Experimental Physics and
Industrial Control System

Michael Davidsaver via Core-talk <core-talk at aps.anl.gov> · Mon, 20 Sep 2021 14:33:00 -0700

On 9/20/21 1:42 PM, Mark Rivers via Core-talk wrote:
> Hi Andrew,
> 
>  
> 
> Ø  How repeatable is the failure?
> 
>  
> 
> I seem to have repeatable problems, but they are not exactly what I saw before.
> 
>  
> 
> This is the output when I do “make clean uninstall” and then “make –sj8”, putting the output in file build3.txt.

Is '-s' necessary to reproduce?  If not, omitting this might give
some idea of exactly what the malformed command is.

Also maybe worthwhile asking a couple of ~silly background questions.

Which 'make' build/version is in use?

Which perl disto/version?

Is H: a local disk or network share?

Any notable environment variables effecting make?  (eg. '%SHELL%' or '%MAKESHELL%')

I'm wondering if ".." is a symptom of using 'cmd' instead of 'bash'
shell.

> H:\epics-devel\base-7.0.6>make -sj clean uninstall
> 
>  
> 
> H:\epics-devel\base-7.0.6>make -sj8 >  build3.txt
> 
> toolchain.c
> 
> '..' is not recognized as an internal or external command,
> 
> operable program or batch file.
> 
>  
> 
> I have attached the output file from that build, build3.txt.  I don’t see any errors.  However, it did not build the pvAccess, pvDatabase, etc. modules.  These are the only DLLs in bin\windows-x64.
> 
>  
> 
> H:\epics-devel\base-7.0.6>dir bin\windows-x64\*.dll
> 
> Volume in drive H is local_rw
> 
> Volume Serial Number is 6DDD-530A
> 
>  
> 
> Directory of H:\epics-devel\base-7.0.6\bin\windows-x64
> 
>  
> 
> 09/20/2021  03:15 PM           438,784 Com.dll
> 
> 09/20/2021  03:18 PM           317,440 ca.dll
> 
> 09/20/2021  03:21 PM           549,888 dbCore.dll
> 
> 09/20/2021  03:24 PM           307,200 dbRecStd.dll
> 
>                4 File(s)      1,613,312 bytes
> 
>                0 Dir(s)  700,124,401,664 bytes free
> 
>  
> 
> That does not seem correct, there should be many more DLLs.
> 
>  
> 
> I then just run “make –sj” again without cleaning first, putting the output in build4.txt
> 
>  
> 
> H:\epics-devel\base-7.0.6>make -sj8 >  build4.txt
> 
>  
> 
> There was no command line output.  The build did not create the missing DLLs.  I have attached build4.txt as well.
> 
>  
> 
> Thanks,
> 
> Mark
> 
>  
> 
>  
> 
>  
> 
> *From:*Johnson, Andrew N. <anj at anl.gov>
> *Sent:* Wednesday, September 15, 2021 5:45 PM
> *To:* Mark Rivers <rivers at cars.uchicago.edu>
> *Cc:* EPICS core-talk <core-talk at aps.anl.gov>
> *Subject:* Re: Building base 7.0.6 with parallel make fails on windows-x64
> 
>  
> 
> Hi Mark,
> 
>  
> 
> On Sep 11, 2021, at 9:02 AM, Mark Rivers via Core-talk <core-talk at aps.anl.gov <mailto:core-talk at aps.anl.gov>> wrote:
> 
>      
> 
>     I tried to build base 7.0.6 on windows-x64 using “make –sj8” on an 8-core machine.  It failed as follows:
> 
>      
> 
>     softIoc_registerRecordDeviceDriver.cpp
> 
>     softMain.cpp
> 
>        Creating library dbRecStd.lib and object dbRecStd.exp
> 
>     Generating code
> 
>     Finished generating code
> 
>     Generating code
> 
>     Finished generating code
> 
>  
> 
> The above files and dbCommon.dbd are built within modules/database/src.
> 
> 
> 
>     dbdExpand.pl: Can't find file 'dbCommon.dbd'
> 
>       while reading 'base.dbd' to create 'recTestIoc.dbd'
> 
>     dbdExpand.pl: Exiting due to errors
> 
>     make[5]: *** [../../../../../../configure/RULES.Db:403: ../O.Common/recTestIoc.dbd] Error 2
> 
>     make[5]: *** Waiting for unfinished jobs....
> 
>     make[4]: *** [../../../../../configure/RULES_ARCHS:58: install.windows-x64] Error 2
> 
>     make[3]: *** [../../../configure/RULES_DIRS:85: std/rec.install] Error 2
> 
>     make[3]: *** Waiting for unfinished jobs....
> 
>  
> 
> The recTestIoc.dbd file is created within modules/database/test/std/rec which is built as a subdirectory of modules/database/test. There is a hard dependency specified in modules/database/Makefile so it isn’t supposed to be possible to start building in the test directory until the src directory has finished building:
> 
>  
> 
>     DIRS += src
> 
>      
> 
>     DIRS += test
> 
>     test_DEPEND_DIRS = src
> 
> 
> 
> Thus I’m having difficulty imagining how that failure can occur (I’m not saying it couldn’t, I just don’t see how at the moment). How repeatable is the failure?
> 
>  
> 
>  
> 
> 
> 
>     Running make –sj8 again appeared to work although it started with this message:
> 
>      
> 
>     H:\epics-devel\base-7.0.6>make -sj8
> 
>     '..' is not recognized as an internal or external command,
> 
>     operable program or batch file.
> 
>     asTestLib.c
> 
>     dbHeaderTest.c
> 
>     arrayOpTest.c
> 
>     …
> 
>  
> 
> Those .c files are compiled in the modules/database/test/std/rec directory, which makes sense given the failure above. I’ll see if I can work out what the ‘..’ program might have been in case it provides more clues.
> 
>  
> 
> - Andrew
> 
>  
> 
> -- 
> 
> Complexity comes for free, simplicity you have to work for.
> 
>  
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System