Michael,
Is the error every time you attempt to set a value or does it just pop up at IOC start?
It looks like having multiple CP links can process the record while at least one of the links isn't fully connected. Since the link isn't connected, the field for that value reports a nil value which causes the error when it attempts to concatenate. I need
to look into the behavior and coding of the calc-style records and see how they behave and what I'm doing wrong, but once all the CP links are connected, the record does correctly grab the string value rather than the index value. If that weren't the case
you wouldn't be getting an error, you'd just be getting a string with a number in the middle instead of the text you wanted.
Keenan
From: Sintschuk, Michael <michael.sintschuk at bam.de>
Sent: Thursday, February 4, 2021 4:14 AM
To: Lang, Keenan C. <klang at anl.gov>; Johnson, Andrew N. <anj at anl.gov>
Cc: tech-talk at aps.anl.gov <tech-talk at aps.anl.gov>
Subject: AW: combining long strings
Hello Andrew, Keenan,
I pulled now the recent luascript-master and rebuild the module/IOC. Now I get the following error:
Calling return table.pack(string.byte(AA..CC, 1, -1)) resulted in error: 1: attempt to concatenate a nil value (global 'AA')
I think, it tries to concatenate the index-value of the mbbo-record and not the string-value. If I write my template like this:
record(luascript,"$(exp):FilePath") {
field(INAA,"$(exp):PathMenu.ZRST CP")
field(INCC,"$(exp):SampleMenu.ZRST CP")
field(CODE,"return table.pack(string.byte(AA..CC, 1, -1))")
field(OUT,"$(cam):$(format)1:FilePath PP")}
It concatenates and puts the string correctly, without errors. But this way I always look at the same string…
Von: Lang, Keenan C. <klang at anl.gov>
Gesendet: Donnerstag, 4. Februar 2021 01:31
An: Johnson, Andrew N. <anj at anl.gov>
Cc: Sintschuk, Michael <michael.sintschuk at bam.de>; tech-talk at aps.anl.gov
Betreff: Re: combining long strings
I wasn't before, but I have now changed the errLog function to also write out using errlogPrintf.
On Feb 3, 2021, at 12:05 PM, Lang, Keenan C. <klang at anl.gov> wrote:
The ERR field has an error in it, so the record did try to process, it just never successfully completed. I haven't seen that error before, but it appears
that it points to some sort of issue with invalid memory access during the code execution. I'm trying to track down what went wrong.
Thanks for correcting me, I didn’t notice that PACT was still 1.
Does it also log lua errors using errlogPrintf() or an equivalent? I would want the IOC's error log to have a record of this kind of problem, and to not have to monitor the ERR fields of all the lua records.
Hi Michael,
I like the luascript script solution. But unfortunately it won’t work for me. The record is always in SEVR: invalid mode and there is no input to the AA, BB … fields:
PINI: NO PJ: 0 PJJ: 0x88d770 PPN: (nil)
PPNR: (nil) PREC: 0 PRIO: LOW PROC: 0
PSVL: PUTF: 0 PVAL: 0 RDES: 0x80c5f0
RELO: Every New File RPRO: 0 RPVT: 0x88d570
RSET: 0x7f8e4c1d9e40 SCAN: Passive SDIS:CONSTANT
SEVR: INVALID SPVT: (nil) STAT: UDF STATE: 0x88d7a8
SVAL: SYNC: Sync TIME: <undefined> TPRO: 0
TSE: 0 TSEL:CONSTANT UDF: 1 UDFS: INVALID
Did I missed something with the record settings?
Your record doesn’t appear to have been processed at all yet (STAT is UDF, UDF is 1 and TIME is <undefined>). Does processing it change the result? You can
have that happen automatically when the IOC boots by setting PINI to YES, or manually with this iocsh command:
dbpf Topo:FilePath.PROC 1
Complexity comes for free, simplicity you have to work for.
Complexity comes for free, simplicity you have to work for.
|