Re: The Haskell Error Index
Lassi Kortela 02 Jul 2024 09:31 UTC
> The problem is that Haskell is one interpreter with strict error messages.
> With Scheme you a lot of different interpreters and each one give different
> error messages. R7RS say nothing about error messages.
https://errors.haskell.org/ is not limited to GHC errors. They say that
errors from other tools can be included, and there are already two from
Haskell Stack (a build system).
If we made a similar site for Scheme, we could make one page per
message, which explains that message for every Scheme implementation
(and tool) that has it.
A good point raised by the Haskell folks is that novices don't
necessarily understand which tool an error message is coming from.
> Maybe it will be good idea to create SRFI with common error messages for
> every error, or most common errors. So it will be feasible to create such
> page.
The SRFI could be a good idea, but it would take many years to be widely
adopted. I think the page would work fine even without standard error
messages. In fact the page might be more useful for non-standard errors,
since standard errors would at least be explained in the SRFI.
Non-standard errors are often not explained anywhere.