Patch Management

[Santhosh Edukulla <sa...@citrix.com>]

Team,

Currently, it seems we have few patches missing in master but available in current running branch, could be attributed for various reasons.

Its hard to track some times that all changes made to make issue fixes work, are all available in master at any given time. I believe we can have an entry gate, that if the patch does not apply cleanly on either of the branch(  and if it is supposed for both ), then its better we don't push it to both the branches, until the clean patch for both master and running branch are submitted again.

This way, review will not be closed to submission. Otherwise, from bug fixes perspective, fixes should go hand in hand to both master and running branch i believe. Tracking manually through review\git log for missing patches may be little tedious and error prone again.

You can see if there is a better way to track patch to completion for both and is proper.

Thanks!
Santhosh

RE: Patch Management

[Santhosh Edukulla <sa...@citrix.com>]

Iam not sure, its stretching a limit, but i have seen many patches missing in master or only applied to one branch. 

One RCA for many failures or issues reported during initial days of every new branch is that it has these fixes are missing. I believe if we can enforce a patch submitter to provide working patch for atleast master and concerned branch, we should not put closure to bug, review and patch submission i believe. Or have review board submissions considered "only" containing patches for both master and individual branch.

Santhosh
________________________________________
From: Daan Hoogland [daan.hoogland@gmail.com]
Sent: Saturday, May 17, 2014 5:06 AM
To: dev
Subject: Re: Patch Management

plus one; the review board does not easily facilitate multiple patches
so I would use the 'depend on' feature to point to the other
branch-patches

On Thu, May 15, 2014 at 9:18 AM, Santhosh Edukulla
<sa...@citrix.com> wrote:
> Team,
>
> Currently, it seems we have few patches missing in master but available in current running branch, could be attributed for various reasons.
>
> Its hard to track some times that all changes made to make issue fixes work, are all available in master at any given time. I believe we can have an entry gate, that if the patch does not apply cleanly on either of the branch(  and if it is supposed for both ), then its better we don't push it to both the branches, until the clean patch for both master and running branch are submitted again.
>
> This way, review will not be closed to submission. Otherwise, from bug fixes perspective, fixes should go hand in hand to both master and running branch i believe. Tracking manually through review\git log for missing patches may be little tedious and error prone again.
>
> You can see if there is a better way to track patch to completion for both and is proper.
>
> Thanks!
> Santhosh



--
Daan

Re: Patch Management

[Daan Hoogland <da...@gmail.com>]

plus one; the review board does not easily facilitate multiple patches
so I would use the 'depend on' feature to point to the other
branch-patches

On Thu, May 15, 2014 at 9:18 AM, Santhosh Edukulla
<sa...@citrix.com> wrote:
> Team,
>
> Currently, it seems we have few patches missing in master but available in current running branch, could be attributed for various reasons.
>
> Its hard to track some times that all changes made to make issue fixes work, are all available in master at any given time. I believe we can have an entry gate, that if the patch does not apply cleanly on either of the branch(  and if it is supposed for both ), then its better we don't push it to both the branches, until the clean patch for both master and running branch are submitted again.
>
> This way, review will not be closed to submission. Otherwise, from bug fixes perspective, fixes should go hand in hand to both master and running branch i believe. Tracking manually through review\git log for missing patches may be little tedious and error prone again.
>
> You can see if there is a better way to track patch to completion for both and is proper.
>
> Thanks!
> Santhosh



-- 
Daan