| yes, we will need to keep an eye on PRs.
However, I noticed that all old PRs were closed when I replaced develop. I had to change the default branch to new-develop to delete it, then push a new branch
called develop, then finally change the default branch back to develop. This had the effect of closing all open PRs.
It might be the case that only PRs taken from the new develop will be accepted by the system.
Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy
Maynooth University
Ireland
> On 8 Sep 2018, at 16:50, Steven Yi wrote:
>
> I reviewed and merged. They were fixes relevant to MSYS2 build. I
> felt a bit anxious, but I did the following:
>
> 1. Reviewed contents of merge
> 2. Followed github instructions to do local test of merging changes.
> 3. Checked 'git rev-list --count HEAD' to ensure count did not explode
> 4. Checked gitk visualization to see parent commits.
> 5. Pulled trigger and used github merge.
>
> It probably would have been safer to push the merge from local to
> github rather than use Github's merge in step 5. Probably for the
> short term we should be very careful to double check PR's locally
> before any merges of the PR's.
>
>
>> On Sat, Sep 8, 2018 at 11:09 AM Victor Lazzarini wrote:
>>
>> Excellent.
>>
>> I noticed that there is a new PR, and it's to do with Windows. Would you or Stephen be able to review it?
>>
>> Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy
>> Maynooth University
>> Ireland
>>
>>> On 8 Sep 2018, at 15:46, Steven Yi wrote:
>>>
>>> Hi Victor,
>>>
>>> Thanks for all of this! There's an option to reset --hard a branch to
>>> origin/develop too but it's much safer to use a fresh clone. I'd also
>>> mention I just merged master into develop so that we won't have
>>> problems when we do the next release. (I tested locally merging
>>> develop back to master and it had no conflicts.)
>>>
>>> Just another note, to double check your local repos, you can always use:
>>>
>>> git rev-list --count HEAD
>>>
>>> and it will list the number of commits you have locally. It should be
>>> just over 15000 (bad states should see something like 44000)
>>>
>>> stevenOn Fri, Sep 7, 2018 at 5:09 PM Victor Lazzarini
>>> wrote:
>>>>
>>>> I have now completed the edits to git history in the develop branch. The old develop branch has been
>>>> saved as old-develop, and torched. In its place, I put a newly-created develop (also present as new-develop)
>>>>
>>>> My suggestion is that a fresh clone is used by everyone in all cases. This is because I tried rebasing
>>>> one of the other local repos I had here and I got conflicts. I suspect these will be a pain to resolve.
>>>>
>>>> There is only a small number of write accesses to the repo, so most of you will not be concerned
>>>> by this, but for those writing, a reminder not to bring in the tainted commits again.
>>>>
>>>> Finally, the process of killing the old develop branch has eliminated all pending PRs (I think
>>>> there were only 4, one being Felipe’s which was not for production). These will have to be
>>>> submitted again.
>>>>
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 3 |