December 17th, 2021 | notes | kipras melnikovas

(hidden) ecwBkX2_S [[(hidden) iCbE7UjOk]]
- (hidden) vjsD7w2of #[[(hidden) S-ygYfi0j]] #bs
(hidden) ROUfwHntI
(hidden) mrumXxG23
(hidden) HsMrKZnC2
(hidden) 536wX09ce #[[(hidden) S-ygYfi0j]]
- {{[[DONE]]}} (hidden) menhG7vp0
  - (hidden) xnvOnn2a8
- {{[[DONE]]}} (hidden) 4ncVbp6r1 [[(hidden) tv1zOVzED]]
  - (hidden) dKxioSwPB
  - (hidden) 6hve8yJv2
- {{[[DONE]]}} (hidden) 6gGuZKnFD [[(hidden) tv1zOVzED]]
  - (hidden) n3-mLC3Dt
- (hidden) Lo-0Nn2kM [[(hidden) SZp9S9TdD]] #[[(hidden) okke3m7KB]]
  - {{[[TODO]]}} (hidden) YFcpi0pyU [[(hidden) jHk2xs5Yn]] [[(hidden) a1_YmqrZr]]
- {{[[DONE]]}} (hidden) SOM__46Ue [[(hidden) pHco5hQZn]]
  - (hidden) EFWjx7an7
  - (hidden) rMUWxAj-3
  - (hidden) jIY8rYgpn
- (hidden) sQOO09ODy #TIL
- (hidden) C2TsbTwUv
- [[codemods]] now
  - {{[[DONE]]}} ((kQlpZosXx))
    - nope
  - {{[[DONE]]}} ((j-RN6ORqM))
- (hidden) 7yC5Qa8V3
- [[codemods]] soon
  - {{[[DONE]]}} can we get rid of the bs @ `run.js`? ~~at least when end user is using it?~~ ugh but that's bad if they change stuff up themselves & it doesn't get updated. need a better solution.
    - oh yes
    - {{[[DONE]]}} move the tsconfig's `.tmp` ignore earlier (to 1st PR)
    - {{[[TODO]]}} mv the speed fix earlier
      - [[git-stacked-rebase]]
        
        {{[[TODO]]}} `--push -f` would help, lots of rebasing atm
        
        {{[[TODO]]}} extract `--apply` into separate fn
        
        {{[[TODO]]}} (sometime later) integration/e2e tests because atm idk if anything broke or not lol
        
        {{[[TODO]]}} `return <code>` instead of `process.exit(<code>)`
        
        #observation i now fully understand the [[(hidden) U5k5oJ6YK]] pattern of returning an exit code, and if success, writing into the function argument's reference. the [[(hidden) U5k5oJ6YK]] fellas are even more [[round-trips]] ahead than the #fp crew? 5Head
        
        okay now we have created our own monad or similar, that works like a Promise. sadly, now i'd need to have a `.map` or similar like `.then`, so that in case it does fail, we don't run the "expect to succeed" part.
        
        we now have a situation like this:
        
```typescript const goodCommands: Exitable<GoodCommand[]> = parseTodoOfStackedRebase({ pathToStackedRebaseTodoFile, }); if (goodCommands.code === ExitCode.FAIL) { return goodCommands; } logGoodCmds(goodCommands); ```

but, "goodCommands" is wrapped inside a monad `Exitable`. it's not easy to re-cast the type, it kinda sucks checking for it; it's not even clear what we're returning.

instead, imagine this:

```typescript ```
        
        fuck, this doesn't play well w/ [[typescript]].
        
        {{[[TODO]]}} is this also what [[(hidden) pokav_dad]] is about? 👀
        
        oh okay more [[observation]]s:
        
        1. sad that ts cannot infer that if type is `[A | null] | [null | B]`, if you verify one side is null, then other should not be, but doesn't infer automatically yet...
        
        2. could use a type assertion fn that could verify this. sadly more syntax.
        
        2.1 but, i also now see where the `Some(x)` comes from in [[(hidden) KMBIAt8MO]]. damn, these people have went thru many [[round-trips]] haven't they?:D
        
        .
        
        i guess this is the best we can do for now?
        
