[Grub-dev] do we even really need a native client

jer jeremie at jabber.org
Fri Jan 11 05:16:47 UTC 2008


>> So, I think you're right and it's missing a \n, but maybe it's
>> missing TWO of them?
>>
>> doc == <nl><URL-record><nl><network_doc>
>>
>> URL-record-v1 == <url><sp>
>> <ip-address><sp>
>> <archive-date><sp>
>> <content-type><sp>
>> <length><nl>
>>
>> So, there should be a \n before each URL record, and two of them
>> after it, one defined as the terminator in URL-record-v1, and one
>> defined as the separator between URL-record and network_doc.  Is that
>> correct?
>>
>> print $arc "\nhttp://$host$path $ip 19691231175959 $ctype",length
>> ($body),"\n\n$body";
>>
>> Is that correct?  Can anyone else verify?
>>
> So it seems.

Anyone else can verify this is correct?  \n URL-stuff \n \n CONTENT ?

>> The workunits can (someday) start to define HTTP/1.1 with a
>> Connection: close, and an Accept-encoding: gzip.  A client supporting
>> the current workunit format shouldn't care or know any different,  
>> right?
>>
> The client's bandwidth might care ;-)

Yep, easy enough to add these headers in the workunits as well :)

>> Doh!  My bad, I can fix it when I generate some more workunits :)
> Aren't they generated on-the-fly?

Heh, nope, there's no DB in this back-end so it's much faster and  
easier to pre-generate batches of these from flat lists right now.

Jer


More information about the Grub-dev mailing list