Re: ending the HTML parser loop

From: Henrik Frystyk Nielsen <frystyk@w3.org>
To: Mathew Cucuzella <mat@mcg.mot.com>; libwww <www-lib@w3.org>
Date: Wednesday, February 17, 1999 4:44 PM
Subject: Re: ending the HTML parser loop


>At 17:24 2/17/99 -0700, Mathew Cucuzella wrote:
>>I'm using the showtags.c example program for the HTML parser. I want
>>to be able to enter the event loop, parse the page, and then come
>>back to my program. Here is the code I changed:
>>

    ...
>>
>>This seems to work, because I now come back from the HTEventList_loop(),
>>but I'm not sure if it is the right call, or if there is a better
>>way to do this. I would also like to call this code multiple times for
>>different URI's which dumps core right now, but I haven't looked into
>>that problem yet.
>
>Stopping the loop is fine - but I wouldn't terminate libwww (delete the
>profile) if I were you if you know that you are entering it again later.
>
>That may also be the problem with the core.
>
>Henrik
>--


    I do something similar, but I delete the request and cleanup the
    library when I come back from the loop. Also, I start other requests
    in the loop (i.e. to load images) and call the HTEventList_stopLoop
    when I don't have any more requests (well, this means I DO delete
    the request in the terminat_handler).

    I cleanup the library when I'm done, and restart it to process new
    requests. Everything seems to basically work, but I'm currently having
    problems in Windows (not crashes though).

    Mathew, if you can be more specific on which kind of crashes and
    problems you are having, I can tell you if they resemble mine (and
    in that case, there could be something related to libwww).

-- Raffaele

Received on Wednesday, 17 February 1999 20:29:29 UTC