makefile crashes

I have a very strange problem, but lets talk about my make file problem instead ;o) Hp-Ux 10.20 Patched to June 99 HP9000 E25 Informix Versions > cat $INFORMIXDIR/etc/*-cr INFORMIX-4GL Interactive Debugger Version 7.20.UD8 INFORMIX-4GL Version 7.20.UD8 INFORMIX-4GL Rapid Development System Version 7.20.UD8 Informix Dynamic Server Version 7.30.UC7 INFORMIX-SQL Version 7.20.UD8 Ok, I applied the Hp-Ux patch bundle XSW800GR1020 over the weekend. Now when I try to make my 4gl code, which has not changed, I get memory fault(coredump) errors. I get this with all programs, the error always occurs after the first few 4gl modules have been compiled, if I do not remove the .o files I can eventually get a successful compilation by repeatedly running the make as each time it picks up from where it left of. To try and determine where the errors were occurring I added echos to the make.rules file (see below) adding the "echo starting-up" prevents this error and allows the make to succeed!!! The "echo dunnit" lines were added first with no effect. I've searched the Hp knowledge base with no luck, I'm not sure what to tell the Hp tech support people, where does the error lie? I installed the patch bundle using swintall in its "text user interface" mode and used the "match what target has" function. All patches were installed successfully, a few were excluded because I'd already loaded a superceeding patch, I'll supply details if required. swlist -l fileset -a state shows all patches as configured. The error; hpe25:/progs/mars2/inv_gen> mu Debugging compilation c4gl -anyerr -a -c i_globs.4gl rm -f i_globs.c i_globs.ec c4gl -anyerr -a -c i_main.4gl rm -f i_main.c i_main.ec sh: 1963 Memory fault(coredump) *** Error exit code 139 A typical makefile #+ # invoice generation Makefile # # NOTE :- The leading space on the rule lines # (i.e. everything on line after a 'tag:') # MUST lead with a TAB (^I) character, else 'make' will fall over. # ==== #- system: inv_gen inv_gen: @make $(SYSTEM) -f $(PROGS)/dev_tools/make.rules \\ "BINDIR=/opt/invoicing/bin" "SQLDIR=/opt/invoicing/sql" \\ "MODULE=inv_gen.4ge" \\ "OBJS= i_globs.o i_main.o i_utils.o i_prechks.o i_reps.o \\ i_slacc.o LIBgetperm.o \\ LIB_about.o LIBround.o LIBstatus.o LIBgetcode.o LIBtrans.o LIBprompt.o LIBsem.o" \\ "FORMS=i_params.frm i_in_use.frm" The make.rules file ESQLC=$(INFORMIXDIR)/lib/fglc2 -4GL C4GL=$(INFORMIXDIR)/bin/c4gl FORM4GL=$(INFORMIXDIR)/bin/form4gl .C.o: cp $< $*.c $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.c rm -f $*.c echo dunnit1 .EC.o: cp $< $*.ec $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.ec rm -f $*.ec echo dunnit2 .4gl.o: echo starting-up c4gl $(MISYS_FLAGS) -c $< rm -f $*.c $*.ec echo dunnit3 .per.frm: $(FORM4GL) $< cp $@ $(SQLDIR) system: $(BINDIR)/$(MODULE) $(FORMS) clean: rm -f $(MODULE) $(OBJS) $(BINDIR)/$(MODULE) $(BINDIR)/$(MODULE): $(OBJS) $(LIBS) $(C4GL) -o $(MODULE) $(OBJS) $(LIBS) strip $(MODULE) mv $(MODULE) $(BINDIR)/$(MODULE) Any advice would be much appreciated. -- --------------------------------------- Tony Flaherty aef@mfs.misys.co.uk Analyst Programmer Misys Financial Systems All statements and opinions are my own, Misys don't pay me enough to have opinions on their behalf .

