Experimental Physics and
Industrial Control System

Finally, I think I'm beginning to get closer to the root problems I've been having using the Sequencer.

 

First, a setup:

0. This is all being implemented on my NetBSD port.  Everything so far has been fairly straightforward to convert and get running.

1. Using EPICS base 3.14.9 (3.14.10 is an upcoming project in itself).

2. I've got the source to EPICS installed in /usr/local/EPICSSRC  Directory is write protected, and I've got a rw overlay mounted over it, to allow writing the O.netbsd-x86 directories.

3. I'm installing everything in /usr/local/EPICS.  This allows the source to be remote from the application.

4. Base builds and runs with no errors.

5. Sequencer is installed at /usr/local/EPICSSRC/modules/soft/seq/seq.  INSTALL_LOCATION in <sequencer base>/configure/CONFIG_SITE points at /usr/local/EPICS

6. Sequencer builds and installs correctly, with no problem.

7. Async 4.9 is installed at /usr/local/EPICSSRC/modules/soft/asyn/asyn.

8. SNCSEQ in <asyn base>/configure/RELEASE points at /usr/local/EPICS

9. *here's the first problem.

When I build asyn, it fails saying that it can't figure out how to make target 'ipSNCServer.o', which is a .stt program.  Looking through the results of the configure/O.netbsd-x86/RULES_INCLUDE, there's nothing that will pull in the RULES_BUILD file from the sequencer, that has the definitions for the .stt build.  It'll pull in the RUILES_BUILD from the base, but it doesn't have any sequencer definitions in it.

 

If I modify the RULES_INCLUDE file (I know, you shouldn't ;-) generated in the asyn configure/O.netbsd-x86 directory, so that RULES_TOP points to where the source directory is at (/usr/local/EPICSSRC/modules/soft/seq/seq), and then modify the configure/RULES_BUILD file in that directory so that the SNC compiler comes from SNCSEQ instead of RULES_TOP, I can build things with no problem.

 

How about a proper fix, though?

 

David Dudley