On Wednesday, August 05, 2020 07:43:53 PM SJW wrote: > On Wednesday, 5 August 2020 at 23:15:49 UTC+10 rhkr...@gmail.com wrote: > > Interesting puzzle ;-) > > > > Do you create one file named feature.txt, or do you create a different > > file for > > each feature with a unique name? > > > > (I'm trying to understand "all these text files" from below.) > > > > One thing I'd consider is just having one file named feature.txt, or > > notes.txt > > and modify that as appropriate and commit it with each change. > > > > Then when you checkout a feature branch, you get the notes (DB ALTER > > statements) associated with that feature, and it stays associated with > > that > > feature branch. > > > > You could choose another name for the text file, perhaps: db_alter.txt > > I have a different file for each feature - problem is, when I merge the > feature and delete the branch, the notes get merged so if I have 5 feature > branchs, I end up with 5 rogue feature.txt files
I guess you mean you have a <feature>.txt file for each feature, each with a different name. In retrospect (retrospect is easy ;-), not the way I'd do it. I don't know how big your repository is and how hard / time consuming it would be to go back and rename each of those <feature>.txt files with the same name, and then, as discussed earlier, merge those files at the same time you merge the features. (E.g., you probably don't need this, but to clarify: you might have a file sort of like a history file (probably in reverse chronological sequence?)>: (e.g.) file features.txt <date> Feature <name of feature 1> <description of feature 1> <notes about feature 1> <DB ALTER statements for feature 1> ... <date> Feature <name of feature n> <description of feature n> <notes about feature n> <DB ALTER statements for feature n> Now that you're in the present condition, I don't have another suggestion to offer.
