[Grub-dev] problem of understanding "crawl-corruption" - anyone care to explain?

[Balinny]

ab wrote:
> i am fighting with this problem of so-called corrupted crawls.
>
> i have already extensively reported and tried to track down reasons and 
> details about this issue, but is somehow really dont understand this 
> whole idea of how a set of httpserver-replies can get "corrupted", and 
> therefore denied from the soap-server (upload).
>
> i would be really glad if someone would care to enlighten me:
>
> see comments onwards from:
> <http://dev.grub.org/cgi-bin/trac.cgi/ticket/9#comment:31>
>
> this "corruption" is driving me crazy, and the grub-crawler is simply a 
> failed experiment to me at its present state and design if its really 
> supposed to be like this. :(
>
> cheers
Maybe the server is giving  wrong headers? My client had problems when 
the content-type was
given empty, or there were \n in it. I now verify the MIME content-type 
and set it to application/x.wrong.content-type
if it's not acceptable. Haven't had servers problems any more using 
valid workunits.
If you send me the workunit and resulting arc.gz file, i'll take it a 
look. It's a bit hard wondering what could have gone
wrong.
Or you could try the C client ;-)