Bug 98617
| Summary: | mysql41-server 4.1.20 compile problem on sparc64 | ||
|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Michael Scheidell <scheidell> |
| Component: | Individual Port(s) | Assignee: | Alex Dupre <ale> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | ||
| Priority: | Normal | ||
| Version: | Latest | ||
| Hardware: | Any | ||
| OS: | Any | ||
Responsible Changed From-To: freebsd-ports-bugs->ale Over to maintainer More information:
Make clean THEN make BUILD_STATIC works.
It appears its the BUILD_OPTIMIZED which makes -O3 option to cc.
This causes problems.
googles for other issues with sparc64 show same thing with a lot of
other ports.
Some prs seem to be closed, some still open.
I have yet to find one with a real answer.
I find that If I cc into that directory
/usr/ports/databases/mysql41-server/work/mysql-4.1.20/myisam
and compile THAT ONE OBJECT BY HAND, using -O2 and ont -O3 everything
else works fine.
cc -DMAP_TO_USE_RAID -I. -I. -I.. -I../include -I../include -DDBUG_OFF
-O -pipe -O2 -fno-omit-frame-pointer -fno-gcse -MT ft_nlq_search.o -MD
-MP -MF ".deps/ft_nlq_search.Tpo" -c -o ft_nlq_search.o ft_nlq_search.c
everything else compiles fine with -O3 (only that ONE piece of code)
Making all in myisam
if cc -DMAP_TO_USE_RAID -I. -I. -I.. -I../include -I../include
-DDBUG_OFF -O -pipe -O3 -fno-omit-frame-pointer -fno-gcse -MT sort.o
-MD -MP -MF ".deps/sort.Tpo" -c -o sort.o sort.c; then mv -f
".deps/sort.Tpo" ".deps/sort.Po"; else rm -f ".deps/sort.Tpo"; exit 1; fi
this patch COULD be applied to Makefile: (or we could figure out what is
wrong, and patch the source for sparc64 only)
Any sparc guru's, chime in:
--- Makefile.orig Wed Jun 7 14:47:27 2006
+++ Makefile Wed Jun 7 14:52:48 2006
@@ -90,7 +90,11 @@
CXX= ${CC}
.endif
.if defined(BUILD_OPTIMIZED)
+.if ${ARCH} != "i386"
CFLAGS+= -O2 -fno-omit-frame-pointer
+.else
+CFLAGS+= -O3 -fno-omit-frame-pointer
+.endif
.if ${OSVERSION} > 500000 || (defined(USE_GCC) && (${USE_GCC} == 3.0 ||
${USE_GCC} == 3.1 || ${USE_GCC} == 3.2 || ${USE_GCC} == 3.3))
CFLAGS+= -fno-gcse
.endif
--
Michael Scheidell, CTO
SECNAP Network Security / www.secnap.com
scheidell@secnap.net / 1+561-999-5000, x 1131
State Changed From-To: open->closed Simply do not use BUILD_OPTIMIZED. -O2 is already the default. They why does it let you do that? why would it hurt to add that patch? don't you think it would help users? is there some problem with that patch? If you think maybe make file should check for BUILD_OPTIMIZED and ARCH = sparc64 and fail, would that be acceptable? there ie already a test for USE_MYSQL that failes. |
critical security update for mysql41-server 4.1.20 won't compile on sun sparc box Making all in myisam if cc -DMAP_TO_USE_RAID -I. -I. -I.. -I../include -I../include -DDBUG_OFF -O -pipe -O3 -fno-omit-frame-pointer -fno-gcse -MT ft_nlq_search.o -MD -MP -MF ".deps/ft_nlq_search.Tpo" -c -o ft_nlq_search.o ft_nlq_search.c; then mv -f ".deps/ft_nlq_search.Tpo" ".deps/ft_nlq_search.Po"; else rm -f ".deps/ft_nlq_search.Tpo"; exit 1; fi {standard input}: Assembler messages: {standard input}:217: Error: Illegal operands: There are only 32 single precision f registers; [0-31] *** Error code 1 Stop in /usr/ports/databases/mysql41-server/work/mysql-4.1.20/myisam. How-To-Repeat: cd /usr/ports/databases/mysql41-server make BUILD_OPTIMIZED=yes BUILD_STATIC=yes or make BUILD_STATIC=yes or just: make