2009/8/25 Martin Zobel-Helas <[email protected]>: > reopen 543263 > retitle 543263 "Fails with Assembler messages" > version 543263 0.8.2-3 > thanks > > Hi, > > now it fails with > > | libtool: compile: gcc -DHAVE_CONFIG_H -I. -I../../libcr -I.. -D_GNU_SOURCE > -D_REENTRANT -I../include -I../../include -I../../libcr/arch/sparc/ -Wall > -Wno-unused-function -Wall -g -O2 -MT libcr_la-cr_async.lo -MD -MP -MF > | +.deps/libcr_la-cr_async.Tpo -c ../../libcr/cr_async.c -fPIC -DPIC -o > .libs/libcr_la-cr_async.o > | /tmp/cckixYGU.s: Assembler messages: > | /tmp/cckixYGU.s:619: Error: Architecture mismatch on "membar". > | /tmp/cckixYGU.s:619: (Requires v9|v9a|v9b; requested architecture is > sparclite.) > | /tmp/cckixYGU.s:620: Error: Architecture mismatch on "cas". > | /tmp/cckixYGU.s:620: (Requires v9|v9a|v9b; requested architecture is > sparclite.) > | /tmp/cckixYGU.s:621: Error: Architecture mismatch on "membar". >
I spotted this one again too. Looks like it won't work on sparclite without some fairly serious porting work (I've been talking to upstream about portability issues, including this). It really needs async-safe atomic compare and swap routines as things currently stand. Which would you see as the best option for the time being then? Marking package as 'not for sparc', or compile with -mcpu=v9 (is that the right option?) and fail gracefully at run time on older sparcs? There's no way I can get a patch done to support sparclite without some major work to the library itself (which upstream is interested in and helpful in discussing) I'm not 100% sure I agree with your justification of serious though - it's not a miss-build, and it's never successfully built on sparc. Alan -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
