[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
core dump
From: |
PATTON, BILLY \(SBCSI\) |
Subject: |
core dump |
Date: |
Fri, 17 Mar 2006 06:27:47 -0600 |
Paul,
I got rid of all the gmsl. 3.79.1 complains about recursion in the
merge function.
With eval changed to warnings on the following lines :
$(foreach proj,$(PROJECTS),$(eval $(call proj_template,$(proj))))
$(foreach proj,$(PROJECTS),$(eval $(call build_proj_template,$(addprefix
build+,$(proj)))))
I get the following results
3.79 perfect results
3.79.1 perfect results
3.8.0 perfect results
3.8.1 perfect results
change the warning back to eval and I get
3.79 nothing
3.79.1 nothing
3.8.0 segemntation fault core dump
3.8.1 prefect results
Some more complex lines: too much stuff ? The result of this line is
where the core dump
points to. ldb is the proj it failes on. BB_ldb has 27 elements.
Explanation of the following line
$(1) = build+ldb where proj is one of 7 names ldb lde ldg ldm ldf ldw
lbf
BB_ldb := a b c d e f g ...
The result of the following line is
build+ldb+a build+ldb+b build+ldb+c ...
$(foreach bb,$(BB_$(patsubst build+%,%,$(1))),$(addsuffix
+$(bb),$(addprefix build+,$(patsubst build+%,%,$(1)))))
The define :
##
## build+ldb : build+ldb+bb
## @echo
define build_proj_template
$(1): $(foreach bb,$(BB_$(patsubst build+%,%,$(1))),$(addsuffix
+$(bb),$(addprefix build+,$(patsubst build+%,%,$(1)))))
@echo "build_proj_template $(1)"
endef
The rule produced by warning looks like :
Makefile:792: build+ldb: build+ldb+inftp build+ldb+opi build+ldb+schapi
build+ldb+tuxapi build+ldb+ldngn build+ldb+infgn build+ldb+ldgn
build+ldb+ldcli build+ldb+lddim build+ldb+tuxcom build+ldb+tuxjul
build+ldb+env build+ldb+ldac build+ldb+ldsrv build+ldb+opign
build+ldb+ldtup build+ldb+celltools build+ldb+lddim32
build+ldb+marshaler build+ldb+ldop build+ldb+inftl build+ldb+ldmrs
build+ldb+ldgdd build+ldb+sla build+ldb+lddl build+ldb+infsync
@echo "build_proj_template build+ldb"
This puts me in a very hard place. We can't put a beta 3.81 onto a
production platform.
The code I'm writing doesn't work on 3.79 and I get a seg fault on 3.80
I request a fix of 3.8.0 to fix the segmentation violation. I've sent
the core dumps.
I can send my makefile 811 lines, @ 675 are lists half the remainder
are comments.
So I probably have < 50 lines of interest.
>From the core dump
##5 0x1f738 in xrealloc (
# ptr=0x40018f78 "build+ldb : build+ldb+inftp build+ldb+opi
build+ldb+schapi build+ldb+tuxapi build+ldb+ldngn build+ldb+i\
n\
#fgn build+ldb+ldgn build+ldb+ld" <Address 0x40019000 out of bounds>,
size=4294965501) at misc.c:384
mics.c:384 it is a fixed length array?
-----Original Message-----
From: Paul Smith [mailto:address@hidden On Behalf Of Paul D. Smith
Sent: Thursday, March 16, 2006 4:04 PM
To: PATTON, BILLY (SBCSI)
Subject: RE: How can I see what is built with the eval
%% "PATTON, BILLY \(SBCSI\)" <address@hidden> writes:
pb> Excellent
pb> That will help a bunch.
pb> I'm also learning to use the gmsl stuff
pb> Is this bad or good?
Is what bad or good? GMSL? I think it's pretty cool stuff.
--
------------------------------------------------------------------------
-------
Paul D. Smith <address@hidden> Find some GNU make tips at:
http://www.gnu.org http://make.paulandlesley.org
"Please remain calm...I may be mad, but I am a professional." --Mad
Scientist
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- core dump,
PATTON, BILLY \(SBCSI\) <=