On 7/25/06, John E. Malmberg <[EMAIL PROTECTED]> wrote:
I am still attempting to get a VMS build of the current blead.
One approach would be to configure with:
$ @configure -"Duser_c_flags=/warnings=informationals"
Though I haven't tried that, it should mimic the behavior of other
platforms that consider all warnings "informational" and no reason to
stop the build and pay attention to what you're doing. Not that I
necessarily recommend this since ignoring warnings is how we ended up
with code in the repository that compares a global variable with a
random value off the stack.
Of course people brave enough to delve into the regex engine have
other things to worry about and a swing or two have already been taken
at this particular problem:
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2006-07/msg00714.html
Perhaps we should apply the regexec.c bits from:
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2006-07/msg00639.html
but not the regcomp.c bits. I'm afraid it may be a day or two before
I can get to trying a build with that.
-John
[EMAIL PROTECTED]
Personal Opinion Only
1 61623 REGCP_SET(lastcp);
....1
%CC-W-UNINIT1, (1) The scalar variable "lastcp" is fetched but not
initialized.
And there may be other such fetches of this variable that have not
been reported in this compilation.
E do { if ((PL_debug & 0x00000200)) do {if
(re_debug_flags & 0x040000) if (lastcp != PL_savestack_ix) PerlIO_print
E f(Perl_PerlIO_stderr(), " Setting an EVAL scope,
savestack=%""ld""\n", (IV)PL_savestack_ix);} while (0); } while (0
E ); lastcp = PL_savestack_ix
2 61624 if (regmatch(reginfo, prog->program + 1)) {
E S_regmatch
2 61625 prog->endp[0] = PL_reginput - PL_bostr;
E PL_reg_state.re_state_reginput
E PL_reg_state.re_state_bostr
