Re: Release 3.14.8: What goes in it and when?

[Ralph Lange <Ralph.Lange@bessy.de>]

Hi Andrew,

I'm taking a few days off now - but I might be back by the end of the 
first week of November or at least able to appear at work for a day or 
two to do the HP tests.

Note that while I'm doing the tests, most of the magic is being done by 
Jeff. (He knows spells that no earthling ever heard.)

Looking at my calendar: I'm at BESSY Nov 3rd for an important meeting 
anyway, so I could try to do tests on that Thursday afternoon. (With 
possible continuation on Friday 4th or during the 7th-11th week.)

Cheers,
Ralph


Andrew Johnson wrote:

> We need to get a new release of 3.14 out soon; I'm getting pressured 
> from more than one site, and there are enough R3.14.7 bugs already 
> fixed to make this worthwhile.  At this point I think a release date 
> in mid-November is probably more important than what extra bug fixes 
> or features we put into it.
>
> I am aware that in the short term Ralph will be far too busy with his 
> paternal responsibilities to do his usual HP-UX fixes, so my 
> suggestion is that 3.14.8 be announced as not supporting HP-UX, but we 
> do a 3.14.9 as soon as Ralph is able to perform the necessary HP-UX 
> magic.
>
> Here's a list of features that I'm aware are pending or that could be 
> added before this release:
>
> * Include Shifu Xu's real-time priority-based scheduling for Linux 
> (Posix?) and make enabling this a site-configuration option.
> * Simulation Mode enhancements from Peter Denison (Diamond)
> * Device Support eXtension Table: hardware address changes at runtime
>
> In addition these are the entries in Mantis that we should look at:
>
> 54: makeBaseApp.pl -i should ask for app name, not just use IOC name
>     Janet committed a fix for this, I'm revising it slightly.
>
> 111: non-preemptive clients disconnect if ca_poll() isnt called regularly
>     APS has been running a Gateway with Ken Evans' fix for this
>     problem for 6 months.
>
> 179: R3.14.7 does not build with vxWorks 6.0
>     This looks like it's been fixed, just not marked as such.
>
> 184: On a powerPC IOC during iocInit, a crash may occur during calls 
> to errlog functions
>     This looks like it's been fixed, just not marked as such.
>     Entries 182 and 188 may be duplicates of this.
> 182: strange behavior of outRoutine() in osdStdio.c associated with 
> crash for PPC targets.
> 188: errlogPrintf for vxWorks cause crash
>
> 200: errlogPrintf shouldnt be called from the core routines used to 
> implement it
>
> 216: Build fails with gcc 4 due to missing forward class declarartions
>     Patch provided in report, should be a quick fix.
>
>
> Jeff, can you comment on the status of 111, 179 and 216.
> Marty, can you look at 182, 184 and 188 and comment.
>
> Have I missed anything?
>
> I would appreciate feedback on whether to incorporate my DSXT work 
> into this release or not (it's currently available on the main trunk).
>
>
> Suggested Dates:
> Features and Mantis fixes committed on or before November 4th
> Platform testing during the week of November 7th-11th
> Final release on November 14/15th.
>
> If you have any issues with these dates please speak up.
>
> - Andrew