The error message presented when a source file contains a tab character \t is a bit cryptic:
Error in examples/repro.ic:
Encountered unprintable character with integral value '9' encountered in source.
Ideally, tabs would be treated like regular whitespace. Barring that, we should have a better error message that (a) points to the location in the source, and (b) attempts to display the unprintable character in a more human-friendly way (ala man ascii).
Finally, the same error is output for every instance of a tab character in the source, which can be pretty overwhelming for a large file.
Parsing tab characters as normal whitespace seems more than reasonable. Feel free to spelunk in frontend/lex/lex.cc, probably starting looking at NextToken.
The error message presented when a source file contains a tab character
\t
is a bit cryptic:Ideally, tabs would be treated like regular whitespace. Barring that, we should have a better error message that (a) points to the location in the source, and (b) attempts to display the unprintable character in a more human-friendly way (ala
man ascii
).Finally, the same error is output for every instance of a tab character in the source, which can be pretty overwhelming for a large file.