Op een zonnige winterdag (Saturday 01 March 2003 19:57), schreef Abe Timmerman: > I'm afraid that here's another one to get "PERL_COPY_ON_WRITE" compiling > [needs regen_headers again]. (Another smoke started.) It needs more, again F<makedef.pl> :-( So is the whole thing (including the previous one): diff -ru perl-current/embed.fnc perl-new/embed.fnc --- perl-current/embed.fnc 2003-03-01 10:27:57.000000000 +0100 +++ perl-new/embed.fnc 2003-03-01 18:46:48.000000000 +0100 @@ -945,7 +945,7 @@ Ap |char * |custom_op_desc|OP* op #if defined(PERL_COPY_ON_WRITE) -pM |int |sv_release_IVX |SV *sv +ApM |int |sv_release_IVX |SV *sv #endif Adp |void |sv_nosharing |SV * diff -ru perl-current/makedef.pl perl-new/makedef.pl --- perl-current/makedef.pl 2003-03-01 10:27:57.000000000 +0100 +++ perl-new/makedef.pl 2003-03-01 21:55:00.000000000 +0100 @@ -562,8 +562,9 @@ unless ($define{'PERL_COPY_ON_WRITE'}) { skip_symbols [qw( - Perl_sv_setsv_cow - )]; + Perl_sv_setsv_cow + Perl_sv_release_IVX + )]; } unless ($define{'PERL_FLEXIBLE_EXCEPTIONS'}) { Good luck, Abe -- And this was cleaner and easier to understand how...? :-) The eval sprintf is certainly 'cleverer' but 'clever' is not equivalent with 'good'... -- Jarkko Hietaniemi on p5p @ 2001-11-14Thread Previous | Thread Next