Try adding .4gl and .ec to the .SUFFIXES. Otherwise all looks well to me! Art S. Kagel PS- I may know someone who can help with that OTHER problem ;-} Tony Flaherty wrote: > > I have a very strange problem, but lets talk about my make file problem > instead ;o) > > Hp-Ux 10.20 Patched to June 99 > HP9000 E25 > > Informix Versions > > > cat $INFORMIXDIR/etc/*-cr > INFORMIX-4GL Interactive Debugger Version 7.20.UD8 > INFORMIX-4GL Version 7.20.UD8 > INFORMIX-4GL Rapid Development System Version 7.20.UD8 > Informix Dynamic Server Version 7.30.UC7 > INFORMIX-SQL Version 7.20.UD8 > > Ok, I applied the Hp-Ux patch bundle XSW800GR1020 over the weekend. Now > when I try to make my 4gl code, which has not changed, I get memory > fault(coredump) errors. I get this with all programs, the error always > occurs after the first few 4gl modules have been compiled, if I do not > remove the .o files I can eventually get a successful compilation by > repeatedly running the make as each time it picks up from where it left of. > > To try and determine where the errors were occurring I added echos to the > make.rules file (see below) adding the "echo starting-up" prevents this > error and allows the make to succeed!!! The "echo dunnit" lines were added > first with no effect. > > I've searched the Hp knowledge base with no luck, I'm not sure what to tell > the Hp tech support people, where does the error lie? > > I installed the patch bundle using swintall in its "text user interface" > mode and used the "match what target has" function. All patches were > installed successfully, a few were excluded because I'd already loaded a > superceeding patch, I'll supply details if required. swlist -l fileset -a > state shows all patches as configured. > > The error; > > hpe25:/progs/mars2/inv_gen> mu > Debugging compilation > c4gl -anyerr -a -c i_globs.4gl > rm -f i_globs.c i_globs.ec > c4gl -anyerr -a -c i_main.4gl > rm -f i_main.c i_main.ec > sh: 1963 Memory fault(coredump) > *** Error exit code 139 > > A typical makefile > > #+ > # invoice generation Makefile > # > # NOTE :- The leading space on the rule lines > # (i.e. everything on line after a 'tag:') > # MUST lead with a TAB (^I) character, else 'make' will fall over. > # ==== > #- > > system: inv_gen > > inv_gen: > @make $(SYSTEM) -f $(PROGS)/dev_tools/make.rules \\ > "BINDIR=/opt/invoicing/bin" "SQLDIR=/opt/invoicing/sql" \\ > "MODULE=inv_gen.4ge" \\ > "OBJS= i_globs.o i_main.o i_utils.o i_prechks.o i_reps.o \\ > i_slacc.o LIBgetperm.o \\ > LIB_about.o LIBround.o LIBstatus.o LIBgetcode.o LIBtrans.o > LIBprompt.o LIBsem.o" \\ > "FORMS=i_params.frm i_in_use.frm" > > The make.rules file > > ESQLC=$(INFORMIXDIR)/lib/fglc2 -4GL > C4GL=$(INFORMIXDIR)/bin/c4gl > FORM4GL=$(INFORMIXDIR)/bin/form4gl > .C.o: > cp $< $*.c > $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.c > rm -f $*.c > echo dunnit1 > > .EC.o: > cp $< $*.ec > $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.ec > rm -f $*.ec > echo dunnit2 > > .4gl.o: > echo starting-up > c4gl $(MISYS_FLAGS) -c $< > rm -f $*.c $*.ec > echo dunnit3 > > .per.frm: > $(FORM4GL) $< > cp $@ $(SQLDIR) > > system: $(BINDIR)/$(MODULE) $(FORMS) > > clean: > rm -f $(MODULE) $(OBJS) $(BINDIR)/$(MODULE) > > $(BINDIR)/$(MODULE): $(OBJS) $(LIBS) > $(C4GL) -o $(MODULE) $(OBJS) $(LIBS) > strip $(MODULE) > mv $(MODULE) $(BINDIR)/$(MODULE) > > Any advice would be much appreciated. > > -- > --------------------------------------- > Tony Flaherty aef@mfs.misys.co.uk > Analyst Programmer > Misys Financial Systems > All statements and opinions are my own, > Misys don't pay me enough to have opinions > on their behalf > > .

