Re: callbackRequestDelayed question

[Michael Davidsaver via Core-talk <core-talk at aps.anl.gov>]

On 6/14/26 6:14 PM, Mark Rivers wrote:

• There is even a release notes entry!

Yes, that is what prompted me to re-do the test. 

However, the release notes just say "However this also resulted in timers expiring slightly [earlier than requested]". 

It is not clear from that statement that it fixed the problem of callbackRequestDelayed returning immediately for delays less than epicsThreadQuantum/2.  IMO that is more than "slightly earlier than requested", since it made delays less than 10 ms on Linux unavailable.

Also, "This affects several facilities which use epicsTimer, including record delays".  Still, I take your point.  Mentioning the local level "epicsTimer" name does not convey the full list of effected APIs and features.  And "slightly earlier" is a fuzzy statement.

I am anyway happy to hear that this long-standing issue has been resolved for you.

Mark

---

From: Michael Davidsaver <mdavidsaver at gmail.com>
Sent: Sunday, June 14, 2026 7:15 PM
To: Mark Rivers <rivers at cars.uchicago.edu>
Cc: core-talk at aps.anl.gov <core-talk at aps.anl.gov>; Johnson, Andrew N. <anj at anl.gov>
Subject: Re: callbackRequestDelayed question

 

On 6/14/26 4:51 PM, Mark Rivers via Core-talk wrote:

Hi Andrew,

I reported in November 2025 issues with callbackRequestDelayed returning essentially immediately for delays less that 0.01 second.

I am happy to report that this works fine in base 7.0.10 on Linux and Windows.

cf. github issue 106 and PR 704.  There is even a release notes entry!

https://github.com/epics-base/epics-base/blob/c5d1668895094bea44aee7655c386b2d9192cfa9/documentation/RELEASE-7.0.10.md?plain=1#L373-L389

...

...  I changed sscanTecord.c to remove that restriction except for vxWorks. ...

This tracks well with the change to epics-base.