Issue: flux resource status is incorrectly merging drain messages with specific field formatting options

[kkier]

Reference:

[status.formats.long]
description = "Long flux-resource status format string"
format = "{state:>12} {color_up}{up:>2}{color_off} {nnodes:>6} +:{reason:<30.30+} {nodelist}"

Flux resource status is incorrectly merging drain messages in the {reason} field, apparently due to an interaction between the +: (do not truncate) and the + (display a plus sign if field is truncated) format options.

Sample output:

[root@ahwahnee1:6.1.0-82]# flux resource drain -f ahwahnee[1002-1003] "123456789012345678901234567890123"
[root@ahwahnee1:6.1.0-82]# flux resource status -o long -i ahwahnee[1002-1003]
       STATE UP NNODES REASON                            NODELIST
     drained  ✔      2 123456789012345678901234567890123 ahwahnee[1002-1003]
[root@ahwahnee1:6.1.0-82]# flux resource drain -f ahwahnee1002 "123456789012345678901234567890123456789"
[root@ahwahnee1:6.1.0-82]# flux resource status -o long -i ahwahnee[1002-1003]
       STATE UP NNODES REASON                                  NODELIST
     drained  ✔      2 123456789012345678901234567890123456789 ahwahnee[1002-1003]

[grondo]

I think Py was correct and the expand field +: should not be allowed with show truncation (+ suffix). For now a fix would be to remove the + from 30.30+ in the custom format.

Meanwhile I'll get a fix to silently ignore the truncation character if it is provided with +: (and hopefully that's not too difficult)