In article <937999801.501.0.nnrp-14.c1ed1f69@news.demon.co.uk>, Tony Flaherty <aef@mfs.misys.co.uk> writes >I have a very strange problem, but lets talk about my make file problem >instead ;o) > >Hp-Ux 10.20 Patched to June 99 >HP9000 E25 > >Informix Versions > >> cat $INFORMIXDIR/etc/*-cr >INFORMIX-4GL Interactive Debugger Version 7.20.UD8 >INFORMIX-4GL Version 7.20.UD8 >INFORMIX-4GL Rapid Development System Version 7.20.UD8 >Informix Dynamic Server Version 7.30.UC7 >INFORMIX-SQL Version 7.20.UD8 > > >Ok, I applied the Hp-Ux patch bundle XSW800GR1020 over the weekend. Now >when I try to make my 4gl code, which has not changed, I get memory >fault(coredump) errors. I get this with all programs, the error always >occurs after the first few 4gl modules have been compiled, if I do not >remove the .o files I can eventually get a successful compilation by >repeatedly running the make as each time it picks up from where it left of. > >To try and determine where the errors were occurring I added echos to the >make.rules file (see below) adding the "echo starting-up" prevents this >error and allows the make to succeed!!! The "echo dunnit" lines were added >first with no effect. > >I've searched the Hp knowledge base with no luck, I'm not sure what to tell >the Hp tech support people, where does the error lie? > >I installed the patch bundle using swintall in its "text user interface" >mode and used the "match what target has" function. All patches were >installed successfully, a few were excluded because I'd already loaded a >superceeding patch, I'll supply details if required. swlist -l fileset -a >state shows all patches as configured. > > > >The error; > >hpe25:/progs/mars2/inv_gen> mu >Debugging compilation > c4gl -anyerr -a -c i_globs.4gl > rm -f i_globs.c i_globs.ec > c4gl -anyerr -a -c i_main.4gl > rm -f i_main.c i_main.ec >sh: 1963 Memory fault(coredump) >*** Error exit code 139 > > > > >A typical makefile > >#+ ># invoice generation Makefile ># ># NOTE :- The leading space on the rule lines ># (i.e. everything on line after a 'tag:') ># MUST lead with a TAB (^I) character, else 'make' will fall over. ># ==== >#- > >system: inv_gen > >inv_gen: > @make $(SYSTEM) -f $(PROGS)/dev_tools/make.rules \\ > "BINDIR=/opt/invoicing/bin" "SQLDIR=/opt/invoicing/sql" \\ > "MODULE=inv_gen.4ge" \\ > "OBJS= i_globs.o i_main.o i_utils.o i_prechks.o i_reps.o \\ > i_slacc.o LIBgetperm.o \\ > LIB_about.o LIBround.o LIBstatus.o LIBgetcode.o LIBtrans.o >LIBprompt.o LIBsem.o" \\ > "FORMS=i_params.frm i_in_use.frm" > > > >The make.rules file > >ESQLC=$(INFORMIXDIR)/lib/fglc2 -4GL >C4GL=$(INFORMIXDIR)/bin/c4gl >FORM4GL=$(INFORMIXDIR)/bin/form4gl >.C.o: > cp $< $*.c > $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.c > rm -f $*.c > echo dunnit1 > >.EC.o: > cp $< $*.ec > $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.ec > rm -f $*.ec > echo dunnit2 > >.4gl.o: > echo starting-up > c4gl $(MISYS_FLAGS) -c $< > rm -f $*.c $*.ec > echo dunnit3 > >.per.frm: > $(FORM4GL) $< > cp $@ $(SQLDIR) > >system: $(BINDIR)/$(MODULE) $(FORMS) > >clean: > rm -f $(MODULE) $(OBJS) $(BINDIR)/$(MODULE) > >$(BINDIR)/$(MODULE): $(OBJS) $(LIBS) > $(C4GL) -o $(MODULE) $(OBJS) $(LIBS) > strip $(MODULE) > mv $(MODULE) $(BINDIR)/$(MODULE) > > >Any advice would be much appreciated. > What does file core give? >-- >--------------------------------------- >Tony Flaherty aef@mfs.misys.co.uk >Analyst Programmer >Misys Financial Systems >All statements and opinions are my own, >Misys don't pay me enough to have opinions >on their behalf > >. > -- David Williams

