The issue message for the $<tmp coalesce var>101 is not understandable

[jorgsowa]

https://psalm.dev/r/ec8201d718

The description of the issues is not understandable. The term $<tmp coalesce var>101 doesn't explain what's wrong exactly. The phrase $<tmp coalesce var>101 should be probably annotated differently.

ERROR: [RedundantCondition](https://psalm.dev/122) - 7:10 - Type array<array-key, mixed> for $<tmp coalesce var>101 is never null

ERROR: [TypeDoesNotContainNull](https://psalm.dev/090) - 7:33 - Cannot resolve types for $<tmp coalesce var>101 - array<array-key, mixed> does not contain null

[psalm-github-bot]

I found these snippets:

https://psalm.dev/r/ec8201d718

<?php

function test_function(?int $number):array {
    return $number ? [$number] : [];
}

var_dump(test_function(null) ?? []);

Psalm output (using commit f90118c):

ERROR: RedundantCondition - 7:10 - Type array<array-key, mixed> for $<tmp coalesce var>101 is never null

ERROR: TypeDoesNotContainNull - 7:33 - Cannot resolve types for $<tmp coalesce var>101 - array<array-key, mixed> does not contain null

ERROR: ForbiddenCode - 7:1 - Unsafe var_dump

[orklah]

It comes from here:

psalm/src/Psalm/Internal/Analyzer/Statements/Expression/BinaryOp/CoalesceAnalyzer.php

Line 46 in 1c19260

$left_var_id = '$<tmp coalesce var>' . (int) $left_expr->getAttribute('startFilePos');

When processing a coalesce operator, Psalm will create a fake expression that looks like this: isset($<tmp coalesce var>XXX) ? $<tmp coalesce var>XXX : [];

the $<tmp coalesce var>XXX must be absolutely unique in the scope because otherwise it may clash with an existing variable in the scope. That's why the XXX is actually the position of the first char of the expression in the file, that way there won't be two clashing variables.

I agree the current syntax is confusing, feel free to make a PR and propose something better (keep in mind that it should be a unique name and must not be something that could already be a variable in the scope)

[AkshatHotCode]

@orklah I've noticed an issue related to Psalm errors and code readability in our project. I'm planning to work on this by introducing a more descriptive variable name in a coalesce operation.

Would appreciate any thoughts or collaboration on this. Thanks!

[orklah]

CoalesceAnalyzer is in need of a lot of love!

We have multiple bugs related to the logic in here and the fact that we transform the coalesce into a fake ternary is causing some of those.

But if you want to focus on the readability issue, my previous message summarized what is needed, a unique string that can't conflict with any variable in the scope nor any other similar ternary (for example, imagine your have $a ?? []; in two consecutive lines, the string you generate must be different between the two

We could maybe go with Type array<array-key, mixed> for **$a at line 108 char 5486** is never null that way it is both unique and descriptive

[AkshatHotCode]

Got it, will fix this issue and raise a PR for the same, At the same time feel you to share more issues that are related to this. Will try and help on them too!