Experimental Physics and Industrial Control System
Has anyone else run into a build dependency problem with base-3,14,12?
What I've been finding is that I can build an IOC or module that depends on
*.db or *.template files from another module successfully the first time, but
it fails on subsequent builds.
Example build output:
% make
perl /reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/bin/linux-x86_64/makeMakefile.pl O.linux-x86_64 ../../..
perl /reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/bin/linux-x86_64/makeMakefile.pl O.RTEMS-beatnik ../../..
perl /reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/bin/linux-x86_64/makeMakefile.pl O.linux-x86 ../../..
mkdir O.Common
/usr/bin/make -C O.linux-x86_64 -f ../Makefile TOP=../../.. T_A=linux-x86_64 install
make[1]: Entering directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.linux-x86_64'
perl /reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/bin/linux-x86_64/makeDbDepends.pl ../O.Common/iocAdmin.db ../iocAdmin.substitutions >> iocAdmin.db.d
echo "../O.Common/iocAdmin.db : " >> iocAdmin.db.d
Inflating database from ../iocAdmin.substitutions
/reg/g/pcds/package/epics/3.14/extensions/current/bin/linux-x86/msi -I. -I.. -I../O.Common -I ../../../db -I/reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/db -I/reg/g/pcds/package/epics/3.14/modules/autosave/R4.2.1.2-2.7.0/db -I/reg/g/pcds/package/epics/3.14/modules/iocAdmin/R3.0.0-1.6.0/db -I/reg/neh/home/bhill/wa2/trunk/pcds/epics/modules/event/current/db -S../iocAdmin.substitutions > iocAdmin.tmp
mv iocAdmin.tmp ../O.Common/iocAdmin.db
perl /reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/bin/linux-x86_64/makeDbDepends.pl ../O.Common/evr.db ../evr.substitutions >> evr.db.d
echo "../O.Common/evr.db : " >> evr.db.d
Inflating database from ../evr.substitutions
/reg/g/pcds/package/epics/3.14/extensions/current/bin/linux-x86/msi -I. -I.. -I../O.Common -I ../../../db -I/reg/g/pcds/package/epics/3.14/base/R3.14.12-0.4.0/db -I/reg/g/pcds/package/epics/3.14/modules/autosave/R4.2.1.2-2.7.0/db -I/reg/g/pcds/package/epics/3.14/modules/iocAdmin/R3.0.0-1.6.0/db -I/reg/neh/home/bhill/wa2/trunk/pcds/epics/modules/event/current/db -S../evr.substitutions > evr.tmp
mv evr.tmp ../O.Common/evr.db
Installing created db file ../../../db/iocAdmin.db
Installing created db file ../../../db/evr.db
make[1]: Leaving directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.linux-x86_64'
/usr/bin/make -C O.RTEMS-beatnik -f ../Makefile TOP=../../.. T_A=RTEMS-beatnik install
make[1]: Entering directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.RTEMS-beatnik'
make[1]: Nothing to be done for `install'.
make[1]: Leaving directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.RTEMS-beatnik'
/usr/bin/make -C O.linux-x86 -f ../Makefile TOP=../../.. T_A=linux-x86 install
make[1]: Entering directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.linux-x86'
make[1]: Nothing to be done for `install'.
make[1]: Leaving directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.linux-x86'
(232) 04:29 PM psusr115:~/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db
% make
/usr/bin/make -C O.linux-x86_64 -f ../Makefile TOP=../../.. T_A=linux-x86_64 install
make[1]: Entering directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.linux-x86_64'
make[1]: *** No rule to make target `ioc.db', needed by `../O.Common/iocAdmin.db'. Stop.
make[1]: Leaving directory `/reg/neh/home/bhill/wa2/epics/iocTest/test-R3.14.12/evr1/myEvrApp/Db/O.linux-x86_64'
make: *** [install.linux-x86_64] Error 2
The first build succeeds because RELEASE_DBFLAGS has a list of -I directives
which includes the released iocAdmin module's db dir, so msi is able to use that
to successfully build my iocAdmin.db file.