```typescript const [exit, goodCommands] = parseTodoOfStackedRebase(pathToStackedRebaseTodoFile); if (!goodCommands) return exit!; ```

...more discoveries... and they all suck...

oooh, maybe?

[[roam-traverse-graph]] [[export-html-pages]]

{{[[TODO]]}} #[[(hidden) jOMmlCP4S]] have a folder w/ literally all blocks as filenames, and inside them have js/html to redirect to the actual page that contains the block lmao

{{[[TODO]]}} cleanup `--help` - instead of so far to the right, have newline & some spacing (like 4 spaces or smthn)

{{[[TODO]]}} hmm. think if we still want to change the args (e.g. `--apply`) into commands (e.g. `apply`) -- instead we could make them compatible w/ each other, e.g. `--apply --push --force` / `-apf`

{{[[TODO]]}} fix whatever's missing from `forcePush` that gets fixed when `rebase` gets ran (w/o making any changes)
- educational vid for team (& beyond?)
  - {{[[TODO]]}} ((X6SUo3kI8))
- {{[[TODO]]}} complete documentation
  - {{[[DONE]]}} ask [[(hidden) pHco5hQZn]] to enable issues in the [[github]] repo
    - {{[[TODO]]}} get response
  - {{[[TODO]]}} mention that "the expected results depend on the current cui5map (mapping generated from figma, that will likely get updates)", & to follow the tests for the current implementation (tho again, mind that the cui5map might not be updated in the repo lol)
    - https://github.com/pipedrive/CodeshiftCommunity/blob/e8cc027e4b0b02d7b650f3fe0d3e8329a8321331/community/%40pipedrive__convention-ui-react/src/5.0.0/wip-transforms/rename-style-tokens/rename-style-tokens.spec.ts#[[(hidden) WBl6TSRcS]]17
      
      {{[[TODO]]}} update to proper branch once merged
  - {{[[TODO]]}} merge
- {{[[TODO]]}} merge PRs (up until & including 5th)
  - {{[[TODO]]}} tho care, might have fixups @ 6th. cc ((VjnnbrdvH))
  - {{[[TODO]]}} maybe shorten Symbol names since now they appear in e.g. "whom map to:"
  - {{[[DONE]]}} rename into `composeTransforms`
  - {{[[TODO]]}} ((eUJOaMbG_)) ?
- {{[[TODO]]}} merge cui5map
- {{[[TODO]]}} integrate [[(hidden) tv1zOVzED]]'s (post/s)css transform
- {{[[TODO]]}} finish up 6th
  - {{[[DONE]]}} tests
  - {{[[TODO]]}} extract fn `addOrExtendDestructuredImportOrRequire`
- {{[[TODO]]}} do a test run w/ someone to try out the codemods
- {{[[TODO]]}} afterwards, announce an ad-hoc meeting @ fe-devs & do a demo
- {{[[TODO]]}} handle styled() wraps at least, if not graph-aware yet
- oh my god the `renameJsxComponent` is killing me the garbage collector is a literal time terrorist wtf
  - {{[[TODO]]}} fixed it. how the f? do i need to learn this stuff in depth now?
- from chat w/ [[(hidden) pHco5hQZn]]
  - {{[[TODO]]}} merge approved PRs
  - {{[[TODO]]}} try out if codemods-cli works on separate published repo
  - {{[[TODO]]}} move code into separate repo
  - {{[[TODO]]}} create documentation on where to find resources, links, how to develop the codemods etc. - exactly what i planned w/ the vid
  - {{[[TODO]]}} upstream improvements w/o the actual codemods
- (hidden) rtK-3zrfO
(hidden) suUyUFUUJ
(hidden) m4NIdIVzx
(hidden) 6_iG3RFjO
(hidden) IeoUeX9Wm
(hidden) frCfoF65k
(hidden) bmlkXgiqN
(hidden) m65hytoiV
(hidden) ri6mMZQFk
(hidden) Lg5mEI7Tl
(hidden) Sb-I8P73K
(hidden) 5u7sypuX9
- (hidden) 8oWORenT8
- (hidden) ruRCHUGtn
- (hidden) EBgMV65K9
- (hidden) b8kysIfeC
(hidden) iCMUmFK3b
{{[[TODO]]}} (hidden) Dpo1kLOb2 #watch
- (hidden) 7JR5uy8M3
(hidden) 1wbWoxerf
[[roam-traverse-graph]]
- {{[[TODO]]}} fuck. currently marking as #public makes it a public-recursive one. yep, should proceed w/ the public & publicRecursive idea, & ditch publicGlobalTags
(hidden) sip9QxflS
[[git-stacked-rebase]]
- {{[[TODO]]}} handle case where origin/<branch> (initial branch) gets updated
  - currently get `AssertionError [ERR_ASSERTION]: very first commit has a branch.`:
    - ```javascript AssertionError [ERR_ASSERTION]: very first commit has a branch. at /Users/kiprasmelnikovas/projects/git-stacked-rebase/dist/git-stacked-rebase.js:355:34 at Array.map (<anonymous>) at createInitialEditTodoOfGitStackedRebase (/Users/kiprasmelnikovas/projects/git-stacked-rebase/dist/git-stacked-rebase.js:353:10) at async gitStackedRebase (/Users/kiprasmelnikovas/projects/git-stacked-rebase/dist/git-stacked-rebase.js:219:13) { generatedMessage: false, code: 'ERR_ASSERTION', actual: false, expected: true, operator: '==' }```
  - providing a commit also doesn't help ofc, since we need a branch
  - it seems here that we don't even detect the new initial branch and we pick up a lot of commits if not all.
(hidden) tDjurtIgy
(hidden) J4gdCm58d
(hidden) vvNqd_U4c
(hidden) prB5uvbXf
(hidden) dKK5eckQD

words:	927 + 0 = 927
last update: