[Ros-kinect] Fwd: Integration update, or OH GOD WHAT DID YOU DO TO MY REPO

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[Ros-kinect] Fwd: Integration update, or OH GOD WHAT DID YOU DO TO MY REPO

rusu
Administrator
Relevant. Our patches are integrated it seems! We should sync.

Cheers,
Radu.



-------- Original Message --------
Subject: Integration update, or OH GOD WHAT DID YOU DO TO MY REPO
Date: Thu, 18 Nov 2010 23:27:16 -0800
From: Kyle Machulis <[hidden email]>
Reply-To: [hidden email]
To: [hidden email]



So, I just brought the following requests into master

- Pull Request 30 - ros-pkg-git - Motor/LED functions for the new API, multiple cameras
- Pull Request 34 - stephenlovegrove - 10-bit camera access
- Pull Request 2 - jpgr87 - cmake updates
- Pull Request 31 - stephenlovegrove - C# updates

Now, everyone gather around, it's git story time:

You may notice that the pull requests don't exactly seem to be closing themselves out, and that you usually get a nice
little note from me talking about what commit you actually came in at relative to the pull request commit. Basically,
we're trying to avoid the problem outlined at

http://blog.xebia.com/2010/09/20/git-workflow/

Where you get a VERY nasty history and because nigh unto unusable when it comes to trying to do bisects and other handy
things for finding bugs and issues. The rate we're receiving pull requests at right now means actually kicking back all
requests until they're pulled up to the head of master also means we'd get very little if anything in quickly, because
we'd have to bring one commit in then wait for everyone else to pull up to the head.

I could queue-ize the pull requests and warn whoever is up next that they need to rebase before I can bring their stuff
in, but, for the moment, it's faster for me to do it (and yes I realize how many bad things have come out of this phrase
being said), and I am doing my very damnedest not to change any code here outside of small conflict resolves. There may
be a few merge spots or fixes here or there, but otherwise it should come in pretty much clean, and if not I'll make a
note of it and let the merge requester fix if needed.

FOR THOSE OF YOU COMMITTING TO YOUR MASTER BRANCH: This is probably going to turn violent for you very quickly as you
try to pull from our master back into yours. If you find yourself having problems, please let me know on this thread and
we'll try to get a workflow together ASAP. Or, check out that link above and their recommended workflow. If you don't
understand it, ask in this thread or the IRC channel.

FOR THOSE OF YOU THAT ALREADY KNOW HOW TO USE GIT, PLEASE SPEAK UP AND HELP OUT. Seriously, this is WAY better than
having to do mass integration via SVN, but it's still tough for a lot of people who just want to get their code in, and
I can only answer so many questions while also trying to integrate and work on the project myself. I would seriously
appreciate any help anyone could pitch in on the support side.

If you see anything go in that's completely wrong or that I screwed up, or think this is a completely stupid way of
doing integration, feel free to speak up here. I'm really just trying to make our history usable, make sure everyone
gets their credit their due as commiters, and keep new stuff coming in as quick as possible without everyone having to
learn git first. This is quickly turning into the largest project I've worked as integrator on though, so I'm learning
here too, and am happy to take advice or criticism, either here or on the IRC channel.

Other than that, keep the changes comin' and I'll try to keep up!

Kyle
_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect
Reply | Threaded
Open this post in threaded view
|

Re: [Ros-kinect] Fwd: Integration update, or OH GOD WHAT DID YOU DO TO MY REPO

Tully Foote
I've taken care of it. 

As can be seen in this emali there was a bit of a mess with integrating the upstream patches.  I have managed to update our fork of the repo without breaking anything.  However for better future work, I have changed the default branch to "ros" and if you have any local checkouts please switch to the ros branch as well.  We will continue developing there, pulling in changes from master as they occur and submitting pull requests either from the ros branch or a custom branch made just for pull requests.  That way we won't get ahead of the master branch and possibly have terrible conflicts when pulling. 

I've also updated the README with more extensive instructions on this and other things. 

Also please remember to sign commits into our fork of libfreenect so they are easy to push back. 

Tully

On Fri, Nov 19, 2010 at 9:17 AM, Radu Bogdan Rusu <[hidden email]> wrote:
Relevant. Our patches are integrated it seems! We should sync.

Cheers,
Radu.