The problem arises because the first build also creates a dependency file
for iocAdmin.db using makeDbDepends.pl that looks like this.
% cat O.linux-x86_64/iocAdmin.db.d
../O.Common/iocAdmin.db: ioc.db access.db iocCluster.db
../O.Common/iocAdmin.db :
Subsequent builds fail because ioc.db is over in the released iocAdmin module's db dir,
which doesn't end up on the vpath for .db files.
I think I have a fix for it, but I can't figure out why this hasn't been a problem for us
and others before now.
I'd like to avoid customizing my base configure files to the extent possible,
but these changes seem to fix the problem.
% svn diff
Index: CONFIG_APP_INCLUDE
===================================================================
--- CONFIG_APP_INCLUDE (revision 8621)
+++ CONFIG_APP_INCLUDE (working copy)
@@ -16,6 +16,8 @@
RELEASE_INCLUDES += $$(addprefix -I,$$(wildcard $$(strip $$($(1)))/include))
RELEASE_DBDFLAGS += $$(addprefix -I,$$(wildcard $$(strip $$($(1)))/dbd))
RELEASE_DBFLAGS += $$(addprefix -I,$$(wildcard $$(strip $$($(1)))/db))
+ RELEASE_DIRS_DBD += $$(wildcard $$(strip $$($(1)))/dbd)
+ RELEASE_DIRS_DB += $$(wildcard $$(strip $$($(1)))/db)
RELEASE_PERL_MODULE_DIRS += $$(wildcard $$($(1)_LIB)/perl)
endef
$(foreach top, $(RELEASE_TOPS), $(eval $(call RELEASE_FLAGS_template,$(top)) ))
Index: RULES.Db
===================================================================
--- RULES.Db (revision 8757)
+++ RULES.Db (working copy)
@@ -11,10 +11,10 @@
##################################################### vpath
vpath %.dbd $(USR_VPATH) $(GENERIC_SRC_DIRS) $(dir $(DBD))
-vpath %.db $(USR_VPATH) $(GENERIC_SRC_DIRS) $(dir $(DB))
-vpath %.vdb $(USR_VPATH) $(GENERIC_SRC_DIRS) $(dir $(DB))
+vpath %.db $(USR_VPATH) $(GENERIC_SRC_DIRS) $(dir $(DB)) $(RELEASE_DIRS_DB)
+vpath %.vdb $(USR_VPATH) $(GENERIC_SRC_DIRS) $(dir $(DB)) $(RELEASE_DIRS_DBD)
vpath %.substitutions $(USR_VPATH) $(GENERIC_SRC_DIRS) $(COMMON_DIR)
-vpath %.template $(USR_VPATH) $(GENERIC_SRC_DIRS) $(COMMON_DIR)
+vpath %.template $(USR_VPATH) $(GENERIC_SRC_DIRS) $(COMMON_DIR) $(RELEASE_DIRS_DB)
vpath bpt%.data $(USR_VPATH) $(GENERIC_SRC_DIRS) $(COMMON_DIR)
vpath %.acf $(USR_VPATH) $(GENERIC_SRC_DIRS) $(COMMON_DIR)
vpath %.acs $(USR_VPATH) $(GENERIC_SRC_DIRS) $(COMMON_DIR)
Any comments would be appreciated.
Thanks,
- Bruce
- Replies:
- RE: Build problem with db dependencies Emma Shepherd
- Navigate by Date:
- Prev:
Re: BOY screens for motor controls Tim Mooney
- Next:
Obsolete code in configure/os/CONFIG.Common.linux-arm michael.abbott
- Index:
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
<2012>
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
- Navigate by Thread:
- Prev:
Re: BOY screens for motor controls Tim Mooney
- Next:
RE: Build problem with db dependencies Emma Shepherd
- Index:
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
<2012>
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024