Add line numbers to parser error messages #574
Assignees
Labels
assembly
Related to Miden assembly
enhancement
New feature or request
good first issue
Good for newcomers
Comments
bobbinth
added
enhancement
bobbinth
changed the title
|
I would like to work on this. |
vlopes11
added a commit
that referenced
this issue
Apr 14, 2023
Prior to this commit, [ParsingError] tracked the step of the source instead of the line number. This is not friendly to the user as he won't be able to find the location of the error. A common standard is to display the line number, so the user can easily spot the location of the error. This commit introduces such track. It keeps a 1:1 map on [TokenStream] from the token to the parsed line. As in common text editors, line number starts at `1`. It also removes the `pos` from [Token] as its only purpose was to display the step to the user. The introduction of `line` completely supersedes that. closes #574
vlopes11
added a commit
that referenced
this issue
Apr 16, 2023
Prior to this commit, [ParsingError] tracked the step of the source instead of the line number. This is not friendly to the user as he won't be able to find the location of the error. A common standard is to display the line number, so the user can easily spot the location of the error. This commit introduces such track. It keeps a 1:1 map on [TokenStream] from the token to the parsed line. As in common text editors, line number starts at `1`. It also removes the `pos` from [Token] as its only purpose was to display the step to the user. The introduction of `line` completely supersedes that. closes #574
|
Closed by #854. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
To improve debugging experience, it may be helpful to know which lines in the source file contain malformed instructions. This can be achieved by adding line numbers to the
ParsingErrormessages.To do this, we need to add line number to the Token struct. This may be in addition to, or instead of, the current
posfield. This also means that we should collect line number info when we construct a token stream (i.e., here).It would probably make sense to track line numbers is a separate map inside a
TokenStream. This map would map a token position to a line number. This may result in a non-negligible performance overhead, and thus, we may want to introduce an option to instantiate token streams in debug mode. Then, line number info would be collected only in debug mode.Note: this would help only with errors generated by the parsers. For better run-time debugging experience a more comprehensive solution is needed.
The text was updated successfully, but these errors were encountered: