Re: make cores on 5.4-RELEASE amd64 but not on i386, need a maintainer.

Harti Brandt wrote:
> On Mon, 30 May 2005, Julian Stacey wrote:
> 
> JS>Any make maintainers reading this ? ... I attach a gdb backtrace.
> 
> yes.
> 
> JS>
> JS>Make cores on 5.4-RELEASE amd64 but not on i386.  With a virgin
> JS>unhacked release-5.4 make executable, as built by a couple
> JS>of `make worlds'.  The hardware is OK, no other troubles.
> JS>
> JS>I could try strip & mail a minimal uuencoded tar.gz to 
> JS>provoke the core, but that's a little tricky:
> JS>	- Data tree is not /usr/src, but my personal data.  
> JS>	- Some of my macros are also needed & some of my tools, but 
> JS>	  all sources for those are public on the web.
> JS>	- A recipient would need a login on an amd64 machine.  
> JS>	  Mine's behind a firewall.
> JS>
> JS>#0  0x000000000043b2a2 in memcpy ()
> JS>#1  0x00000000004122e0 in SuffFindNormalDeps (gn=0x579d80, slst=0x561200)
> JS>    at suff.c:1797
> 
> This could be a hint to the bug in suffix parsing that Matt found some 
> time ago. Is there any chance you can test this with a make from -current?

Sure, I forgot to do that before posting, Thanks, done.
I just built make from 
	FreeBSD/branches/-current/src/usr.bin/make
with a CTM date stamp on last delta received of 
	214243 May 30 10:05 src-cur.6870.gz	(TZ=GMT+01:00)
& built on my 5.4-RELEASE based system (ie libs inc/ & kernel = 5.4).

Using current make it does Not Segmentation fault,
using 5.4 release make it does Segmentation fault.

So it seems its's a cleared bug :-)	Don't know if that's what
you were expecting.
What I've submited might at best now just be used as something for a
regression test, (though it calls my make macros & builds tools &
includes variables from other inc/Makefile type stuff, so Not a
nice easy self contained reegression test.

> I'll have a look tomorrow.

Thanks.
If you want me to test anything, shout.
My Makefile structure will change later, & it may then be
hard to tickle this bug again, but I'll keep a frozen version for
a while in case needed.

> harti
> 
> JS>#2  0x0000000000412c09 in SuffFindDeps (gn=0x579d80, slst=0x561200)
> JS>    at suff.c:2167
> JS>#3  0x0000000000412aac in Suff_FindDeps (gn=0x579d80) at suff.c:2115
> JS>#4  0x0000000000402a40 in CompatMake (gnp=0x579d80, pgnp=0x570980)
> JS>    at compat.c:453
> JS>#5  0x0000000000419224 in Lst_ForEachFrom (l=0x5754c0, ln=0x58a2a0, 
> JS>    proc=0x4029d0 <CompatMake>, d=0x570980)
> JS>    at /usr1/5.4-RELEASE/src/usr.bin/make/lst.lib/lstForEachFrom.c:94
> JS>#6  0x00000000004191b0 in Lst_ForEach (l=0x5754c0, proc=0x4029d0 <CompatMake>, 
> JS>    d=0x570980) at /usr1/5.4-RELEASE/src/usr.bin/make/lst.lib/lstForEach.c:73
> JS>#7  0x0000000000402a56 in CompatMake (gnp=0x570980, pgnp=0x570880)
> JS>    at compat.c:454
> JS>#8  0x0000000000419224 in Lst_ForEachFrom (l=0x575100, ln=0x57d680, 
> JS>    proc=0x4029d0 <CompatMake>, d=0x570880)
> JS>    at /usr1/5.4-RELEASE/src/usr.bin/make/lst.lib/lstForEachFrom.c:94
> JS>#9  0x00000000004191b0 in Lst_ForEach (l=0x575100, proc=0x4029d0 <CompatMake>, 
> JS>    d=0x570880) at /usr1/5.4-RELEASE/src/usr.bin/make/lst.lib/lstForEach.c:73
> JS>#10 0x0000000000402a56 in CompatMake (gnp=0x570880, pgnp=0x570880)
> JS>    at compat.c:454
> JS>#11 0x0000000000402f79 in Compat_Run (targs=0x587240) at compat.c:701
> JS>#12 0x000000000040bfbd in main (argc=1, argv=0x7fffffffdcf0) at main.c:897
> JS>
> JS>-
> JS>Julian Stacey        Net & Sys Eng Consultant, Munich       http://berklix.com
> JS> Mail in Ascii (Html=Spam).  Ihr Rauch = mein allergischer Kopfschmerz.
> JS>_______________________________________________
> JS>freebsd-current_at_freebsd.org mailing list
> JS>http://lists.freebsd.org/mailman/listinfo/freebsd-current
> JS>To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
> JS>
> JS>
> JS>
> 
> 

-
Julian Stacey        Net & Sys Eng Consultant, Munich       http://berklix.com
Mail in Ascii (Html=Spam).  Ihr Rauch = mein allergischer Kopfschmerz.