<* PRAGMA LL *>

INTERFACE WebVBT;

IMPORT HTML, HTMLVBT, TextEditVBT, TextList, PixmapVBT, VBT, Web;

CONST
  DefaultURL = "http://www.research.digital.com/SRC/home.html";

TYPE
  T <: Public;
  Private <: VBT.T;
  Public = Private OBJECT
             <* LL = VBT.mu *>
             url : TEXT := NIL;
             page: Page := NIL;
           METHODS
             <* LL <= VBT.mu *>
             init (): T;

             <* LL = VBT.mu *>
             fetch (url      : TEXT    := DefaultURL;
                    style    : Style   := Style.Normal;
                    zippers  : BOOLEAN := FALSE;
                    reload   : BOOLEAN := FALSE;
                    server   : Web.T   := NIL;
                    scrollBar: BOOLEAN := TRUE);
             fromText (contents      : TEXT;
                       contentType   : Web.MIMEType := Web.MIMEType.Text;
                       contentSubType: TEXT         := "html";
                       url           : TEXT         := "text:";
                       style         : Style        := Style.Normal;
                       zippers       : BOOLEAN      := FALSE;
                       reload        : BOOLEAN      := FALSE;
                       server        : Web.T        := NIL;
                       scrollBar     : BOOLEAN      := TRUE);

             stop ();
             getLinks(): TextList.T;
             search(pattern: TEXT): BOOLEAN;

             (* the following are intended to be overridden by clients: *)
             ready (remImages: CARDINAL);

             hotlink (link: TEXT; READONLY cd: VBT.MouseRec);
             ismap   (link: TEXT; READONLY cd: VBT.MouseRec);
             isindex (typein: TEXT);
             form    ();
           END;

  Style =
    {Ugly (* text-only, fixed font *),
     NoImages (* multiple fonts, but no images *),
     Normal (* multiple fonts, images fetched before anything displayed *),
     Background                  (* images are fetched and displayed in the
                                    background *)
    };

The method v.fetch(url, style, reload, server) (eventually) displays the contents of url. More precisely, v.fetch cancels any fetching in progress and then forks a thread to actually retrieve the contents of url, passing to Web.Fetch the reload and server arguments. After the contents of url are fetched (and, in the case of an HTML page, also parsed), the following events take place atomically locked by VBT.mu: the url and page fields are set, the VBT displaying the url (i.e., page.vbt) is updated, and the v.ready(0) method is invoked.

Almost. When fetch is called with style=Background, the ready method is called before any images are fetched. The parameter to ready is the number of images remaining to be fetched; the ready method will be called repeatedly, each time that an image is fetched.

Not quite. The style parameter given to the fetch method allows 4 different styles of fetching: When style=Ugly, a page is displayed with no graphics, using a single font. This is pretty ugly, but it's probably fast. For all other styles, the page is displayed using multiple fonts. When style=NoImages, no attempt is made to display images; the textual alt field is displayed, if one is given in the html. When style=Normal (the default), images are fetched (using the reload and server arguments passed to fetch) and displayed before the page is displayed. Finally, when style=Background, the page is displayed before all the images have been retrieved. As images are retrieved, v.ready(ct) is called repeatedly, with ct set to the number of images left to retrieve.

The method v.stop() is used to cancel any previous call to v.fetch that has not yet caused v.ready(0) to be called.

The ready method is called after a URL fetched by a call to v.fetch(...) has been fetched and displayed. As described above, when fetch is called with style equal to Ugly, NoImages, or Normal, v.ready(0) is simply called. However, if fetch is called with style equal to Background, the ready method is called repeatedly, with a descreasing number. The ready method is not called if the call to fetch was pre-empted by another call to fetch or the call was stopped by a call to v.stop(). The default ready method is a noop.

The method v.getLinks returns a list of URLs contained in v. This list is NIL if v is not an HTML page or if fetching v is still pending.

The method v.search(pattern) returns whether pattern appears in the text displayed by v. The search is case-insensitive.

The method v.hotlink(link, cd) is called whenever the user has clicked on a link in an HTML page. The default hotlink method is a noop.

TYPE

  Page = Web.Page OBJECT END;

  TextPage = Page OBJECT
    vbt: TextEditVBT.T
  END;

  HTMLPage = Page OBJECT
     html: HTML.T;
     vbt: HTMLVBT.T;
  END;

  ImagePage = Page OBJECT
     vbt: PixmapVBT.T;
  END;

END WebVBT.