-------- Original Message --------
Subject:        Integration update, or OH GOD WHAT DID YOU DO TO MY REPO
Date:   Thu, 18 Nov 2010 23:27:16 -0800
From:   Kyle Machulis <[hidden email]>
Reply-To:       [hidden email]
To:     [hidden email]



So, I just brought the following requests into master

- Pull Request 30 - ros-pkg-git - Motor/LED functions for the new API, multiple cameras
- Pull Request 34 - stephenlovegrove - 10-bit camera access
- Pull Request 2 - jpgr87 - cmake updates
- Pull Request 31 - stephenlovegrove - C# updates

Now, everyone gather around, it's git story time:

You may notice that the pull requests don't exactly seem to be closing themselves out, and that you usually get a nice
little note from me talking about what commit you actually came in at relative to the pull request commit. Basically,
we're trying to avoid the problem outlined at

http://blog.xebia.com/2010/09/20/git-workflow/

Where you get a VERY nasty history and because nigh unto unusable when it comes to trying to do bisects and other handy
things for finding bugs and issues. The rate we're receiving pull requests at right now means actually kicking back all
requests until they're pulled up to the head of master also means we'd get very little if anything in quickly, because
we'd have to bring one commit in then wait for everyone else to pull up to the head.

I could queue-ize the pull requests and warn whoever is up next that they need to rebase before I can bring their stuff
in, but, for the moment, it's faster for me to do it (and yes I realize how many bad things have come out of this phrase
being said), and I am doing my very damnedest not to change any code here outside of small conflict resolves. There may
be a few merge spots or fixes here or there, but otherwise it should come in pretty much clean, and if not I'll make a
note of it and let the merge requester fix if needed.

FOR THOSE OF YOU COMMITTING TO YOUR MASTER BRANCH: This is probably going to turn violent for you very quickly as you
try to pull from our master back into yours. If you find yourself having problems, please let me know on this thread and
we'll try to get a workflow together ASAP. Or, check out that link above and their recommended workflow. If you don't
understand it, ask in this thread or the IRC channel.

FOR THOSE OF YOU THAT ALREADY KNOW HOW TO USE GIT, PLEASE SPEAK UP AND HELP OUT. Seriously, this is WAY better than
having to do mass integration via SVN, but it's still tough for a lot of people who just want to get their code in, and
I can only answer so many questions while also trying to integrate and work on the project myself. I would seriously
appreciate any help anyone could pitch in on the support side.

If you see anything go in that's completely wrong or that I screwed up, or think this is a completely stupid way of
doing integration, feel free to speak up here. I'm really just trying to make our history usable, make sure everyone
gets their credit their due as commiters, and keep new stuff coming in as quick as possible without everyone having to
learn git first. This is quickly turning into the largest project I've worked as integrator on though, so I'm learning
here too, and am happy to take advice or criticism, either here or on the IRC channel.

Other than that, keep the changes comin' and I'll try to keep up!

Kyle
_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect



--
Tully Foote
Systems Engineer
Willow Garage, Inc.
[hidden email]
(650) 475-2827

_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect
Reply | Threaded
Open this post in threaded view
|

Re: [Ros-kinect] Fwd: Integration update, or OH GOD WHAT DID YOU DO TO MY REPO

Ivan Dryanovski
Just for clarification, this does not apply to ros-pkg-git/kinect, correct?

Ivan


On Sat, Nov 20, 2010 at 3:51 AM, Tully Foote <[hidden email]> wrote:

> I've taken care of it.
>
> As can be seen in this emali there was a bit of a mess with integrating the
> upstream patches.  I have managed to update our fork of the repo without
> breaking anything.  However for better future work, I have changed the
> default branch to "ros" and if you have any local checkouts please switch to
> the ros branch as well.  We will continue developing there, pulling in
> changes from master as they occur and submitting pull requests either from
> the ros branch or a custom branch made just for pull requests.  That way we
> won't get ahead of the master branch and possibly have terrible conflicts
> when pulling.
>
> I've also updated the README with more extensive instructions on this and
> other things.
>
> Also please remember to sign commits into our fork of libfreenect so they
> are easy to push back.
>
> Tully
>
> On Fri, Nov 19, 2010 at 9:17 AM, Radu Bogdan Rusu <[hidden email]>
> wrote:
>>
>> Relevant. Our patches are integrated it seems! We should sync.
>>
>> Cheers,
>> Radu.
>>
>>
>>
>> -------- Original Message --------
>> Subject:        Integration update, or OH GOD WHAT DID YOU DO TO MY REPO
>> Date:   Thu, 18 Nov 2010 23:27:16 -0800
>> From:   Kyle Machulis <[hidden email]>
>> Reply-To:       [hidden email]
>> To:     [hidden email]
>>
>>
>>
>> So, I just brought the following requests into master
>>
>> - Pull Request 30 - ros-pkg-git - Motor/LED functions for the new API,
>> multiple cameras
>> - Pull Request 34 - stephenlovegrove - 10-bit camera access
>> - Pull Request 2 - jpgr87 - cmake updates
>> - Pull Request 31 - stephenlovegrove - C# updates
>>
>> Now, everyone gather around, it's git story time:
>>
>> You may notice that the pull requests don't exactly seem to be closing
>> themselves out, and that you usually get a nice
>> little note from me talking about what commit you actually came in at
>> relative to the pull request commit. Basically,
>> we're trying to avoid the problem outlined at
>>
>> http://blog.xebia.com/2010/09/20/git-workflow/
>>
>> Where you get a VERY nasty history and because nigh unto unusable when it
>> comes to trying to do bisects and other handy
>> things for finding bugs and issues. The rate we're receiving pull requests
>> at right now means actually kicking back all
>> requests until they're pulled up to the head of master also means we'd get
>> very little if anything in quickly, because
>> we'd have to bring one commit in then wait for everyone else to pull up to
>> the head.
>>
>> I could queue-ize the pull requests and warn whoever is up next that they
>> need to rebase before I can bring their stuff
>> in, but, for the moment, it's faster for me to do it (and yes I realize
>> how many bad things have come out of this phrase
>> being said), and I am doing my very damnedest not to change any code here
>> outside of small conflict resolves. There may
>> be a few merge spots or fixes here or there, but otherwise it should come
>> in pretty much clean, and if not I'll make a
>> note of it and let the merge requester fix if needed.
>>
>> FOR THOSE OF YOU COMMITTING TO YOUR MASTER BRANCH: This is probably going
>> to turn violent for you very quickly as you
>> try to pull from our master back into yours. If you find yourself having
>> problems, please let me know on this thread and
>> we'll try to get a workflow together ASAP. Or, check out that link above
>> and their recommended workflow. If you don't
>> understand it, ask in this thread or the IRC channel.
>>
>> FOR THOSE OF YOU THAT ALREADY KNOW HOW TO USE GIT, PLEASE SPEAK UP AND
>> HELP OUT. Seriously, this is WAY better than
>> having to do mass integration via SVN, but it's still tough for a lot of
>> people who just want to get their code in, and
>> I can only answer so many questions while also trying to integrate and
>> work on the project myself. I would seriously
>> appreciate any help anyone could pitch in on the support side.
>>
>> If you see anything go in that's completely wrong or that I screwed up, or
>> think this is a completely stupid way of
>> doing integration, feel free to speak up here. I'm really just trying to
>> make our history usable, make sure everyone
>> gets their credit their due as commiters, and keep new stuff coming in as
>> quick as possible without everyone having to
>> learn git first. This is quickly turning into the largest project I've
>> worked as integrator on though, so I'm learning
>> here too, and am happy to take advice or criticism, either here or on the
>> IRC channel.
>>
>> Other than that, keep the changes comin' and I'll try to keep up!
>>
>> Kyle
>> _______________________________________________
>> Ros-kinect mailing list
>> [hidden email]
>> https://code.ros.org/mailman/listinfo/ros-kinect
>
>
>
> --
> Tully Foote
> Systems Engineer
> Willow Garage, Inc.
> [hidden email]
> (650) 475-2827
>
> _______________________________________________
> Ros-kinect mailing list
> [hidden email]
> https://code.ros.org/mailman/listinfo/ros-kinect
>
>
_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect
Reply | Threaded
Open this post in threaded view
|

Re: [Ros-kinect] Fwd: Integration update, or OH GOD WHAT DID YOU DO TO MY REPO

Tully Foote

Correct.  We should think about that sort of thing. But since our contributor list is smaller its much less of an issue.

