As always, my environment is DEC C V5.2-003 on OpenVMS Alpha V7.1.
Building an extension (DBI) against 5.5.670 fails with:
#include "objXSUB.h"
.^
%CC-F-NOINCLFILE, Cannot find file "objXSUB.h" specified in #include directive.
at line number 234 in file PERL_ROOT:[LIB.VMS_AXP.5_005670.CORE]XSUB.H;1
This should correct that although I won't have a chance to test it tonight:
--- vms/descrip_mms.template;-0 Wed Mar 1 00:28:41 2000
+++ vms/descrip_mms.template Wed Mar 1 21:55:53 2000
@@ -1091,6 +1091,9 @@
$(ARCHCORE)xsub.h : xsub.h
@ If F$Search("$(ARCHDIR)CORE.dir").eqs."" Then Create/Directory $(ARCHCORE)
Copy/Log $(MMS$SOURCE) $(MMS$TARGET)
+$(ARCHCORE)objxsub.h : objxsub.h
+ @ If F$Search("$(ARCHDIR)CORE.dir").eqs."" Then Create/Directory $(ARCHCORE)
+ Copy/Log $(MMS$SOURCE) $(MMS$TARGET)
$(ARCHCORE)$(DBG)libperl$(OLB) : $(DBG)libperl$(OLB) $(DBG)perlshr_xtras.ts
@ If F$Search("$(ARCHDIR)CORE.dir").eqs."" Then Create/Directory $(ARCHCORE)
Copy/Log $(MMS$SOURCE) $(MMS$TARGET)
[End of Patch.]
Any tests involving tainting blow up big time, as they did under 660.
I believe it went away when I built a debug version which is why I
forgot about it until now. If we can't build in debug, how would we
track this down -- hack the link so it builds with /TRACE and hope we
get a clue? Oh, and if I run the test script outside of the test
harness it doesn't blow up, although it does fail a number of tests.
$ @[.vms]test .exe "" "-v" [.op]taint.t
%DELETE-I-FILDEL, DISK8:[BERRYC.PERL-5_5_670.T]PERL.;1 deleted (8 blocks)
%COPY-S-COPIED, DISK8:[BERRYC.PERL-5_5_670]PERL.EXE;1 copied to
DISK8:[BERRYC.PERL-5_5_670.T]PERL.;1 (6 blocks)
[accounting stuff snipped]
[.OP]TAINT..............%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual
address=0000000056005058, PC=FFFFFFFF8050E990, PS=0000001B
Improperly handled condition, image exit forced.
Signal arguments: Number = 0000000000000005
Name = 000000000000000C
0000000000010000
0000000056005058
FFFFFFFF8050E990
000000000000001B
Register dump:
R0 = 00000000000438A5 R1 = 000000000000001B R2 = 0000000000054188
R3 = 0000000000000000 R4 = 00000000000438A3 R5 = 0000000000000000
R6 = 000000007AF67198 R7 = 0000000000000000 R8 = 0000000000000000
R9 = 00000000000438A5 R10 = 0000000000000000 R11 = 0000000000172170
R12 = 000000000000001F R13 = 0000000056005058 R14 = 00000073252F5F25
R15 = 0000000000000000 R16 = 0000000056005058 R17 = 000000000000004E
R18 = 000000007AF671E0 R19 = 0000000000000000 R20 = 0000000000000073
R21 = 000000000000005D R22 = 00000073252F5F25 R23 = 0000000000000018
R24 = 0000000000000001 R25 = 0000000000000019 R26 = 0000000000120CF4
R27 = 0000000000120290 R28 = 0000000000120CD0 R29 = 000000007AF66F60
SP = 000000007AF66F60 PC = FFFFFFFF8050E990 PS = 200000000000001B
FAILED on test 0
Failed 1 test, 0.00% okay.
u=17.35 s=0 cu=0 cs=0 files=0 tests=0
1-MAR-2000 21:56:17.67 User: BERRYC Process ID: 00007930
Node: BUBBA Process name: "BERRYC"
Accounting information:
Buffered I/O count: 14516 Peak working set size: 13328
Direct I/O count: 4410 Peak virtual size: 180608
Page faults: 16210 Mounted volumes: 0
Images activated: 137
Elapsed CPU time: 0 00:00:17.35
Connect time: 0 00:48:28.92
%DELETE-I-FILDEL, DISK8:[BERRYC.PERL-5_5_670.T]ECHO.EXE;1 deleted (8 blocks)
____________________________________________
Craig A. Berry
mailto:[EMAIL PROTECTED]
