Paul Eggert wrote:
Jeannot Langlois <[EMAIL PROTECTED]> writes:
Are there any mechanisms already in place to communicate the presence
of this "-ly" option to the skeleton files? (Like any M4 macros
already defined, or any symbol?).
The -y option (not -ly -- that's for the linker) sets the yacc_flag
variable in the C code. If that's not communicated to the templates
via an M4 macro you can add something to do it.
Hi Paul, I have installed Bison-2.0a and patched my M4 macros in
successfully. I have checked the testsuite and indeed, tests cases (now
offset by two, as I guess two testcases have been added between 2.0 and
2.0a) #119 and #128 were NOT using bison option "-y" (and I couldn't
find any of these testcases linking with "-ly"). So I have changed
these testcase's yyerror() return value to "void".
Now I have added necessary code to propagate the presence of the
"yacc_flag" option to the yacc.c and glr.c skeleton files. When I run
the testsuite, now the following test cases do fail: 67, 71, 73, 74, 75,
81, 85, 87, 88, and 89 -- still for a similar reason: they use "%yacc"
(which is the equivalent of bison option "-y" AFAIK), and therefore
cause my M4 macros to generate a yyerror() declaration that use an "int"
return value -- but of course, they still declare a "static void
yyerror(...)" prototype in the prologue section.
Are these declarations another possible mistake by the test case
writer? Shouldn't these test case use "static int yyerror(...)" since
they assert "%yacc" ? (Should I change these test cases to use "void"
too? Or maybe there's something I didn't get...?)
Thanks,
--
*Jeannot Langlois*
B. Sc. Computer Science / B. Sc. Informatique
Software Developer / Programmeur-Analyste
System/Network Administrator / Administrateur Syst�me/R�seau
**
LINUX_LOGO
