use_if_null_to_convert_nulls_to_bools has quickfix which does not respect operator precedence of the surrounding expression

[Chaos02]

If this is a bug report, please help us by providing:

Example:

  SingleChildRenderObjectWidget createArticleImage(Orientation orientation) => article!.image?.isNotEmpty == true && orientation == Orientation.portrait
        ? Padding(
            padding: const EdgeInsets.only(left: 5, bottom: 8, right: 5),
            child: integrationTestMode
                ? Image.asset('assets/icon/logo.png') // Use a static image during tests
                : Image.network('$baseImgUrl${article!.imageMobile!}'),
          )
        : emptyBox;

this correctly triggers the Dart Analyzer use_if_null_to_convert_nulls_to_bools lint.

After applying it the code looks like this:

  SingleChildRenderObjectWidget createArticleImage(Orientation orientation) => article!.image?.isNotEmpty ?? false && orientation == Orientation.portrait
        ? Padding(
            padding: const EdgeInsets.only(left: 5, bottom: 8, right: 5),
            child: integrationTestMode
                ? Image.asset('assets/icon/logo.png') // Use a static image during tests
                : Image.network('$baseImgUrl${article!.imageMobile!}'),
          )
        : emptyBox;

BUT it registers the && orientation == Orientation.portrait part as dead code (which is false, since nullable.isNotEmpty == true is the same as nullable.isNotEmpty ?? false

Dart SDK version: 3.9.2 (stable) (Wed Aug 27 03:49:40 2025 -0700) on "windows_x64"

Flutter 3.35.3 • channel stable • https://github.com/flutter/flutter.git
Framework • revision a402d9a437 (13 days ago) • 2025-09-03 14:54:31 -0700
Engine • hash 672c59cfa87c8070c20ba2cd1a6c2a1baf5cf08b (revision ddf47dd3ff) (12 days ago) • 2025-09-03 20:02:13.000Z
Tools • Dart 3.9.2 • DevTools 2.48.0

[mraleph]

That's because a ?? b && c is parsed as a ?? (b && c), not as (a ?? b) && c because ?? has lower precedence compared to &&. Seems like quick fix for the lint is incorrect.

[FMorschel]

I'll take a look at this

[FMorschel]

Thanks for reporting @Chaos02! This is a relatively new fix I introduced for this diagnostic at #47436.

Here is the CL for this fix https://dart-review.googlesource.com/c/sdk/+/449921.

[mraleph]

@FMorschel maybe it worth auditing all quickfixes which can perform expression -> expression rewrites and possibly introduce a helper for properly handling precedence?

[FMorschel]

I'll take a look at what other cases we have and think of a better way to handle them all together. Thanks for the idea. I'll track this at #61508

[Chaos02]

Hey, thanks for the really quick responses!
In the meantime I've noticed, that the lint indeed does not appear if you I close (a ?? b) in brackets...
Handling precedence sounds like a good plan? Especially since the "invert if statement" quick fixes already seem to do so (at least partly?).

Thanks again and have a good time!

[FMorschel]

In the meantime I've noticed, that the lint indeed does not appear if you I close (a ?? b)

Just to clarify, so you can avoid being misunderstood in the future:

• diagnostic: anything on the "problems" tab on VS Code or that dart analyze will show (squiggles in general)
• lint: this refers specifically to diagnostics created by a lint rule (e.g. use_if_null_to_convert_nulls_to_bools) - (the ones with blue squiggles)
• warning: any diagnostic made by analyzer that doesn't stop you from compiling code, but should be looked at (the ones with yellow squiggles) - (e.g. dead_code)
• error: any diagnostic that says you can't compile code (the ones with red squiggles)

Handling precedence sounds like a good plan? Especially since the "invert if statement" quick fixes already seem to do so (at least partly?).

Sure thing. I'll be sure to review those, too. Please open issues if you find any case that produces something wrong.

---

One last thing, this issue will automatically close once the CL above lands on main, so you don't have to worry about it.