I re-ran this this morning and I'm still getting core dumps but now the message is Bus error(coredump) file core gives core file from 'make' - received SIGBUS or running on a different 4gl app. still gives the memory fault; core file from 'make' - received SIGSEGV To reiterate this 4gl code has not changed in any way! -- --------------------------------------- Tony Flaherty aef@mfs.misys.co.uk Analyst Programmer Misys Financial Systems All statements and opinions are my own, Misys don't pay me enough to have opinions on their behalf . David Williams wrote in message ... >In article <937999801.501.0.nnrp-14.c1ed1f69@news.demon.co.uk>, Tony >Flaherty <aef@mfs.misys.co.uk> writes >>I have a very strange problem, but lets talk about my make file problem >>instead ;o) >> >>Hp-Ux 10.20 Patched to June 99 >>HP9000 E25 >> >>Informix Versions >> >>> cat $INFORMIXDIR/etc/*-cr >>INFORMIX-4GL Interactive Debugger Version 7.20.UD8 >>INFORMIX-4GL Version 7.20.UD8 >>INFORMIX-4GL Rapid Development System Version 7.20.UD8 >>Informix Dynamic Server Version 7.30.UC7 >>INFORMIX-SQL Version 7.20.UD8 >> >> >>Ok, I applied the Hp-Ux patch bundle XSW800GR1020 over the weekend. Now >>when I try to make my 4gl code, which has not changed, I get memory >>fault(coredump) errors. I get this with all programs, the error always >>occurs after the first few 4gl modules have been compiled, if I do not >>remove the .o files I can eventually get a successful compilation by >>repeatedly running the make as each time it picks up from where it left of. >> >>To try and determine where the errors were occurring I added echos to the >>make.rules file (see below) adding the "echo starting-up" prevents this >>error and allows the make to succeed!!! The "echo dunnit" lines were added >>first with no effect. >> >>I've searched the Hp knowledge base with no luck, I'm not sure what to tell >>the Hp tech support people, where does the error lie? >> >>I installed the patch bundle using swintall in its "text user interface" >>mode and used the "match what target has" function. All patches were >>installed successfully, a few were excluded because I'd already loaded a >>superceeding patch, I'll supply details if required. swlist -l fileset -a >>state shows all patches as configured. >> >> >> >>The error; >> >>hpe25:/progs/mars2/inv_gen> mu >>Debugging compilation >> c4gl -anyerr -a -c i_globs.4gl >> rm -f i_globs.c i_globs.ec >> c4gl -anyerr -a -c i_main.4gl >> rm -f i_main.c i_main.ec >>sh: 1963 Memory fault(coredump) >>*** Error exit code 139 >> >> >> >> >>A typical makefile >> >>#+ >># invoice generation Makefile >># >># NOTE :- The leading space on the rule lines >># (i.e. everything on line after a 'tag:') >># MUST lead with a TAB (^I) character, else 'make' will fall over. >># ==== >>#- >> >>system: inv_gen >> >>inv_gen: >> @make $(SYSTEM) -f $(PROGS)/dev_tools/make.rules \\ >> "BINDIR=/opt/invoicing/bin" "SQLDIR=/opt/invoicing/sql" \\ >> "MODULE=inv_gen.4ge" \\ >> "OBJS= i_globs.o i_main.o i_utils.o i_prechks.o i_reps.o \\ >> i_slacc.o LIBgetperm.o \\ >> LIB_about.o LIBround.o LIBstatus.o LIBgetcode.o LIBtrans.o >>LIBprompt.o LIBsem.o" \\ >> "FORMS=i_params.frm i_in_use.frm" >> >> >> >>The make.rules file >> >>ESQLC=$(INFORMIXDIR)/lib/fglc2 -4GL >>C4GL=$(INFORMIXDIR)/bin/c4gl >>FORM4GL=$(INFORMIXDIR)/bin/form4gl >>.C.o: >> cp $< $*.c >> $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.c >> rm -f $*.c >> echo dunnit1 >> >>.EC.o: >> cp $< $*.ec >> $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.ec >> rm -f $*.ec >> echo dunnit2 >> >>.4gl.o: >> echo starting-up >> c4gl $(MISYS_FLAGS) -c $< >> rm -f $*.c $*.ec >> echo dunnit3 >> >>.per.frm: >> $(FORM4GL) $< >> cp $@ $(SQLDIR) >> >>system: $(BINDIR)/$(MODULE) $(FORMS) >> >>clean: >> rm -f $(MODULE) $(OBJS) $(BINDIR)/$(MODULE) >> >>$(BINDIR)/$(MODULE): $(OBJS) $(LIBS) >> $(C4GL) -o $(MODULE) $(OBJS) $(LIBS) >> strip $(MODULE) >> mv $(MODULE) $(BINDIR)/$(MODULE) >> >> >>Any advice would be much appreciated. >> > > What does > > file core > > give? > >>-- >>--------------------------------------- >>Tony Flaherty aef@mfs.misys.co.uk >>Analyst Programmer >>Misys Financial Systems >>All statements and opinions are my own, >>Misys don't pay me enough to have opinions >>on their behalf >> >>. >> > >-- >David Williams

