Program logic bugs vs input/environmental errors

[ketmar via Digitalmars-d]

On Fri, 03 Oct 2014 19:25:53 -0700
Brad Roberts via Digitalmars-d <digitalmars-d at puremagic.com> wrote:

> Where's the contradiction?  The compilers state hasn't been corrupted 
> just because it encounters errors in the text file.
but compiler is in unknown state. it can't do telepathy, and it's tries
are annoying. there is no reason to guess what code programmer meant to
write, it's just a source of mystic/nonsensical error messages
("garbage", in other words).

the original source of "trying to recover and continue analysis" was
slow compilation times, it was really painful to restart compiler after
each error. but D compilation times are good enough to stop this
"guess-and-miss" nonsence. and we have good IDEs that can analyse code
on background and highligh errors, so there are virtually no reasons
for telepathy left.

yet many compilers (including D) still tries to do telepathy (and
fails). 'cmon, it's best to improve compiling times, not attempting to
guess. PL/1 fails at this, and all other compilers since then fails too.

it's strange to me that Walter telling us that program should stop once
it enters unknown state, but forcing D compiler to make uneducated
guesses when D compiler enters unknown state. something is very wrong
with one of this things.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://lists.puremagic.com/pipermail/digitalmars-d/attachments/20141004/943f6684/attachment.sig>