`(nullable?.methodOrProperty ?? defaultValue) === defaultValue` should be a type guard about typescript

[MartinJohns]

Duplicate of #54825.

[typescript-bot]

This issue has been marked as "Duplicate" and has seen no recent activity. It has been automatically closed for house-keeping purposes.

[uhyo]

I doubt this is a duplicate. The linked issue is all about whether TS should know that 0 < 1, while this issue only requires the knowledge that 0 === 0, which TS already uses in many places.

Also the linked issue mentions another issue ( #48536 (comment) ) but that's a different topic too. This issue does not include any request to reason about “or” logics.

[MartinJohns]

The underlying issue is the same.

[tats-u]

@MartinJohns You have said that that issue is difficult to be fixed for a "design limitation", but this is a more limited form and easier to be fixed. This issue must not be closed just because of the combination of the "duplication" and that "design limitation".

[tats-u]

I have added an additional limitation that the fallback value is the same as the value of the other operand of the equality or non-or-equal comparison (< or >) operator.

[tats-u]

Why no advanced notice?

[MartinJohns]

Reopen and remove Duplilate

Ryan considers it a duplicate, that's why he marked the issue as such. He surely read your comment, but it doesn't change anything. The underlaying issue is still the same. Type narrowing is done on variables, not expressions. By adding the ?? 0 to bar?.length you're not checking anything bar related for truthiness anymore, but the entire expression, so bar is not narrowed. Changing that would require a lot of changes in the code base.

Why no advanced notice?

It's been marked as Duplicate for 4 days. That's the advance notice.

[tats-u]

@MartinJohns Do you mean #48536 (comment) ?
a ?? "" === "" is as simple as typeof a === "b". The only differences are:

• the fallback value propagated from the left operand expression when a is nullish is vary from a single value ("undefined")
• ?? is a binary instead of unary

FYI C# 8.0+'s null flow analysis can recognize this pattern. You will be able to get its specification from https://ecma-international.org/publications-and-standards/standards/ecma-334/ next year. This patern might as well be considered to be added to the backlog.

TS may not be able to handle the type number (not nullish) | 0 (nullish) unlike ("string" | "nunber" | ...) | "undefined" in a expression.

It's been marked as Duplicate for 4 days. That's the advance notice.

FYI, https://github.com/actions/stale, which has been famous, politely writes a post before closing, but your bot does not. However I do not require the bot to be improved because this improvement is just a bullshit job for your team and even hell can be home.

[MartinJohns]

Please note that I am not a member of the TypeScript team. It's not my bot.

Besides that, the bot did write a polite comment before closing the issue.

[tats-u]

Oh no, you are not Owner or Member. I have misunderstood and just noticed it. Sorry.