Code review comment for lp:~epics-core/epics-base/spinlockfix

I’m not a fan of assertions. Either a check is important or it isn’t. Not something to come and go between debugging and production versions.

On Jul 21, 2014, at 2:09 PM, Andrew Johnson <email address hidden> wr

> In the RTEMS and VxWorks implementations of epicsSpinLock() and epicsSpinUnlock(), is it wise to make an assertion with both interrupts and thread dispatch disabled? Won't that result in a silent lock-up of the IOC if the assertion fails? Not helpful in the presence of a nasty bug that sprays memory, or of a bad memory chip.
>
> The cantProceed() messages in the Posix implementation seem a little minimalistic, a little more detail might be helpful (put yourself in the position of an IOC developer who gets one of these, how useful are these to finding out what happened).
>
> - Andrew
> --
> https://code.launchpad.net/~epics-core/epics-base/spinlockfix/+merge/220845
> Your team EPICS Core Developers is requested to review the proposed merge of lp:~epics-core/epics-base/spinlockfix into lp:epics-base.

--
Eric Norum
<email address hidden>