Mailing List Archive

Thank you for your help.

Jeff> What is Select_fd_set_t #defined to in config.h? It should
Jeff> be int *.
Andreas>I can't answer that, it was Harri's machine... Harri?

Select_fd_set_t is defined to int * .

Andreas>Harri, can you test this patch, please?
Andreas>--- DIST/MsqlPerl-1.02/Msql.xs Sun Aug 20 12:53:46 1995
Andreas>+++ Msql.xs Sun Sep 24 11:27:28 1995
Andreas>@@ -345,7 +345,7 @@
Andreas> curField = msqlFetchField(result);
Andreas> if (cur[off]){
Andreas>- PUSHs(sv_2mortal((SV*)newSVpv(cur[off], strlen(cur[off]))));
Andreas>+ PUSHs(sv_2mortal((SV*)newSVpv((char*)cur[off], (STRLEN)strlen(cur[off]))
Andreas>));
Andreas> }else{
Andreas> PUSHs(&sv_undef);
Andreas>@@ -543,7 +543,8 @@
Andreas> while ( cur = msqlFetchRow(result) ){
Andreas> EXTEND(sp,1);
Andreas> curField = msqlFetchField(result);
Andreas>- PUSHs(sv_2mortal((SV*)newSVpv(cur[0], strlen(cur[0]))));
Andreas>+ PUSHs(sv_2mortal((SV*)newSVpv(cur[0], (STRLEN)strlen(cur[0]))));
Andreas>+ PUSHs(sv_2mortal((SV*)newSVpv((char*)cur[0], strlen(cur[0]))));
Andreas> }
Andreas> msqlFreeResult(result);
Andreas> }
Andreas>@@ -567,7 +568,7 @@
Andreas> while ( cur = msqlFetchRow(result) ){
Andreas> EXTEND(sp,1);
Andreas> curField = msqlFetchField(result);
Andreas>- PUSHs(sv_2mortal((SV*)newSVpv(cur[0], strlen(cur[0]))));
Andreas>+ PUSHs(sv_2mortal((SV*)newSVpv((char*)cur[0], (STRLEN)strlen(cur[0]))));
Andreas> }
Andreas> msqlFreeResult(result);
Andreas> }

I applied the patch you sent me. It didn't seem to help. I still get
warnings.

cc -c -Ae +O3 +Onolimit +Z -I/usr/local/lib/perl5/hpux/CORE -I/custom/b
in/Minerva/include Msql.c
cc: "Msql.c", line 500: warning 604: Pointers are not assignment-compatible.
cc: "Msql.c", line 500: warning 563: Argument #1 is not the correct type.
cc: "Msql.c", line 933: warning 604: Pointers are not assignment-compatible.
cc: "Msql.c", line 933: warning 563: Argument #1 is not the correct type.
cc: "Msql.c", line 933: warning 604: Pointers are not assignment-compatible.
cc: "Msql.c", line 933: warning 563: Argument #1 is not the correct type.
cc: "Msql.c", line 935: warning 604: Pointers are not assignment-compatible.
cc: "Msql.c", line 935: warning 563: Argument #1 is not the correct type.
cc: "Msql.c", line 990: warning 604: Pointers are not assignment-compatible.
cc: "Msql.c", line 990: warning 563: Argument #1 is not the correct type.

Are those warnings of any importance?


Also I did a quick and dirty fix into Makefile and changed ldflags from
-b -s -a shared to -b -s -a archive and it seemed to help. At least
ld now finds libmsql.a from path. Adding another -L entry had no effect.

Now I get following ld output from make:
LD_RUN_PATH="/custom/bin/Minerva/lib:/node/kaprun/disc01/users/ideas_tea
m/custom/bin/Minerva/lib" ld -o ./blib/auto/Msql/Msql.sl -b -s -a archive Msql.o
-L/custom/bin/Minerva/lib -L/node/kaprun/disc01/users/ideas_team/custom/bin/M
inerva/lib -lmsql -lmisc

So now MsqlPerl seems to work allright, but the "make test" still hangs.

Jeff> Hangs in what way?
Jeff> Jeff

Test script starts tp execute and after a few seconds nothing happens anymore.
After that msqld takes all of processor's time (idle 0%) and nothing happens
anymore. Test program passes checkpoint 24 (ok 24), but hangs just before
checkpoint 25 when opening the test database once again. Mystifying :-O

Best regards,
Harri

---------------------------------------------------------------------------
Harri Piili harri.piili@ntc.nokia.com
CAD Engineer Phone: +368 0 511 27923
Nokia Telecommunications Oy Fax: +368 0 511 27582
P.O.Box 12
FIN-02611 Espoo, Finland
---------------------------------------------------------------------------