Autosave tse#50
Conversation
Merging from areaDetector:master
…ds_slac Merging in changes from Mark's ADCore master
Merging pcds_slac into master branch
… pcds_slac Updating from github
Bringing in updates from pcds_slac branch
Merging in recent work from Mark's githum ADCore master branch
… that derive from image updates
|
Hi Bruce, Thanks for the Pull Request. I have a few practical comments and suggestions that will make it easier to review and eventually merge: It seems that there are 10 commits in the Pull Request at the moment - but most of them are some merges, perhaps of the upstream "master". The real changes seems to me are all happening in the last commit: f8b0bd0 - is this correct? In order to maintain a clean/tidy history of the main ADCore repository it would be helpful if you could provide a simpler Pull Request with only the real changes. This can be done in a couple of ways:
Cheers, |
|
Hi Ulrik I'm not confident that cherry pick will work either, as I'd still have at least Let me know if you have other suggestions. My other option might Thanks,
On 01/08/2015 03:38 AM, Ulrik Kofoed Pedersen wrote:
Bruce Hill |
|
Hi Bruce, I have just had a go at it. I agree the rebase/squash is a bit difficult to get working - and potentially a bit dangerous to rewrite history anyway. I gave up on that after playing around with it a bit. Instead I had some success with cherry-pick - the trick is just to identify the exact commits where the real changes happen. I think I figured that out (including one which has spaces instead of tabs). What you will have to do is create a new clean branch off master and then cherry pick the exact useful commits one-by-one from your old branches into this new one. I have done that as an example with these commands: You can see the result here: https://github.com/ulrikpedersen/ADCore/tree/cherries and review the comparison from there. I think it's even possible for you to create a Pull Request straight off my fork if you wish! Cheers, |
|
Hi Ulrik I've tried rebase but haven't been able to find a way to get a clean master branch. I tried renaming the branch locally, thinking I could create a new one, but can't I ended up pushing the bitsPerPixel change from your cherries branch, which worked. I'd like to move to a workflow where my bhill-slac/ADCore master branch only reflects I'm thinking maybe my best option is just to delete the bhill-slac/ADCore fork on github Any recommendations? Thanks,
On 01/09/2015 02:40 AM, Ulrik Kofoed Pedersen wrote:
Bruce Hill |
|
Hi Bruce, Yes, I see you have got your master branch into a bit of a mess over several months with many merges and I think you are probably right, the best (certainly easier) way may be to simply delete you entire fork, and then create a new one. Your workflow on your new fork is also good: only use the master branch for synchronising with the upstream (i.e. original) ADCore. That is also how I work at the moment. It is a bit of a nuisance to start over, but consider it a hard earned git lesson - I've done it a couple of times before as well :-) Just make sure that you don't accidentally lose the changes in your Pull Request: i.e. hold on with the delete until we have accepted and merged (or rejected) your Pull Requests. Cheers, |
|
It looks to me that this breaks backwards compatibility. The default value of TSE is 0, so all existing IOCs will have TSE=0. But this pull request sets the default value of TSE to -2 for some records, which will change the timestamps on these records to be the time the image was collected, rather than the time the record processed. I think we need to preserve backwards compatibility. |
|
Hi @MarkRivers I think we can close this specific Pull Request - but your comment about TSE=0 vs TSE=-2 should apply to #53. Cheers, |
|
I'm closing this request as it's a mess and the changes are also in #53. |
Adds support for initialization and autosave of TSE field for records that update w/ each new image.