Damien - Hello guys,We use two clusters for d...
Hello guys,
We use two clusters for dev (with orgs : dev, test and integ) and for prod (with preprod and prod orgs).
The team develops on dev org, commits to dev branch regularly, pull requests to test branch, test branch is deployed to test org, etc ....
We have a lot of issues with this setup right now and I'm looking for your help and advice on those issues :
* Today we can't have multiple dev branches because of guid management. This is a pain when multiple features are developped at the same time : if we want to deploy first a feature that was commited last to the dev branch , we have to deal with mulitple commits mixed up in both dev and test branches
* Github pull requests generate additional commits (merge commits) that make moving between orgs very complex as the production branch will never be equal to any other branch.
* When deploying to an org, objects removed from git are not deleted from the org if I'm right, what is the best way to deal with them ?
Ideally we would like to be able to create a branch from master when we start a new feature, connect it to a new org, then after test is over, merge it to master, delete the original feature branch and the associated org. If I understand correctly, the main issue with this workflow is GUID mapping. There is a new feature called ObjectIds, does this allow us to get rid of guid mapping ?
Regarding our issues with Github, Gitlab handles merge requests differently and would solve this issue. Do you have an updated ETA for gitlab support ?
Thanks in advance,
Damien
We use two clusters for dev (with orgs : dev, test and integ) and for prod (with preprod and prod orgs).
The team develops on dev org, commits to dev branch regularly, pull requests to test branch, test branch is deployed to test org, etc ....
We have a lot of issues with this setup right now and I'm looking for your help and advice on those issues :
* Today we can't have multiple dev branches because of guid management. This is a pain when multiple features are developped at the same time : if we want to deploy first a feature that was commited last to the dev branch , we have to deal with mulitple commits mixed up in both dev and test branches
* Github pull requests generate additional commits (merge commits) that make moving between orgs very complex as the production branch will never be equal to any other branch.
* When deploying to an org, objects removed from git are not deleted from the org if I'm right, what is the best way to deal with them ?
Ideally we would like to be able to create a branch from master when we start a new feature, connect it to a new org, then after test is over, merge it to master, delete the original feature branch and the associated org. If I understand correctly, the main issue with this workflow is GUID mapping. There is a new feature called ObjectIds, does this allow us to get rid of guid mapping ?
Regarding our issues with Github, Gitlab handles merge requests differently and would solve this issue. Do you have an updated ETA for gitlab support ?
Thanks in advance,
Damien