Ok guys, panic over! I found a patch on the Hp web site PHCO_18638 which fixes this, it was a bug in make. Thanks for the replies. -- --------------------------------------- Tony Flaherty aef@mfs.misys.co.uk Analyst Programmer Misys Financial Systems All statements and opinions are my own, Misys don't pay me enough to have opinions on their behalf . Tony Flaherty wrote in message <937999801.501.0.nnrp-14.c1ed1f69@news.demon.co.uk>... >I have a very strange problem, but lets talk about my make file problem >instead ;o) > >Hp-Ux 10.20 Patched to June 99 >HP9000 E25 > >Informix Versions > >> cat $INFORMIXDIR/etc/*-cr >INFORMIX-4GL Interactive Debugger Version 7.20.UD8 >INFORMIX-4GL Version 7.20.UD8 >INFORMIX-4GL Rapid Development System Version 7.20.UD8 >Informix Dynamic Server Version 7.30.UC7 >INFORMIX-SQL Version 7.20.UD8 > > >Ok, I applied the Hp-Ux patch bundle XSW800GR1020 over the weekend. Now >when I try to make my 4gl code, which has not changed, I get memory >fault(coredump) errors. I get this with all programs, the error always >occurs after the first few 4gl modules have been compiled, if I do not >remove the .o files I can eventually get a successful compilation by >repeatedly running the make as each time it picks up from where it left of. > >To try and determine where the errors were occurring I added echos to the >make.rules file (see below) adding the "echo starting-up" prevents this >error and allows the make to succeed!!! The "echo dunnit" lines were added >first with no effect. > >I've searched the Hp knowledge base with no luck, I'm not sure what to tell >the Hp tech support people, where does the error lie? > >I installed the patch bundle using swintall in its "text user interface" >mode and used the "match what target has" function. All patches were >installed successfully, a few were excluded because I'd already loaded a >superceeding patch, I'll supply details if required. swlist -l fileset -a >state shows all patches as configured. > > > >The error; > >hpe25:/progs/mars2/inv_gen> mu >Debugging compilation > c4gl -anyerr -a -c i_globs.4gl > rm -f i_globs.c i_globs.ec > c4gl -anyerr -a -c i_main.4gl > rm -f i_main.c i_main.ec >sh: 1963 Memory fault(coredump) >*** Error exit code 139 > > > > >A typical makefile > >#+ ># invoice generation Makefile ># ># NOTE :- The leading space on the rule lines ># (i.e. everything on line after a 'tag:') ># MUST lead with a TAB (^I) character, else 'make' will fall over. ># ==== >#- > >system: inv_gen > >inv_gen: > @make $(SYSTEM) -f $(PROGS)/dev_tools/make.rules \\ > "BINDIR=/opt/invoicing/bin" "SQLDIR=/opt/invoicing/sql" \\ > "MODULE=inv_gen.4ge" \\ > "OBJS= i_globs.o i_main.o i_utils.o i_prechks.o i_reps.o \\ > i_slacc.o LIBgetperm.o \\ > LIB_about.o LIBround.o LIBstatus.o LIBgetcode.o LIBtrans.o >LIBprompt.o LIBsem.o" \\ > "FORMS=i_params.frm i_in_use.frm" > > > >The make.rules file > >ESQLC=$(INFORMIXDIR)/lib/fglc2 -4GL >C4GL=$(INFORMIXDIR)/bin/c4gl >FORM4GL=$(INFORMIXDIR)/bin/form4gl >.C.o: > cp $< $*.c > $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.c > rm -f $*.c > echo dunnit1 > >.EC.o: > cp $< $*.ec > $(C4GL) $(CFLAGS) $(MISYS_FLAGS) -c $*.ec > rm -f $*.ec > echo dunnit2 > >.4gl.o: > echo starting-up > c4gl $(MISYS_FLAGS) -c $< > rm -f $*.c $*.ec > echo dunnit3 > >.per.frm: > $(FORM4GL) $< > cp $@ $(SQLDIR) > >system: $(BINDIR)/$(MODULE) $(FORMS) > >clean: > rm -f $(MODULE) $(OBJS) $(BINDIR)/$(MODULE) > >$(BINDIR)/$(MODULE): $(OBJS) $(LIBS) > $(C4GL) -o $(MODULE) $(OBJS) $(LIBS) > strip $(MODULE) > mv $(MODULE) $(BINDIR)/$(MODULE) > > >Any advice would be much appreciated. > >-- >--------------------------------------- >Tony Flaherty aef@mfs.misys.co.uk >Analyst Programmer >Misys Financial Systems >All statements and opinions are my own, >Misys don't pay me enough to have opinions >on their behalf > >. >