Fix specs that exclude inputs their functions intentionally handle#15571
Merged
Conversation
Found by re-running the type checker over function bodies with spec-derived argument domains (elixir-lang#15559): * Config.Provider.validate_config_path!/1 was typed by the very type it validates, making its own check trivially true. A validator that raises a descriptive error on malformed input must accept term(). * List.to_float/1 and List.to_integer/1,2 declared charlist(), which admits the empty list, but the underlying BIFs always raise on it. Use nonempty_charlist(). * Logger.Formatter.prune/1 exists to sanitize INVALID chardata (its catch-all replaces unprintable data), yet its spec restricted the input to valid IO.chardata(). Accept term(). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
lukaszsamson
commented
Jul 9, 2026
|
|
||
| Typically called after formatting when the data cannot be printed. | ||
| """ | ||
| @spec prune(IO.chardata()) :: IO.chardata() |
Contributor
Author
There was a problem hiding this comment.
I'm not convinced about this change
Member
There was a problem hiding this comment.
This is wrong as well. It is completely fine for us to say that the input in the spec is a subset of what the function actually handles. The opposite is wrong.
lukaszsamson
commented
Jul 9, 2026
| Validates a `t:config_path/0`. | ||
| """ | ||
| @doc since: "1.9.0" | ||
| @spec validate_config_path!(config_path) :: :ok |
Contributor
Author
There was a problem hiding this comment.
I'm not convinced about this change
Member
There was a problem hiding this comment.
Yeah, it doesn't look correct to me either.
josevalim
reviewed
Jul 10, 2026
josevalim
reviewed
Jul 10, 2026
Co-authored-by: José Valim <jose.valim@gmail.com>
Member
|
💚 💙 💜 💛 ❤️ |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Found by re-running the type checker over stdlib function bodies with spec-derived argument domains (the spec-domain body check from #15559):
Config.Provider.validate_config_path!/1was typed by the very type it validates, making its ownis_binarycheck trivially true per the spec. A validator that raises a descriptive error on malformed input must acceptterm().List.to_float/1andList.to_integer/1,2declaredcharlist(), which admits[], but the underlying BIFs always raise on it. Nownonempty_charlist().Logger.Formatter.prune/1exists to sanitize invalid chardata (its catch-all clause substitutes the replacement char), yet its spec restricted input to validIO.chardata(), making that clause dead per the spec. Now acceptsterm().Assisted-By: Claude Fable 5
🤖 Generated with Claude Code