Re: lynx-dev Curious: is HTParse efficient?

[Leonid Pauzner]

Yes, there are lots of HTParse() call in mainloop and other places,
perhaps url should became a struct of prefix/host/path/etc.,
not just char* as it is now.
I mean `address' field in DocAddress and document structures,
it would be nice if we could parse url only once.


19-Jul-2000 09:34 Henry Nelson wrote:
>>              Lynx Trace Log (2.8.3rel.1)
> [...]
>> Entered HTAnchor_findAddress
>> New anchor 0x81fe6a0 has hash 92 and address `http://www.staples.ca/'
>> HTAccess: loading document http://www.staples.ca/
>> HTParse: aName:`http://www.staples.ca/'
>>    relatedName:`file:'
>> HTParse:      result:http
>> HTParse: aName:`http://www.staples.ca/'
>>    relatedName:`'
>> HTParse:      result:www.staples.ca
>> HTParse: aName:`http://www.staples.ca/'
>>    relatedName:`'
>> HTParse:      result:http
>> HTParse: aName:`http://www.staples.ca/'
>>    relatedName:`'
>> HTParse:      result:www.staples.ca

> Perhaps there is no way around it or it's some safety mechanism, but
> HTParse seems to do a lot of redundant work.  In the above example,
> while the previous section's starting point is different, the result
> is the same "http", such that
>       HTParse:      result:http
>       HTParse: aName:`http://www.staples.ca/'
>          relatedName:`'
>       HTParse:      result:www.staples.ca
> appears twice.

> __Henry

> ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden




; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden