-
{{[[DONE]]}} (hidden) qyQpiD5t_
-
(hidden) GJSHDNTOr
-
(hidden) f_PWNuoPS [[(hidden) F3IIXCSCa]] [[(hidden) s-Ow4R2Dp]]
-
(hidden) D3LUWR3nG
-
(hidden) m4F9wBSgA
-
(hidden) DJj8msH7c
-
{{[[TODO]]}} (hidden) zVQkWf-kY [[(hidden) _hKhPUEak]]
-
(hidden) ep1WBLjxv
-
[[git-stacked-rebase]]
-
{{[[TODO]]}} ffs, the `Exitable` stuff is bad tbh, `exit` was a `TypeError` instead of an exit code / proper error - was hard to find, no good, gotta fix.
-
{{[[TODO]]}} we __really__ need e2e tests
-
also where the initial issue came from (re: [Exitable](((el1TOG8X8)))) - i just found out (by using the tool myself ofc) that i introduced a regression sometime between [that fix](((4R3-dZkR-))) (0.4.1) and now (0.5.0).
-
`git bisect` suspect - `0657970dd2303f264036606e0cdc06b6854a0104`
-
figured it out - it's because in the very ending of the `gitStackedRebase`, we `markThatNeedsToApply()`, and __then__ we call `applyIfNeedsToApply()`, which, of course, will now always think that we need to apply. other notes: it makes it impossible to e.g. `edit` in the rebase-todo - it continues (apply'ing?) w/o stopping and might explode if we start a new git-stacked-rebase for the first time (i.e. there's no `.git/stacked-rebase/` folder, specifically `.git/stacked-rebase/rewritten-list` file in this case) also, it'd be even harder to find because we'd only call the `applyIfNeedsToApply()` in the very end if `configValues.autoApplyIfNeeded` was true (design choice which made & still makes sense; it just makes the current bug harder to find)
-
```typescript if (configValues.autoApplyIfNeeded) { const { userAllowedToApply } = await applyIfNeedsToApply({...}) }```
-
initial fix:
-
https://github.com/kiprasmel/git-stacked-rebase/commit/6eadc9e601523fd39263de13007de56b4f7a991c
-
> currently `canAndShouldBeApplying` is only in that part of the logic. to be fair, the problem arises only when we try to apply before the rebase was finished. but, it very well could be that the user (even tho unlikely) that user runs `git-stacked-rebase` while a regular `git-rebase` is in progress, thus causing the same issue. TODO so we need to handle this everywhere, and the approach might be different as well.
-
-
-
-
-
-
[[codemods]]
-
-
{{[[TODO]]}} start using the `debug` lib
-
{{[[TODO]]}} add `--continue`, `--abort` as aliases to regular git rebase commands, for convenience
-
{{[[TODO]]}} file mappings fn that takes in the repo's workdir and returns the wanted file
-
o fuck, just realised when updating the CLI --help message - the `prompt` for `question` is a no no, at least in the library code (in cli code would be just fine).
-
{{[[TODO]]}} handle failures if --apply was (attempted to be) used more than once - currently errors because file missing; should print (return) info message instead
-
-
(hidden) SbB85_Qsi
-
(hidden) dwCGEsMqa
-
(hidden) 1_Vvzylkm
-
(hidden) jIPHiFqXP
-
(hidden) FWfm4c3eU