Tully

On Nov 20, 2010 4:04 PM, "Ivan Dryanovski" <[hidden email]> wrote:
> Just for clarification, this does not apply to ros-pkg-git/kinect, correct?
>
> Ivan
>
>
> On Sat, Nov 20, 2010 at 3:51 AM, Tully Foote <[hidden email]> wrote:
>> I've taken care of it.
>>
>> As can be seen in this emali there was a bit of a mess with integrating the
>> upstream patches.  I have managed to update our fork of the repo without
>> breaking anything.  However for better future work, I have changed the
>> default branch to "ros" and if you have any local checkouts please switch to
>> the ros branch as well.  We will continue developing there, pulling in
>> changes from master as they occur and submitting pull requests either from
>> the ros branch or a custom branch made just for pull requests.  That way we
>> won't get ahead of the master branch and possibly have terrible conflicts
>> when pulling.
>>
>> I've also updated the README with more extensive instructions on this and
>> other things.
>>
>> Also please remember to sign commits into our fork of libfreenect so they
>> are easy to push back.
>>
>> Tully
>>
>> On Fri, Nov 19, 2010 at 9:17 AM, Radu Bogdan Rusu <[hidden email]>
>> wrote:
>>>
>>> Relevant. Our patches are integrated it seems! We should sync.
>>>
>>> Cheers,
>>> Radu.
>>>
>>>
>>>
>>> -------- Original Message --------
>>> Subject:        Integration update, or OH GOD WHAT DID YOU DO TO MY REPO
>>> Date:   Thu, 18 Nov 2010 23:27:16 -0800
>>> From:   Kyle Machulis <[hidden email]>
>>> Reply-To:       [hidden email]
>>> To:     [hidden email]
>>>
>>>
>>>
>>> So, I just brought the following requests into master
>>>
>>> - Pull Request 30 - ros-pkg-git - Motor/LED functions for the new API,
>>> multiple cameras
>>> - Pull Request 34 - stephenlovegrove - 10-bit camera access
>>> - Pull Request 2 - jpgr87 - cmake updates
>>> - Pull Request 31 - stephenlovegrove - C# updates
>>>
>>> Now, everyone gather around, it's git story time:
>>>
>>> You may notice that the pull requests don't exactly seem to be closing
>>> themselves out, and that you usually get a nice
>>> little note from me talking about what commit you actually came in at
>>> relative to the pull request commit. Basically,
>>> we're trying to avoid the problem outlined at
>>>
>>> http://blog.xebia.com/2010/09/20/git-workflow/
>>>
>>> Where you get a VERY nasty history and because nigh unto unusable when it
>>> comes to trying to do bisects and other handy
>>> things for finding bugs and issues. The rate we're receiving pull requests
>>> at right now means actually kicking back all
>>> requests until they're pulled up to the head of master also means we'd get
>>> very little if anything in quickly, because
>>> we'd have to bring one commit in then wait for everyone else to pull up to
>>> the head.
>>>
>>> I could queue-ize the pull requests and warn whoever is up next that they
>>> need to rebase before I can bring their stuff
>>> in, but, for the moment, it's faster for me to do it (and yes I realize
>>> how many bad things have come out of this phrase
>>> being said), and I am doing my very damnedest not to change any code here
>>> outside of small conflict resolves. There may
>>> be a few merge spots or fixes here or there, but otherwise it should come
>>> in pretty much clean, and if not I'll make a
>>> note of it and let the merge requester fix if needed.
>>>
>>> FOR THOSE OF YOU COMMITTING TO YOUR MASTER BRANCH: This is probably going
>>> to turn violent for you very quickly as you
>>> try to pull from our master back into yours. If you find yourself having
>>> problems, please let me know on this thread and
>>> we'll try to get a workflow together ASAP. Or, check out that link above
>>> and their recommended workflow. If you don't
>>> understand it, ask in this thread or the IRC channel.
>>>
>>> FOR THOSE OF YOU THAT ALREADY KNOW HOW TO USE GIT, PLEASE SPEAK UP AND
>>> HELP OUT. Seriously, this is WAY better than
>>> having to do mass integration via SVN, but it's still tough for a lot of
>>> people who just want to get their code in, and
>>> I can only answer so many questions while also trying to integrate and
>>> work on the project myself. I would seriously
>>> appreciate any help anyone could pitch in on the support side.
>>>
>>> If you see anything go in that's completely wrong or that I screwed up, or
>>> think this is a completely stupid way of
>>> doing integration, feel free to speak up here. I'm really just trying to
>>> make our history usable, make sure everyone
>>> gets their credit their due as commiters, and keep new stuff coming in as
>>> quick as possible without everyone having to
>>> learn git first. This is quickly turning into the largest project I've
>>> worked as integrator on though, so I'm learning
>>> here too, and am happy to take advice or criticism, either here or on the
>>> IRC channel.
>>>
>>> Other than that, keep the changes comin' and I'll try to keep up!
>>>
>>> Kyle
>>> _______________________________________________
>>> Ros-kinect mailing list
>>> [hidden email]
>>> https://code.ros.org/mailman/listinfo/ros-kinect
>>
>>
>>
>> --
>> Tully Foote
>> Systems Engineer
>> Willow Garage, Inc.
>> [hidden email]
>> (650) 475-2827
>>
>> _______________________________________________
>> Ros-kinect mailing list
>> [hidden email]
>> https://code.ros.org/mailman/listinfo/ros-kinect
>>
>>
> _______________________________________________
> Ros-kinect mailing list
> [hidden email]
> https://code.ros.org/mailman/listinfo/ros-kinect

_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect