Re: lynx-dev LYNX: meaning of "Bad partial reference; stripping leading dots"?

[David Combs]

On Wed, Jul 26, 2000 at 12:42:14AM -0700, pAb-032871 wrote:
> In "lynx-dev LYNX: meaning of "Bad partial reference; stripping leading 
> dots"?"
> [25/Jul/2000 Tue 18:40:04]
> David Combs wrote:
> 
> > Man, do I see this ALL the time.
> > 
> > Even O'Reilly's site.
> > 
> > Exactly what is screwy in the html?
> 
> It's not the HTML, I think.  Check your settings in lynx.cfg
> 
> # If STRIP_DOTDOT_URLS is TRUE, Lynx emulates the invalid behavior of many
> # browsers to strip a leading "../" segment from relative URLs in HTML
> # documents with a http or https base URL, if this would otherwise lead to
> # an absolute URLs with those characters still in it.  Such URLs are normally
                                           ^^^^^^^^^^^

Please, someone, show how this could happen -- how following the 
../ (whose meaning is quite clear) could result in an absolute
url WITH THOSE CHARS STILL IN IT.

Of course, thanks for the explanations thus far, but 
that above sentence in the .cfg confuses me, for sure.


I deliberately do not remove the following stuff from 
the prior email, since one of the examples there might
be a good "existing" piece of text from which to show
this "with those characters STILL in it" possibility.

David


> # erroneous and not what is intended by page authors.  Lynx will issue
> # a warning message when this occurs.
> #
> # If STRIP_DOTDOT_URLS is FALSE, Lynx will use those URLs for requests
> # without taking any special actions or issuing Warnings, in most cases
> # this will result in an error response from the server.
> #
> # Note that Lynx never tries to fix similar URLs for protocols other than
> # http and https, since they are less common and may actually be valid in
> # some cases.
> #
> #STRIP_DOTDOT_URLS:TRUE
> 
> I'm not sure what the compilation default is though.
> 
> 
> Here are two relative link examples.  Let's take the *absolute*
> URL to be;
>       http://www.serv.dom/~usr/subdirectory/index.html
> 
>       <A HREF="../top_index.html">My Main Page</A>
> Actually points to;
>       http://www.serv.dom/~usr/top_index.html
> "../" is actually shorthand for "up one directory"
> 
>       <A HREF="./other_file.html">Other File</A>
> Actually points to;
>       http://www.serv.dom/~usr/subdirectory/other_file.html
> 
> With the leading "./" stripped, it would also point to;
>       http://www.serv.dom/~usr/subdirectory/other_file.html
> 
> 
> On many FTP servers, you may notice two "directories" [I think
> that's what they are] in every listing, named "." and ".."  Entering
> "." puts you in the current directory [possibly a refreshed listing?],
> and ".." moves you up one directory.  The system might put these
> here automatically for some reason, I don't know much about it.
> They're in my own login directory and I didn't put them there.../
> 
> BTW, if this is an invalid way of inserting relative links to
> a shorter filepath, what about <A HREF="/~usr/top_index.html">
> and <A HREF="/"> then?
> 
> 
> 
>                         Patrick
>               <mailto:address@hidden>
>  
> 
> ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden

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