FIX: fix autosave pass1 by setting PINI=YES

[ZLLentz]

closes #235

This is a workaround fix for autosave pass1 not behaving as expected for records where PINI=0

Implementation:

• Add a method update_pini_for_autosave on EPICSRecord with an appropriate docstring
• Call this method when creating output records

With the goal of being able to cleanly remove this method if we figure out why the expected behavior did not work at ads-ioc in pcdshub/ioc-common-ads-ioc#65

Resulting diff on my test IOC (extended to show the full record):

$ git diff
diff --git a/iocBoot/ioc-tc-mot-example/tc_mot_example.db b/iocBoot/ioc-tc-mot-example/tc_mot_example.db
index 319a356..cd3628b 100644
--- a/iocBoot/ioc-tc-mot-example/tc_mot_example.db
+++ b/iocBoot/ioc-tc-mot-example/tc_mot_example.db
@@ -13,6 +13,7 @@ record(ai, "PLC:TST:PASS1_RBV") {

 record(ao, "PLC:TST:PASS1") {
   field(DESC, "Main.fAutoPass1")
+  field(PINI, "1")
   field(DTYP, "asynFloat64")
   field(OUT, "@asyn($(PORT),0,1)ADSPORT=851/Main.fAutoPass1=")
   field(PREC, "3")
   info(autosaveFields, "VAL")
   info(autosaveFields_pass0, "PREC")
   info(archive, "VAL")
 }

[klauer]

This looks good. I'd like a bit more time to research this before rolling out the bandaid, but would not complain if it were deemed necessary to.

[ZLLentz]

I'd like to deploy the bandaid during the upcoming PAMM next week if we can't find the real culprit. This would help a lot with the tmo and fee vacuum systems.

[klauer]

Sounds like a plan. Will make a goal today to dive a bit deeper, ensuring there's enough time to tag/rebuild/release IOCs for the PAMM.

[klauer]

I'm convinced this is the correct fix, not a band-aid. Let's move forward with it.