search

epics-modules / iocStats

EPICS IOC Status and Control
Other
12 stars 40 forks source link

Fix timezone warning #38

Closed mark0n closed 1 year ago

mark0n commented 4 years ago

This fixes #37.

jeonghanlee commented 4 years ago

Does anyone have a plan to merge this pull request?

MarkRivers commented 4 years ago

Why is EPICS_TIMEZONE removed from iocAdminSoft.substitutions and why is EPICS_TZ not added there?

jeonghanlee commented 4 years ago

For me, TZ works with typical Soft IOCs with linux-x86_64. However, I saw @anjohnson mentioned that reason. Please see https://github.com/epics-modules/iocStats/issues/37

anjohnson commented 4 years ago

The EPICS_TIMEZONE and EPICS_TZ variables are really only needed on RTEMS or VxWorks IOCs, the workstation OS's use other methods to configure which timezone they should use when displaying a timestamp, so these changes shouldn't really affect anything running on Linux, MacOS or Windows. The iocAdminSoft.substitutions file had a PV showing the EPICS_TIMEZONE variable, but it would never have had any effect on the IOCs loading this file at all so IMHO @mark0n was right to remove it.

I haven't tested these changes myself, has anyone else? As long as someone other than Martin has tried them out I'll be happy to merge this PR. Unfortunately we don't really have anyone who owns this module at the moment so some of the core developers have been working on fixes as we find time, but I think we would prefer that someone else actually takes ownership of it.

jeonghanlee commented 4 years ago

@anjohnson What kind of test do we need to go further? Just running IOC on the Linux is fine? Stress test? Memory usage, and so on? Do we have any standard procedure which each EPICS module can be tested before merging into?

Let me know, I can help anything which I can do if we go further from this.

mark0n commented 4 years ago

@jeonghanlee there are no standard test procedures for EPICS modules and thus their code quality tends to vary. At a minimum someone needs to confirm it compiles without (new) warnings and that my changes actually do what they promise to do. This should already catch most of the stupid mistakes I make everyday and makes it less likely that it only works on my machine. Everything beyond that is up to you - whatever you deem necessary to allow this code to run on your production IOCs. Nobody will blame you if you miss something - it's each lab's responsibility to ensure code they grab from the Internet meets their quality requirements.

simon-ess commented 1 year ago

This has been closed in favour of #49, as @mark0n is no longer working with EPICS.