Re: Xnews: Smart decoding doesnot work too well

blah <blah@xxxxxxxx> wrote in
news:44f46801$0$26631$a82e6bb9@xxxxxxxxxxxxxxxxxxx:

vmw12@xxxxxx wrote in news:44f417f0$0$1376$db0fefd9@xxxxxxxxxxxxxx:

Snipped from Xnews
"you can queue the whole thing normally and Xnews should figure it
out, but the first way may be safer.)"
I understand this to mean that the normal method works usually. My
experience is that it never works as stated in the manual.

Well, yeah IIRC, emphasis on 'usually'.
That's why I always re-queue just the unread parts.
I think there's mention of the 'loop' syndrome somewhere in the manual
or changes.txt Ah yes, I find it for thee (if thou hast not readith it
already) - (it's in changes.txt):
-------------------------------
quote:
6.12.01:

Hi all :)

When a yEnc(oded) article has a CRC error, the program gets stuck in a
loop where it keeps trying to decode the same file if you have
alternate server(s) defined. I tried an hour trying to fix this bug
and gave up (it's frightening, but I keep finding new depths of
dumbosity within). Instead, I decided to Deal With the Problem By Not
Dealing With the Problem(tm). So, if you've been plagued by this bug,
add this to Xnews.ini under [Misc]

IgnoreYEncCRCError=1

Voila!

Actually, this works out pretty well since most of the time, the file
is still usable with PAR2 and what not.

Happy holidays :)

unquote
------------------------------------


If i am at the computer whilst downloading and notice it I do as
mentioned
above. That is not much of a problem, problem is when the download is
unattended, you can back and may find,

multiple instances of partially downloaded files.

Hope the above fix works for you.


If smart decoding is turned off, Xnews will download only upto the
interrupted part, ie it wont skip the interrupted part and carry on
to the end of the fill. Is that correct

I've never turned it off.
Although I do monitor the re-queuing and re-downloading of 'error'
files when I get them. I think I recall that 'loop' happening once but
I was there to stop it.



Spoke too soon!
I had already done, it ages ago when i first saw that fix.
Hasn't worked for me, i'm afraid

my Xnews.ini misc is here (news host blanked out)
[Misc]
SMTPHost=mailhost.xxxx.xxxx.xxx
SMTPUserName=
SMTPPasswd=
LaunchTypes=.BMP|.GIF
SaveTypes=.PNG|.JPEG|.JPG|.BMP|.GIF
GotReplySound=
FileDoneSound=
AutoGetMsgCounts=1
QuoteRegex=^\s?\s?\s?[>:|\]»]
MaxXPostKill=1
KeptScore=8888
AttachDir=C:\PROGRAM FILES\XNEWS\attachments\
DecodeDir=E:\0GRABIT\
SaveDir=C:\PROGRAM FILES\XNEWS\downloads\
IgnoreReWhenSorting=1
ExpandWhenTag=0
AutoSaveAttachments=0
LoadReadArticles=0
HideEmptyGroups=0
OpenNextGroup=0
ScoreFileLoadOpt=2
UpdateCountInterval=0
StorageExp=0
StorageOpt=0
AutoResumeDelay=10
IgnoreYEncCRCError=1
AutoCollapse=1
CompactPercent=8
UseSmartDecode=1
DejaQuery=

.

Relevant Pages

  • Re: Xnews: Smart decoding doesnot work too well
    ... I think there's mention of the 'loop' syndrome somewhere in the manual or changes.txt ... When a yEncarticle has a CRC error, the program gets stuck in a ... So, if you've been plagued by this bug, add this to ... Hope the above fix works for you. ...
    (news.software.readers)
  • Windows Update Refuses to download updates
    ... Does anyone know how to fix this "Infinite Update Now Loop" bug? ...
    (microsoft.public.windowsupdate)
  • Re: LOOP blows!
    ... that it's not a quine. ... Can we expect version Final.1 to fix that bug? ... Your LOOP is obviously broken. ...
    (comp.lang.lisp)
  • Re: shrinker->nr = LONG_MAX means deadlock for icache
    ... If breaking the loop despite we made progress doesn't risk to make us go ... an huge nr value would over-shrink the caches. ... proposed fix was ok. ... status of the kernel that they used has all nr_unused and other icache ...
    (Linux-Kernel)
  • Re: How do you forth experts drop a break_point ?
    ... Say that I believe the problem is at step 2417 of a do loop. ... rummage through it enough to find this error and fix it, chances are ... There's a strong temptation not to do that for commercial code because ... Much more pleasant to think that one last patch will fix ...
    (comp.lang.forth)