Greetings,
I found that during a build of perl@10967 on the VAX the Digest::MD5
extension ran into the following compiler problem:
CC/DECC
/Include=[]/Standard=Relaxed_ANSI/Prefix=All/Obj=.obj/NoList/Define=("VERSION=""2.13""","XS_VERSION=""2.13""")/Include=([---])/Optimize
MD5.c
TALON::PVHP 16:12:55 MMS CPU=00:03:22.01 PF=381314 IO=159418 MEM=2068
TALON::PVHP 11:22:20 MMS CPU=00:03:22.01 PF=381316 IO=159419 MEM=2068
That is from 4:12 in the afternoon until 11:22 this morning (as indicated
by Ctrl-t) it could not compile the xsubpp extracted MD5.c file.
I then tried by hand without the /optimize qualifier:
set def [.ext.digest.md5]
CC/DECC
/Include=[]/Standard=Relaxed_ANSI/Prefix=All/Obj=.obj/NoList/Define=("VERSION=""2.13""","XS_VERSION=""2.13""")/Include=([---])
md5.c
^T
TALON::PVHP 11:52:30 DECC$COMP CPU=00:21:40.22 PF=423376 IO=159870 MEM=11776
^T
TALON::PVHP 12:03:59 DECC$COMP CPU=00:33:08.39 PF=423384 IO=159871 MEM=11784
^T
TALON::PVHP 12:07:11 DECC$COMP CPU=00:36:20.48 PF=423384 IO=159872 MEM=11784
^C
Cancel
%VCG-F-UABORT, Compilation terminated by user.
%VCG-I-NOBJECT, No object file produced.
I cannot convince myself that we ought to be trying to compile Digest::MD5
on the VAX with DEC C and CCVERSION .le. 50390006. Can anyone else refute
this assertion?
Peter Prymmer
