[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
%raw is broken?
From: |
Akim Demaille |
Subject: |
%raw is broken? |
Date: |
20 Sep 2000 14:10:04 +0200 |
User-agent: |
Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands) |
Hi everybody, and Jesse in particular,
I'm backing at working on Bison, and I'm currently setting up a test
suite. One of my tests fails. Since I'm tempted to call this a Bison
bug, I did check this test in CVS Bison, but if I'm wrong, I'll remove
it quickly.
It seems to me that %raw (aka --raw) does not work properly.
`calc.y', which is generated by the test suite, is appended below, but
the following trace should demonstrate the problem:
src/bison/tests % bison calc.y -o calc.c && gcc calc.c -o calc nostromo 14:03
src/bison/tests % echo "1 + 2" | ./calc && echo nostromo 14:03
3
src/bison/tests % bison calc.y -o calc.c --raw && gcc calc.c -o calc
src/bison/tests % echo "1 + 2" | ./calc && echo nostromo 14:03
parse error
I suppose the way the tokens are numbered with %raw is not completely
propagated in bison, or maybe %raw must not be used when you do return
characters (and not only tokens in its %token meaning) in which case
bison should probably refuse to see character literals in the grammar.
I don't know too well, but anyway, I'm pretty sure it should not
happen this way, or at least the documentation should display huge
warnings (I found none).
BTW, I never quite understood the interest of this option, could
someone explain it to me?
Thanks!
Akim
calc.y
Description: Text document
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- %raw is broken?,
Akim Demaille <=