Re: Error handling with Coco/r

Glen Herrmannsfeldt wrote:
(snip on compilers and error messages)


> [Back in the 1970s there was a lot of work on automatic error correction,
> which was of some interest when compilers ran as batch jobs and you wanted
> to get as much out of each job as possible. I agree that these days once
> you've pointed out where the error was, there's little point in doing any
> more work. -John]


I remember some of the ones that the PL/I (F) compiler used to
give me, sometimes good sometimes no so good.


There are many times when one small error can result in a very
large number of messages. In those cases, it might be better for
the compiler to 'correct' the error and then continue.


One that seems obvious is that a mistake in one part of a declaration
statement often causes the whole statement to be ignored, resulting in
messages related to all the other variables. In cases like that,
a little fixup goes a long way.


There are some cases where the wrong message can be very confusing.


Not so long ago in comp.lang.fortran there was a question about an
error message related to a statement function being before executable
statements. The real error was related to a mistake in an array,
declaration. Figuring out the most likely message might not only
depend on the number of changes, but also the likelihood of using
certain constructs.


-- glen

Return to the comp.compilers page.
Search the comp.compilers archives again.