[HECnet] RSTS/E started emitting "?EVTLOG (BLDNIC) -- %Integer error" messages

[Paul Koning]

I see what's going on.

You have the original Evtlog.  That time stamp is from the logger, not FAL (it generates the event but the OS puts on the timestamp and evtlog processes it).  The 2034 date is what evtlog as shipped would do.  The patched version I created fixes it for dates after 2003.  But the fix is incomplete, it fails to handle Julian halfdate values with the upper bit set.

	paul

> On Nov 10, 2021, at 9:16 AM, Supratim Sanyal <supratim at riseup.net> wrote:
> 
> On 11/10/21 9:01 AM, Paul Koning wrote:
> 
>> Curious.  And you're running those with the actual time, not a date in the past?
> FAL says "27-Oct-34", show date looks correct (assuming not 1921).
> 
> 
> 
> Event type 33.0, Remote file access
> Occurred 27-Oct-34 14:11:53.7 on node 31.28 (RST101)
> Access: Local
> Function: OPEN/Read
> Remote node = 0 (RST101)
> Remote process = 0 1 2 NFT002
> Local process = 17 50 1 FAL006
> User = 50,1
> File accessed = 1 SY:[50,1]INFO.TXT
> 
> $
> $
> $
> $ show date
>   10-Nov-21 02:12 PM
> 
> 
> 
>> I wonder if the copy of DECnet I have is from before some patches that fix some of these issues.
>> 
>> 	paul
>> 
>>> On Nov 10, 2021, at 6:47 AM, Supratim Sanyal <supratim at riseup.net <mailto:supratim at riseup.net>> wrote:
>>> 
>>> On 11/10/21 5:14 AM, Keith Halewood wrote:
>>> 
>>>> I’m running RSTS/E 10.1-L out of the box so to speak. I’m not seeing any error messages.
>>> No such errors on 31.28 (RST101) either.
>>> 
>>> /s.
>>> 
>>>>  
>>>> Keith
>>>>  
>>>> From: owner-hecnet at Update.UU.SE <mailto:owner-hecnet at Update.UU.SE> [mailto:owner-hecnet at Update.UU.SE <mailto:owner-hecnet at Update.UU.SE>] On Behalf Of Tony Nicholson
>>>> Sent: 10 November 2021 08:46
>>>> To: hecnet at update.uu.se <mailto:hecnet at update.uu.se>
>>>> Subject: Re: [HECnet] RSTS/E started emitting "?EVTLOG (BLDNIC) -- %Integer error" messages
>>>>  
>>>> Paul Koning is working on a fix.
>>>>  
>>>> The RSTS/E DECnet event logger has an integer overflow too.  He said in a e-mail to me -
>>>>  
>>>> “ I think I found the problem, yes it's a signed overflow.  The "julian halfdays" used in the event log timestamp overflows on 10-Nov-2021.  I fixed an unsigned issue in the INPUT to the function that converts RSTS date to julian halfday, but I didn't notice the unsigned issue in the formula for the output.”
>>>>  
>>>> Tony
>>>>  
>>>> On Wed, 10 Nov 2021 at 19:14, Wilm Boerhout <wboerhout at gmail.com <mailto:wboerhout at gmail.com>> wrote:
>>>> I am receiving the same messages on the console since today. I am running the “latest” DECnet images that I picked up in July 2020.
>>>> -- 
>>>> Tony Nicholson <tony.nicholson at computer.org <mailto:tony.nicholson at computer.org>>
>>> -- 
>>> <SANYALNET-LABS-QR-CODE.png>
>> 
> -- 
> <SANYALNET-LABS-QR-CODE.png>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20211110/7d02a433/attachment.htm>