Code comlpetion not working for variables

[Ilyat1337]

Bug Report

🔎 Search Terms

Completion, IntelliSense

🕗 Version & Regression Information

Tested in 'Nightly' version - the problem is not fixed.

• This changed between versions 5.0.4 and 5.1.3

⏯ Playground Link

Playground link with relevant code

💻 Code

let foo: number = 1
fo

🙁 Actual behavior

Completion list ater fo does not contain foo item.

🙂 Expected behavior

Completion list ater fo contains foo item.

[Andarist]

This has been changed in typescript@5.1.0-dev.20230415 so the likely culprit is #52690 (new commits included in this version can be seen on this diff), cc @zardoy

[zardoy]

Thank you so much for the mention! I checked this and you're totally right. I was also annoyed with this recently, but didn't know that PR is culprit. It happens because you don't have semicolon and contextToken is numeric literal 1 (so here code thinks you're inside of variable declaration for some reason) though I'm looking at #52690 (comment) and it seems we just didn't have such test

[fatcerberus]

Huh, odd - you'd expect that ASI would kick in and have the parser act as if there's a semicolon after the 1

[svipas]

@DanielRosenwasser @zardoy @RyanCavanaugh issue still exists:

const obj = {
  name: '',
} as const // <- missing semicolon

---

If I add ; (semicolon) it starts to work:

const obj = {
  name: '',
} as const; // <- added semicolon

---

TS Version: 5.2.0-dev.20230717 and same with 5.1.6
Playground: https://www.typescriptlang.org/play?ts=5.1.6#code/MYewdgzgLgBCBGArGBeGBvAUDGYCGAtgKYBcMA5OQDSYC+MeEMoksA9GzAIIAmPMAbgZh+AS1gB3EACcA1hEyYgA

[svipas]

@RyanCavanaugh @DanielRosenwasser Let me know if I should open a new issue or we re-open this one

[zardoy]

@svipas I don't think it is related to this issue. This issue is about regression that started to happen from 5.1.x and was patched in 5.1.6. But your issue is happening in 5.0 and in older version as well.

The root of the issue is the same as in #54729 completions are being blocked in that position

[svipas]

Thanks @zardoy, you're right. I wrote a comment in your issue to include this case as well.