[ut2004] 64 bit client problems with downloads

Tom Emerson osnut at pacbell.net
Sat Nov 20 00:15:20 EST 2004


On Thursday 18 November 2004 9:05 am, Ryan C. Gordon wrote:
> > of course, from my point of view, I have to ask, "why is there a bug in
> > the first place?" this is pretty simple stuff we're talking about: you're
> > getting a file, you know how big it is, and you know how many bytes
> > you've received -- not rocket science...
>
> Easy to say from where you're sitting.

OK, I started this, so I'll bite: what is it that I don't know about this 
situation?  Do you or do you not know how big the file is supposed to be that 
you are receiving?  Is it possible that the actual number of bytes returned 
by a "get" operation on a socket does not equal the reported number of bytes 
received?  Is there "supposed to be..." some sort of compression algorithm in 
place that isn't actually there?  [this would explain a lot -- for example, 
suppose you expect to receive 100 bytes of compressed data, but in 
uncompressed form the file is 600 bytes -- physically receiving 600 bytes 
when you expected to receive 100 would indeed report rather oddly]

And yes, I *have* done socket programming on "something other than an intel 
CPU" [HP's PA-Risc, to be exact], so I'm not exactly playing armchair 
quarterback here -- unless there is something grotesque about the way the 
original programmers wrote the code, converting a "percent done" routine 
should NOT be hard to do.  [of course, this leaves open the possibility that 
because it isn't hard to do, it hasn't been a priority to actually do it -- 
if you tell me that's the case, I'll shut up now ;) ]

I'll echo Clint's comment about the fact that the counter does get to "100%" 
generally pretty quickly, then restarts and runs "noticably slower" and has, 
in fact, reached 800% or more before announcing the download is complete [and 
I suspect there have been more than a few cases where the "download" took 
longer than the actual game took, so since the server "switched over" to a 
new map, the current download aborted...]  If indeed the file was "fully 
received" at 100%, but then retries from "some other server", what can *I* do 
to help debug and/or verify this situation?

Tom 

(yes, I really do want to help -- some of these maps sound really cool!)



More information about the ut2004 mailing list