Also tested with the latest byacc from https://invisible-island.net/byacc/:
====
./yacc -V
./yacc - 2.0 20210808
byacc-20210808/yacc gram.y
byacc-20210808/yacc: 21 shift/reduce conflicts, 3 reduce/reduce conflicts.
====
On 3/11/21 19:27, Domingo Alvarez Duarte wrote:
Hello !
I'm trying to build the CFRONT parser with bison instead of byacc and
I've got bison to accept the grammar
https://github.com/mingodad/cfront-3/blob/master/src/gram.y but the
generated parser by bison has a small difference in the conflicts
found in the grammar (see bellow) and also by comparison I tested the
last C++ grammar from GCC 3.3.6
https://github.com/gcc-mirror/gcc/blob/releases/gcc-3.3.6/gcc/cp/parse.y
and there byacc and bison seem to agree on the conflicts.
I tested with two bison versions and both give the same results but
the generated parser do not pass the tests.
I'm pointing this here because it seems to be a good example of a
corner case maybe not noticed so far.
bison (GNU Bison) 3.0.4
bison (GNU Bison) 3.8.2
byacc - 1.9 20140715
====
cfront-3/src$ bison -y gram.y
gram.y: warning: 20 shift/reduce conflicts [-Wconflicts-sr]
gram.y: warning: 4 reduce/reduce conflicts [-Wconflicts-rr]
cfront-3/srv$ byacc gram.y
byacc: 21 shift/reduce conflicts, 3 reduce/reduce conflicts.
====
====
gcc-3.3.6/gcc/cp$ bison -y parse.y
parse.y: warning: 33 shift/reduce conflicts [-Wconflicts-sr]
parse.y: warning: 58 reduce/reduce conflicts [-Wconflicts-rr]
gcc-3.3.6/gcc/cp$ byacc parse.y
byacc: 33 shift/reduce conflicts, 58 reduce/reduce conflicts.
====
Cheers !