10 Nov
2021
10 Nov
'21
6:20 p.m.
It's a very obscure form of flagging... :-)
Well, occasionally, the session itself also is borked and lost, which
could be a hint as well. (Actually, I think it goes hand in hand every
time. When the session is borked, some memory is also lost.)
So yeah. I consider it a bug that RSX isn't handling this better.
But I definitely also claim that latd is buggy and not following the
protocol. Sure, it's understandable that it might not since they didn't
have the protocol spec, but no matter what, it is a problem that the
implementation isn't doing things right.
Johnny
On 2021-11-10 18:10, Thomas DeBellis wrote:
Heh...? I wouldn't consider it a bug; I'd say
it's more of a misfeature:
RSX is flagging an occasional malformed LAT packet by running out of memory.
Oh, /OK/; perhaps the reporting paradigm (going catatonic) could be
enhanced...
Of course, that begs the question of whether I'm being like most
reasonable people, being reasonable at all, or just plain being silly...? ?
On 11/8/21 5:20 PM, Dave McGuire wrote:
> On 11/8/21 4:49 PM, Johnny Billquist wrote:
>> Just as a warning, the lat implementation have some bugs. I've never
>> spent the time to figure it out, but at least when talking to RSX
>> hosts, it triggers some bug in RSX which cause it to leak memory, and
>> eventually RSX becomes catatonic.
>
> ? I think most reasonable people would consider that a bug in RSX, not
> a bug in the Linux LAT implementation. ;)
>
> ???????????? -Dave
>
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol