Re: lynx-dev Re: Lynx-dev Using Win95...THANKS!

[Klaus Weide]

I have played a bit with a jump file (on Linux) and looked at the code,
but not tested anything under Windows.


On Thu, 1 Apr 1999, Leonid Pauzner wrote:
> 31-Mar-99 13:46 Robert D. Dedman wrote:

> > Having problem getting JUMP to work.
> > The .cfg file says some jump samples are included, but they're not.

They are included in the source package (and should be included in any
binary package, definitely if the lynx.cfg that comes with it says so!)

If you don't have them, see under
    <http://sol.slcc.edu/lynx/current/lynx2-8-2/samples/>.

> > 1. On pressing <j> get:"can't locate jump file"
> 
> > 2. This is what I've typed in the lynx.cfg file:
> 
> > JUMPFILE:/c:\progra~1\Lynx/jumps.html
> 
> > 2. I've also tried putting the jumps.html file in the "C:\" directory
> > and typing alternatively:
> 
> > JUMPFILE:/c:\jumps.html
>            ^^ try removing leading slash before the drive letter.

Yes.  The path in JUMPFILE should be given in native file syntax.
(At least I didn't find any place where the name is converted in
any way; a comment in lynx.cfg says "On VMS, use Unix SHELL syntax",
not sure how this can work for VMS if it does.)  It is fed directly
to open() [! not fopen() !].  

> > ...but get the same message.
> 
> 
> > 3. Finally, I don't understand this:
> 
> > "...Make sure your jumps file includes a '?' shortcut for a
> > file://localhost URL to itself:
> > <dt>?<dd><a href="c:\progra~1\Lynx\jumps.html">This Shortcut
> > List</a>..."
> 
> > Does it mean:
> NO idea.
> 
> > (a) create a file "jumps.html"
> > (b) enter "<dt>?<dd><a href="c:\progra~1\Lynx\jumps.html">This Shortcut
> > List</a>"
> > as HTML code
> 
> > or what?

Basically yes, see the sample jumpsUnix.html, but you'll have to change
it appropriately for Windows.

The href value should be given in URL syntax, not in native filename
syntax, so it will look quite different from what you put in lynx.cfg's
JUMPFILE even if both refer to the same file.

For example your (b) should probably be
<dt>?<dd><a href="file://localhost/c:/progra%7E1/Lynx/jumps.html">This Shortcut 
List</a>

I have also URL-escaped the '~' character, that *should* work for Windows
and could avoid potential problems.  (Although the problem may not exist
for '~' in full URLs.)

The format of jump files may be very sensitive, maybe more so than
bookmark files.  Basically they are parsed in two completely different
ways, and both methods are supposed to extract the same URLs.  So keep
your file(s) as close as possible to the sample file, unless you want to
experiment.

One thing I haven't seen documented anywhere: It seems that the "keys"
(shortcut names, what's after the <dt>) have to be sorted (case-insensitive)
otherwise they may not be found at a sortcut prompt.

I still don't quite get what the whole mechanism is good for.
Every shortcut leads to something that can be expressed as a URL;
it seems easier to just put them into an HTML file (with much more
freedom in formatting) and just navigate to the right link and
activate it, rather than learning shortcut names for those URLs.

   Klaus