src which builds it...
surely every developer has a short-cut for "save-and-run-make"?


-- 
Dr Thomas Leonard		ROX desktop / Zero Install
GPG: 9242 9807 C985 3C07 44A6  8B9A AE07 8280 59A5 3CC1

On Sat, Apr 25, 2009 at 03:37:31PM +0100, Thomas Leonard wrote:
> 2009/4/22 Bernard Jungen <bjung50169@...>:
> > Hi list,
> >
> > I've been playing with rox source code lately and came up with 3 patches to be
> > found as attachments. These have been generated from my local git repo, having
> > failed to generate a useable fork at repo.or.cz.
> 
> You might just need to push to it. I can't remember, but they might
> start off empty. Did you get an error?

Cloning it downloads and then gives me this message:
   warning: remote HEAD refers to nonexistent ref, unable to checkout.

Unusable repo.
Apparently there's a problem at repo.or.cz not initializing forks correctly.
Anyone else has a clue?

Otherwise, how much more time-wasting is that for you to integrate generated patches?

> I'm not sure the compile option on the menu is a good idea, though.
> For most users, the directory will be read-only and this option will
> just generate errors. There's a Makefile in src which builds it...
> surely every developer has a short-cut for "save-and-run-make"?

Of course you're right. Let's say that first patch was a test in using git :-)

Cheers,

Bernard.

On Sat, 25 Apr 2009 15:37:31 +0100
Thomas Leonard <talex5@...> wrote:

> 2009/4/22 Bernard Jungen <bjung50169@...>:
> > Hi list,
> >
> > I've been playing with rox source code lately and came up with 3 patches to be
> > found as attachments. These have been generated from my local git repo, having
> > failed to generate a useable fork at repo.or.cz.
> 
> You might just need to push to it. I can't remember, but they might
> start off empty. Did you get an error?

In such cases, I think it will be useful to have the "mob"
feature enabled:

http://repo.or.cz/mob.html

Ciao
-- 
Nicola

2009/4/25 Bernard Jungen <bjung50169@...>:
> On Sat, Apr 25, 2009 at 03:37:31PM +0100, Thomas Leonard wrote:
>> 2009/4/22 Bernard Jungen <bjung50169@...>:
>> > Hi list,
>> >
>> > I've been playing with rox source code lately and came up with 3 patches to be
>> > found as attachments. These have been generated from my local git repo, having
>> > failed to generate a useable fork at repo.or.cz.
>>
>> You might just need to push to it. I can't remember, but they might
>> start off empty. Did you get an error?
>
> Cloning it downloads and then gives me this message:
>   warning: remote HEAD refers to nonexistent ref, unable to checkout.

Try cloning the original (my copy) and then pushing that to your fork. e.g.

git clone git://repo.or.cz/rox-filer.git
cd rox-filer
git remote add mine ssh://repo.or.cz/srv/git/rox-filer/bju.git
git push mine --all

> Otherwise, how much more time-wasting is that for you to integrate generated patches?

The only problem is that if two people apply your patches then GIT
will see that as two separate commits, whereas if we both pull from
your repository then it treats it as a single commit.


-- 
Dr Thomas Leonard		ROX desktop / Zero Install
GPG: 9242 9807 C985 3C07 44A6  8B9A AE07 8280 59A5 3CC1

Quoth Thomas Leonard <talex5@...>:
> 2009/4/25 Bernard Jungen <bjung50169@...>:
> 
> > Otherwise, how much more time-wasting is that for you to integrate
> generated patches?
> 
> The only problem is that if two people apply your patches then GIT
> will see that as two separate commits, whereas if we both pull from
> your repository then it treats it as a single commit.

I thought that if the patches were generated with git-format-patch and
applied with git-am then it was equivalent to a pull? (I thought that
was the whole point.)

Ben

On Sat, Apr 25, 2009 at 07:56:21PM +0100, Thomas Leonard wrote:
> 2009/4/25 Bernard Jungen <bjung50169@...>:
> > On Sat, Apr 25, 2009 at 03:37:31PM +0100, Thomas Leonard wrote:
> >> 2009/4/22 Bernard Jungen <bjung50169@...>:
> >> > Hi list,
> >> >
> >> > I've been playing with rox source code lately and came up with 3 patches to be
> >> > found as attachments. These have been generated from my local git repo, having
> >> > failed to generate a useable fork at repo.or.cz.
> >>
> >> You might just need to push to it. I can't remember, but they might
> >> start off empty. Did you get an error?
> >
> > Cloning it downloads and then gives me this message:
> >   warning: remote HEAD refers to nonexistent ref, unable to checkout.
> 
> Try cloning the original (my copy) and then pushing that to your fork. e.g.
> 
> git clone git://repo.or.cz/rox-filer.git
> cd rox-filer
> git remote add mine ssh://repo.or.cz/srv/git/rox-filer/bju.git
> git push mine --all

Fantastic, that works! You were right, all I needed to do was to push my
original (local) clone to my remote fork.

Here's another patch for rox-filer then:
  http://repo.or.cz/w/rox-filer/bju.git?a=commit;h=a823fca07a3670b378f4374754d7c4d863438481
    Show check sign in Display menu for sort type

Cheers,

Bernard.

2009/4/25 Ben Morrow <ben@...>:
> Quoth Thomas Leonard <talex5@...>:
>> 2009/4/25 Bernard Jungen <bjung50169@...>:
>>
>> > Otherwise, how much more time-wasting is that for you to integrate
>> generated patches?
>>
>> The only problem is that if two people apply your patches then GIT
>> will see that as two separate commits, whereas if we both pull from
>> your repository then it treats it as a single commit.
>
> I thought that if the patches were generated with git-format-patch and
> applied with git-am then it was equivalent to a pull? (I thought that
> was the whole point.)

There seem to be three different ways of sending patches with GIT:

git-diff / git-apply : like "svn diff" and "patch". No log message or
author information (only changes working copy), so tedious to apply
because I have to commit them myself.

git-format-patch / git-am : A series of patches with author
information and commit messages. Gets added as a series of commits
against my HEAD, where the Author is the original author and Committer
is me. Quite easy to use, but if two people commit them they look like
separate commits in the history. The original author then has to
discard their commits and use mine instead (fairly easy to do,
though).

git-push / git-pull : The exact commit objects are shared. Author and
Committer are the original author.

Personally, I prefer the last option because it's the least work for
me, but git-format-patch works well too.


-- 
Dr Thomas Leonard		ROX desktop / Zero Install
GPG: 9242 9807 C985 3C07 44A6  8B9A AE07 8280 59A5 3CC1