export to html and relative links:

export to html and relative links:

[T. V. Raman]

Hi,

org-export turns links of the form 
[[foo][link to relative url foo]]

ends up creating links of the form 
<a href="#foo">...</a>

this means that it becomes impossible to write hyperlinks that
are relative URLs.

-- 
Best Regards,
--raman

      
Email:  raman@users.sf.net
WWW:    http://emacspeak.sf.net/raman/
AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
Google: tv+raman 
IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[Carsten Dominik]

On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:

> Hi,
>
> org-export turns links of the form
> [[foo][link to relative url foo]]
>
> ends up creating links of the form
> <a href="#foo">...</a>
>
> this means that it becomes impossible to write hyperlinks that
> are relative URLs.

Hi Raman,

the url goes into the first pair of square brackets, the description  
into the second.
If you put something into the first bracket pair that does not look  
lie a url, it will
be treated as an internal link.

Have you tried to use something like

[[http:foo.html][link to relative url]]

?

- Carsten

Re: export to html and relative links:

[T. V. Raman]

I've not tried http:foo.html -- but I suggest doing that to
author a relative URL is a bad idea.

I  tried ./foo.html and that didn't work either. One compromise
would be to get ./foo.html to link to a relative url, while
foo.html continues to link to  a local anchor

>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
    Carsten> 
    >> Hi,
    >> 
    >> org-export turns links of the form [[foo][link to relative
    >> url foo]]
    >> 
    >> ends up creating links of the form <a href="#foo">...</a>
    >> 
    >> this means that it becomes impossible to write hyperlinks
    >> that are relative URLs.
    Carsten> 
    Carsten> Hi Raman,
    Carsten> 
    Carsten> the url goes into the first pair of square brackets,
    Carsten> the description into the second.  If you put
    Carsten> something into the first bracket pair that does not
    Carsten> look lie a url, it will be treated as an internal
    Carsten> link.
    Carsten> 
    Carsten> Have you tried to use something like
    Carsten> 
    Carsten> [[http:foo.html][link to relative url]]
    Carsten> 
    Carsten> ?
    Carsten> 
    Carsten> - Carsten

-- 
Best Regards,
--raman

      
Email:  raman@users.sf.net
WWW:    http://emacspeak.sf.net/raman/
AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
Google: tv+raman 
IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[Carsten Dominik]

On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:

> I've not tried http:foo.html -- but I suggest doing that to
> author a relative URL is a bad idea.

Can you explain why you think that this is a bad idea?
Educate me!  What is wrong with writing http:foo.html ??????

- Carsten

>
>
> I  tried ./foo.html and that didn't work either. One compromise
> would be to get ./foo.html to link to a relative url, while
> foo.html continues to link to  a local anchor
>
>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
>    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
>    Carsten>
>>> Hi,
>>>
>>> org-export turns links of the form [[foo][link to relative
>>> url foo]]
>>>
>>> ends up creating links of the form <a href="#foo">...</a>
>>>
>>> this means that it becomes impossible to write hyperlinks
>>> that are relative URLs.
>    Carsten>
>    Carsten> Hi Raman,
>    Carsten>
>    Carsten> the url goes into the first pair of square brackets,
>    Carsten> the description into the second.  If you put
>    Carsten> something into the first bracket pair that does not
>    Carsten> look lie a url, it will be treated as an internal
>    Carsten> link.
>    Carsten>
>    Carsten> Have you tried to use something like
>    Carsten>
>    Carsten> [[http:foo.html][link to relative url]]
>    Carsten>
>    Carsten> ?
>    Carsten>
>    Carsten> - Carsten
>
> -- 
> Best Regards,
> --raman
>
>
> Email:  raman@users.sf.net
> WWW:    http://emacspeak.sf.net/raman/
> AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
> PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
> Google: tv+raman
> IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[T. V. Raman]

I think writing http:foo.html is a bad idea because typing that
string in other contexts is sort of meaningless as a URL.
Until now, everything one types in org-mode sort of has meaning
elsewhere. If you want it to look like a url in this case too
then I'd suggest file:foo.html -- rther than http:file.html.

I  did not suggest that earlier because it's more chars to type,
but that's not a huge concern since emacs does the typing, and we
could always augment c-c c-l to enable authoring of relative
links with file completion.

so file:a/b/c.html would become href="a/b/c.html" in the
generated html.

>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
    Carsten> 
    >> I've not tried http:foo.html -- but I suggest doing that
    >> to author a relative URL is a bad idea.
    Carsten> 
    Carsten> Can you explain why you think that this is a bad
    Carsten> idea?  Educate me!  What is wrong with writing
    Carsten> http:foo.html ??????
    Carsten> 
    Carsten> - Carsten
    Carsten> 
    >> 
>
> I tried ./foo.html and that didn't work either. One compromise
    >> would be to get ./foo.html to link to a relative url,
    >> while foo.html continues to link to a local anchor
    >> 
    >>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
    >>>>>>> writes:
    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
    Carsten> 
    >>>> Hi,
    >>>> 
    >>>> org-export turns links of the form [[foo][link to
    >>>> relative url foo]]
    >>>> 
    >>>> ends up creating links of the form <a
    >>>> href="#foo">...</a>
    >>>> 
    >>>> this means that it becomes impossible to write
    >>>> hyperlinks that are relative URLs.
    Carsten> 
    Carsten> Hi Raman,
    Carsten> 
    Carsten> the url goes into the first pair of square brackets,
    Carsten> the description into the second.  If you put
    Carsten> something into the first bracket pair that does not
    Carsten> look lie a url, it will be treated as an internal
    Carsten> link.
    Carsten> 
    Carsten> Have you tried to use something like
    Carsten> 
    Carsten> [[http:foo.html][link to relative url]]
    Carsten> 
    Carsten> ?
    Carsten> 
    Carsten> - Carsten
    >> 
    >> -- 
    >> Best Regards, --raman
    >> 
    >> 
    >> Email: raman@users.sf.net WWW:
    >> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
    >> tv.raman.tv@gmail.com PGP:
    >> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
    >> tv+raman IRC: irc://irc.freenode.net/#emacs

-- 
Best Regards,
--raman

      
Email:  raman@users.sf.net
WWW:    http://emacspeak.sf.net/raman/
AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
Google: tv+raman 
IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[Carsten Dominik]

Hmmmm,

this is *such* a good idea, that I will implement this retroactively
into all versions since ... 2.0 or so?  Abracadabra! Done.

In fact, it has been working just so for a very long time.

file:foo.html   -> href="foo.html"
file:foo.org    -> href="foo.html"

The second line assumes that the org file foo.org will also
be converted to an html file in the publishing process....

Also, I was mistaken earlier:

http:foo.html  becomes   href="http:foo.html" and not href="foo.html"

so this actually cannot be used for a local link.

So what remains from this exchange?  Should we accept links
that look like absolute file names (including ./foo.html) as
equivalent to file:./foo.html ?

- Carsten

On Apr 13, 2008, at 6:25 PM, T. V. Raman wrote:

>
> I think writing http:foo.html is a bad idea because typing that
> string in other contexts is sort of meaningless as a URL.
> Until now, everything one types in org-mode sort of has meaning
> elsewhere. If you want it to look like a url in this case too
> then I'd suggest file:foo.html -- rther than http:file.html.
>
> I  did not suggest that earlier because it's more chars to type,
> but that's not a huge concern since emacs does the typing, and we
> could always augment c-c c-l to enable authoring of relative
> links with file completion.
>
> so file:a/b/c.html would become href="a/b/c.html" in the
> generated html.
>
>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
>    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
>    Carsten>
>>> I've not tried http:foo.html -- but I suggest doing that
>>> to author a relative URL is a bad idea.
>    Carsten>
>    Carsten> Can you explain why you think that this is a bad
>    Carsten> idea?  Educate me!  What is wrong with writing
>    Carsten> http:foo.html ??????
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>>>
>>
>> I tried ./foo.html and that didn't work either. One compromise
>>> would be to get ./foo.html to link to a relative url,
>>> while foo.html continues to link to a local anchor
>>>
>>>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
>>>>>>>> writes:
>    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
>    Carsten>
>>>>> Hi,
>>>>>
>>>>> org-export turns links of the form [[foo][link to
>>>>> relative url foo]]
>>>>>
>>>>> ends up creating links of the form <a
>>>>> href="#foo">...</a>
>>>>>
>>>>> this means that it becomes impossible to write
>>>>> hyperlinks that are relative URLs.
>    Carsten>
>    Carsten> Hi Raman,
>    Carsten>
>    Carsten> the url goes into the first pair of square brackets,
>    Carsten> the description into the second.  If you put
>    Carsten> something into the first bracket pair that does not
>    Carsten> look lie a url, it will be treated as an internal
>    Carsten> link.
>    Carsten>
>    Carsten> Have you tried to use something like
>    Carsten>
>    Carsten> [[http:foo.html][link to relative url]]
>    Carsten>
>    Carsten> ?
>    Carsten>
>    Carsten> - Carsten
>>>
>>> -- 
>>> Best Regards, --raman
>>>
>>>
>>> Email: raman@users.sf.net WWW:
>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>> tv.raman.tv@gmail.com PGP:
>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>
> -- 
> Best Regards,
> --raman
>
>
> Email:  raman@users.sf.net
> WWW:    http://emacspeak.sf.net/raman/
> AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
> PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
> Google: tv+raman
> IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[T. V. Raman]

two suggestions:

1)  accept paths like "./foo.html" as local links.
2) Augment C-c C-l to react to file: by providing filename
completion relative to the working directory.

The emacs binding to the w3m browser does this if you type file: in the minibuffer
when prompted for a URL.

>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
    Carsten> Hmmmm, this is *such* a good idea, that I will
    Carsten> implement this retroactively into all versions since
    Carsten> ... 2.0 or so?  Abracadabra! Done.
    Carsten> 
    Carsten> In fact, it has been working just so for a very long
    Carsten> time.
    Carsten> 
    Carsten> file:foo.html -> href="foo.html" file:foo.org ->
    Carsten> href="foo.html"
    Carsten> 
    Carsten> The second line assumes that the org file foo.org
    Carsten> will also be converted to an html file in the
    Carsten> publishing process....
    Carsten> 
    Carsten> Also, I was mistaken earlier:
    Carsten> 
    Carsten> http:foo.html becomes href="http:foo.html" and not
    Carsten> href="foo.html"
    Carsten> 
    Carsten> so this actually cannot be used for a local link.
    Carsten> 
    Carsten> So what remains from this exchange?  Should we
    Carsten> accept links that look like absolute file names
    Carsten> (including ./foo.html) as equivalent to
    Carsten> file:./foo.html ?
    Carsten> 
    Carsten> - Carsten
    Carsten> 
    Carsten> On Apr 13, 2008, at 6:25 PM, T. V. Raman wrote:
    Carsten> 
    >> 
> I think writing http:foo.html is a bad idea because typing that
    >> string in other contexts is sort of meaningless as a URL.
    >> Until now, everything one types in org-mode sort of has
    >> meaning elsewhere. If you want it to look like a url in
    >> this case too then I'd suggest file:foo.html -- rther than
    >> http:file.html.
    >> 
    >> I did not suggest that earlier because it's more chars to
    >> type, but that's not a huge concern since emacs does the
    >> typing, and we could always augment c-c c-l to enable
    >> authoring of relative links with file completion.
    >> 
    >> so file:a/b/c.html would become href="a/b/c.html" in the
    >> generated html.
    >> 
    >>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
    >>>>>>> writes:
    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
    Carsten> 
    >>>> I've not tried http:foo.html -- but I suggest doing that
    >>>> to author a relative URL is a bad idea.
    Carsten> 
    Carsten> Can you explain why you think that this is a bad
    Carsten> idea?  Educate me!  What is wrong with writing
    Carsten> http:foo.html ??????
    Carsten> 
    Carsten> - Carsten
    Carsten> 
    >>>> 
>>
>> I tried ./foo.html and that didn't work either. One compromise
    >>>> would be to get ./foo.html to link to a relative url,
    >>>> while foo.html continues to link to a local anchor
    >>>> 
    >>>>>>>>> "Carsten" == Carsten Dominik
    >>>>>>>>> <dominik@science.uva.nl> writes:
    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
    Carsten> 
    >>>>>> Hi,
    >>>>>> 
    >>>>>> org-export turns links of the form [[foo][link to
    >>>>>> relative url foo]]
    >>>>>> 
    >>>>>> ends up creating links of the form <a
    >>>>>> href="#foo">...</a>
    >>>>>> 
    >>>>>> this means that it becomes impossible to write
    >>>>>> hyperlinks that are relative URLs.
    Carsten> 
    Carsten> Hi Raman,
    Carsten> 
    Carsten> the url goes into the first pair of square brackets,
    Carsten> the description into the second.  If you put
    Carsten> something into the first bracket pair that does not
    Carsten> look lie a url, it will be treated as an internal
    Carsten> link.
    Carsten> 
    Carsten> Have you tried to use something like
    Carsten> 
    Carsten> [[http:foo.html][link to relative url]]
    Carsten> 
    Carsten> ?
    Carsten> 
    Carsten> - Carsten
    >>>> 
    >>>> -- 
    >>>> Best Regards, --raman
    >>>> 
    >>>> 
    >>>> Email: raman@users.sf.net WWW:
    >>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
    >>>> tv.raman.tv@gmail.com PGP:
    >>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
    >>>> tv+raman IRC: irc://irc.freenode.net/#emacs
    >> 
    >> -- 
    >> Best Regards, --raman
    >> 
    >> 
    >> Email: raman@users.sf.net WWW:
    >> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
    >> tv.raman.tv@gmail.com PGP:
    >> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
    >> tv+raman IRC: irc://irc.freenode.net/#emacs

-- 
Best Regards,
--raman

      
Email:  raman@users.sf.net
WWW:    http://emacspeak.sf.net/raman/
AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
Google: tv+raman 
IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[Carsten Dominik]

Hi Raman,

On Apr 13, 2008, at 8:48 PM, T. V. Raman wrote:

> two suggestions:
>
> 1)  accept paths like "./foo.html" as local links.

I will look into this.  The problem is the system dependence
of file names, so I am not sure what a good solution would be
that would work on Windows as well as on Unix.

> 2) Augment C-c C-l to react to file: by providing filename
> completion relative to the working directory.

Hmm, that is difficult, but thanks for the pointer, maybe we can steel  
some code.

For now, if you use `C-u C-c C-l', you get file completion, and the  
result will
be a "file:" link with a relative path (unless you enter a path  
starting with "/").
Even if you find the file with ../../../some/path, it will still be  
inserted as a relative link.

- Carsten

>
>
> The emacs binding to the w3m browser does this if you type file: in  
> the minibuffer
> when prompted for a URL.
>
>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
>    Carsten> Hmmmm, this is *such* a good idea, that I will
>    Carsten> implement this retroactively into all versions since
>    Carsten> ... 2.0 or so?  Abracadabra! Done.
>    Carsten>
>    Carsten> In fact, it has been working just so for a very long
>    Carsten> time.
>    Carsten>
>    Carsten> file:foo.html -> href="foo.html" file:foo.org ->
>    Carsten> href="foo.html"
>    Carsten>
>    Carsten> The second line assumes that the org file foo.org
>    Carsten> will also be converted to an html file in the
>    Carsten> publishing process....
>    Carsten>
>    Carsten> Also, I was mistaken earlier:
>    Carsten>
>    Carsten> http:foo.html becomes href="http:foo.html" and not
>    Carsten> href="foo.html"
>    Carsten>
>    Carsten> so this actually cannot be used for a local link.
>    Carsten>
>    Carsten> So what remains from this exchange?  Should we
>    Carsten> accept links that look like absolute file names
>    Carsten> (including ./foo.html) as equivalent to
>    Carsten> file:./foo.html ?
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>    Carsten> On Apr 13, 2008, at 6:25 PM, T. V. Raman wrote:
>    Carsten>
>>>
>> I think writing http:foo.html is a bad idea because typing that
>>> string in other contexts is sort of meaningless as a URL.
>>> Until now, everything one types in org-mode sort of has
>>> meaning elsewhere. If you want it to look like a url in
>>> this case too then I'd suggest file:foo.html -- rther than
>>> http:file.html.
>>>
>>> I did not suggest that earlier because it's more chars to
>>> type, but that's not a huge concern since emacs does the
>>> typing, and we could always augment c-c c-l to enable
>>> authoring of relative links with file completion.
>>>
>>> so file:a/b/c.html would become href="a/b/c.html" in the
>>> generated html.
>>>
>>>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
>>>>>>>> writes:
>    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
>    Carsten>
>>>>> I've not tried http:foo.html -- but I suggest doing that
>>>>> to author a relative URL is a bad idea.
>    Carsten>
>    Carsten> Can you explain why you think that this is a bad
>    Carsten> idea?  Educate me!  What is wrong with writing
>    Carsten> http:foo.html ??????
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>>>>>
>>>
>>> I tried ./foo.html and that didn't work either. One compromise
>>>>> would be to get ./foo.html to link to a relative url,
>>>>> while foo.html continues to link to a local anchor
>>>>>
>>>>>>>>>> "Carsten" == Carsten Dominik
>>>>>>>>>> <dominik@science.uva.nl> writes:
>    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
>    Carsten>
>>>>>>> Hi,
>>>>>>>
>>>>>>> org-export turns links of the form [[foo][link to
>>>>>>> relative url foo]]
>>>>>>>
>>>>>>> ends up creating links of the form <a
>>>>>>> href="#foo">...</a>
>>>>>>>
>>>>>>> this means that it becomes impossible to write
>>>>>>> hyperlinks that are relative URLs.
>    Carsten>
>    Carsten> Hi Raman,
>    Carsten>
>    Carsten> the url goes into the first pair of square brackets,
>    Carsten> the description into the second.  If you put
>    Carsten> something into the first bracket pair that does not
>    Carsten> look lie a url, it will be treated as an internal
>    Carsten> link.
>    Carsten>
>    Carsten> Have you tried to use something like
>    Carsten>
>    Carsten> [[http:foo.html][link to relative url]]
>    Carsten>
>    Carsten> ?
>    Carsten>
>    Carsten> - Carsten
>>>>>
>>>>> -- 
>>>>> Best Regards, --raman
>>>>>
>>>>>
>>>>> Email: raman@users.sf.net WWW:
>>>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>>>> tv.raman.tv@gmail.com PGP:
>>>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>>>
>>> -- 
>>> Best Regards, --raman
>>>
>>>
>>> Email: raman@users.sf.net WWW:
>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>> tv.raman.tv@gmail.com PGP:
>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>
> -- 
> Best Regards,
> --raman
>
>
> Email:  raman@users.sf.net
> WWW:    http://emacspeak.sf.net/raman/
> AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
> PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
> Google: tv+raman
> IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[T. V. Raman]

file:// urls are already designed to be platform independent.
So an org file should never carry in it a path like a\\b\\c.html
-- we should always use a/b/c.html
since that's the syntax used by relative URLs.

>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
    Carsten> Hi Raman, On Apr 13, 2008, at 8:48 PM, T. V. Raman
    Carsten> wrote:
    Carsten> 
    >> two suggestions:
    >> 
    >> 1) accept paths like "./foo.html" as local links.
    Carsten> 
    Carsten> I will look into this.  The problem is the system
    Carsten> dependence of file names, so I am not sure what a
    Carsten> good solution would be that would work on Windows as
    Carsten> well as on Unix.
    Carsten> 
    >> 2) Augment C-c C-l to react to file: by providing filename
    >> completion relative to the working directory.
    Carsten> 
    Carsten> Hmm, that is difficult, but thanks for the pointer,
    Carsten> maybe we can steel some code.
    Carsten> 
    Carsten> For now, if you use `C-u C-c C-l', you get file
    Carsten> completion, and the result will be a "file:" link
    Carsten> with a relative path (unless you enter a path
    Carsten> starting with "/").  Even if you find the file with
    Carsten> ../../../some/path, it will still be inserted as a
    Carsten> relative link.
    Carsten> 
    Carsten> - Carsten
    Carsten> 
    >> 
>
> The emacs binding to the w3m browser does this if you type
    >> file: in the minibuffer when prompted for a URL.
    >> 
    >>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
    >>>>>>> writes:
    Carsten> Hmmmm, this is *such* a good idea, that I will
    Carsten> implement this retroactively into all versions since
    Carsten> ... 2.0 or so?  Abracadabra! Done.
    Carsten> 
    Carsten> In fact, it has been working just so for a very long
    Carsten> time.
    Carsten> 
    Carsten> file:foo.html -> href="foo.html" file:foo.org ->
    Carsten> href="foo.html"
    Carsten> 
    Carsten> The second line assumes that the org file foo.org
    Carsten> will also be converted to an html file in the
    Carsten> publishing process....
    Carsten> 
    Carsten> Also, I was mistaken earlier:
    Carsten> 
    Carsten> http:foo.html becomes href="http:foo.html" and not
    Carsten> href="foo.html"
    Carsten> 
    Carsten> so this actually cannot be used for a local link.
    Carsten> 
    Carsten> So what remains from this exchange?  Should we
    Carsten> accept links that look like absolute file names
    Carsten> (including ./foo.html) as equivalent to
    Carsten> file:./foo.html ?
    Carsten> 
    Carsten> - Carsten
    Carsten> 
    Carsten> On Apr 13, 2008, at 6:25 PM, T. V. Raman wrote:
    Carsten> 
    >>>> 
>> I think writing http:foo.html is a bad idea because typing
    >>>> that string in other contexts is sort of meaningless as
    >>>> a URL.  Until now, everything one types in org-mode sort
    >>>> of has meaning elsewhere. If you want it to look like a
    >>>> url in this case too then I'd suggest file:foo.html --
    >>>> rther than http:file.html.
    >>>> 
    >>>> I did not suggest that earlier because it's more chars
    >>>> to type, but that's not a huge concern since emacs does
    >>>> the typing, and we could always augment c-c c-l to
    >>>> enable authoring of relative links with file completion.
    >>>> 
    >>>> so file:a/b/c.html would become href="a/b/c.html" in the
    >>>> generated html.
    >>>> 
    >>>>>>>>> "Carsten" == Carsten Dominik
    >>>>>>>>> <dominik@science.uva.nl> writes:
    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
    Carsten> 
    >>>>>> I've not tried http:foo.html -- but I suggest doing
    >>>>>> that to author a relative URL is a bad idea.
    Carsten> 
    Carsten> Can you explain why you think that this is a bad
    Carsten> idea?  Educate me!  What is wrong with writing
    Carsten> http:foo.html ??????
    Carsten> 
    Carsten> - Carsten
    Carsten> 
    >>>>>> 
>>>
>>> I tried ./foo.html and that didn't work either. One
    >>>>>> compromise would be to get ./foo.html to link to a
    >>>>>> relative url, while foo.html continues to link to a
    >>>>>> local anchor
    >>>>>> 
    >>>>>>>>>>> "Carsten" == Carsten Dominik
    >>>>>>>>>>> <dominik@science.uva.nl> writes:
    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
    Carsten> 
    >>>>>>>> Hi,
    >>>>>>>> 
    >>>>>>>> org-export turns links of the form [[foo][link to
    >>>>>>>> relative url foo]]
    >>>>>>>> 
    >>>>>>>> ends up creating links of the form <a
    >>>>>>>> href="#foo">...</a>
    >>>>>>>> 
    >>>>>>>> this means that it becomes impossible to write
    >>>>>>>> hyperlinks that are relative URLs.
    Carsten> 
    Carsten> Hi Raman,
    Carsten> 
    Carsten> the url goes into the first pair of square brackets,
    Carsten> the description into the second.  If you put
    Carsten> something into the first bracket pair that does not
    Carsten> look lie a url, it will be treated as an internal
    Carsten> link.
    Carsten> 
    Carsten> Have you tried to use something like
    Carsten> 
    Carsten> [[http:foo.html][link to relative url]]
    Carsten> 
    Carsten> ?
    Carsten> 
    Carsten> - Carsten
    >>>>>> 
    >>>>>> -- 
    >>>>>> Best Regards, --raman
    >>>>>> 
    >>>>>> 
    >>>>>> Email: raman@users.sf.net WWW:
    >>>>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
    >>>>>> tv.raman.tv@gmail.com PGP:
    >>>>>> http://emacspeak.sf.net/raman/raman-almaden.asc
    >>>>>> Google: tv+raman IRC: irc://irc.freenode.net/#emacs
    >>>> 
    >>>> -- 
    >>>> Best Regards, --raman
    >>>> 
    >>>> 
    >>>> Email: raman@users.sf.net WWW:
    >>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
    >>>> tv.raman.tv@gmail.com PGP:
    >>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
    >>>> tv+raman IRC: irc://irc.freenode.net/#emacs
    >> 
    >> -- 
    >> Best Regards, --raman
    >> 
    >> 
    >> Email: raman@users.sf.net WWW:
    >> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
    >> tv.raman.tv@gmail.com PGP:
    >> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
    >> tv+raman IRC: irc://irc.freenode.net/#emacs

-- 
Best Regards,
--raman

      
Email:  raman@users.sf.net
WWW:    http://emacspeak.sf.net/raman/
AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
Google: tv+raman 
IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[Carsten Dominik]

On Apr 13, 2008, at 10:32 PM, T. V. Raman wrote:
> file:// urls are already designed to be platform independent.
> So an org file should never carry in it a path like a\\b\\c.html
> -- we should always use a/b/c.html
> since that's the syntax used by relative URLs.


Yes, but while an Org file is still an Org file, the same
links function is links to open the corresponding file in
Emacs, and my fear is that this can lead to conflicts.

Anyway, this is for the windows people to complain about
and to propose solutions... :-)

- Carsten

>
>
>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
>    Carsten> Hi Raman, On Apr 13, 2008, at 8:48 PM, T. V. Raman
>    Carsten> wrote:
>    Carsten>
>>> two suggestions:
>>>
>>> 1) accept paths like "./foo.html" as local links.
>    Carsten>
>    Carsten> I will look into this.  The problem is the system
>    Carsten> dependence of file names, so I am not sure what a
>    Carsten> good solution would be that would work on Windows as
>    Carsten> well as on Unix.
>    Carsten>
>>> 2) Augment C-c C-l to react to file: by providing filename
>>> completion relative to the working directory.
>    Carsten>
>    Carsten> Hmm, that is difficult, but thanks for the pointer,
>    Carsten> maybe we can steel some code.
>    Carsten>
>    Carsten> For now, if you use `C-u C-c C-l', you get file
>    Carsten> completion, and the result will be a "file:" link
>    Carsten> with a relative path (unless you enter a path
>    Carsten> starting with "/").  Even if you find the file with
>    Carsten> ../../../some/path, it will still be inserted as a
>    Carsten> relative link.
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>>>
>>
>> The emacs binding to the w3m browser does this if you type
>>> file: in the minibuffer when prompted for a URL.
>>>
>>>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
>>>>>>>> writes:
>    Carsten> Hmmmm, this is *such* a good idea, that I will
>    Carsten> implement this retroactively into all versions since
>    Carsten> ... 2.0 or so?  Abracadabra! Done.
>    Carsten>
>    Carsten> In fact, it has been working just so for a very long
>    Carsten> time.
>    Carsten>
>    Carsten> file:foo.html -> href="foo.html" file:foo.org ->
>    Carsten> href="foo.html"
>    Carsten>
>    Carsten> The second line assumes that the org file foo.org
>    Carsten> will also be converted to an html file in the
>    Carsten> publishing process....
>    Carsten>
>    Carsten> Also, I was mistaken earlier:
>    Carsten>
>    Carsten> http:foo.html becomes href="http:foo.html" and not
>    Carsten> href="foo.html"
>    Carsten>
>    Carsten> so this actually cannot be used for a local link.
>    Carsten>
>    Carsten> So what remains from this exchange?  Should we
>    Carsten> accept links that look like absolute file names
>    Carsten> (including ./foo.html) as equivalent to
>    Carsten> file:./foo.html ?
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>    Carsten> On Apr 13, 2008, at 6:25 PM, T. V. Raman wrote:
>    Carsten>
>>>>>
>>> I think writing http:foo.html is a bad idea because typing
>>>>> that string in other contexts is sort of meaningless as
>>>>> a URL.  Until now, everything one types in org-mode sort
>>>>> of has meaning elsewhere. If you want it to look like a
>>>>> url in this case too then I'd suggest file:foo.html --
>>>>> rther than http:file.html.
>>>>>
>>>>> I did not suggest that earlier because it's more chars
>>>>> to type, but that's not a huge concern since emacs does
>>>>> the typing, and we could always augment c-c c-l to
>>>>> enable authoring of relative links with file completion.
>>>>>
>>>>> so file:a/b/c.html would become href="a/b/c.html" in the
>>>>> generated html.
>>>>>
>>>>>>>>>> "Carsten" == Carsten Dominik
>>>>>>>>>> <dominik@science.uva.nl> writes:
>    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
>    Carsten>
>>>>>>> I've not tried http:foo.html -- but I suggest doing
>>>>>>> that to author a relative URL is a bad idea.
>    Carsten>
>    Carsten> Can you explain why you think that this is a bad
>    Carsten> idea?  Educate me!  What is wrong with writing
>    Carsten> http:foo.html ??????
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>>>>>>>
>>>>
>>>> I tried ./foo.html and that didn't work either. One
>>>>>>> compromise would be to get ./foo.html to link to a
>>>>>>> relative url, while foo.html continues to link to a
>>>>>>> local anchor
>>>>>>>
>>>>>>>>>>>> "Carsten" == Carsten Dominik
>>>>>>>>>>>> <dominik@science.uva.nl> writes:
>    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
>    Carsten>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> org-export turns links of the form [[foo][link to
>>>>>>>>> relative url foo]]
>>>>>>>>>
>>>>>>>>> ends up creating links of the form <a
>>>>>>>>> href="#foo">...</a>
>>>>>>>>>
>>>>>>>>> this means that it becomes impossible to write
>>>>>>>>> hyperlinks that are relative URLs.
>    Carsten>
>    Carsten> Hi Raman,
>    Carsten>
>    Carsten> the url goes into the first pair of square brackets,
>    Carsten> the description into the second.  If you put
>    Carsten> something into the first bracket pair that does not
>    Carsten> look lie a url, it will be treated as an internal
>    Carsten> link.
>    Carsten>
>    Carsten> Have you tried to use something like
>    Carsten>
>    Carsten> [[http:foo.html][link to relative url]]
>    Carsten>
>    Carsten> ?
>    Carsten>
>    Carsten> - Carsten
>>>>>>>
>>>>>>> -- 
>>>>>>> Best Regards, --raman
>>>>>>>
>>>>>>>
>>>>>>> Email: raman@users.sf.net WWW:
>>>>>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>>>>>> tv.raman.tv@gmail.com PGP:
>>>>>>> http://emacspeak.sf.net/raman/raman-almaden.asc
>>>>>>> Google: tv+raman IRC: irc://irc.freenode.net/#emacs
>>>>>
>>>>> -- 
>>>>> Best Regards, --raman
>>>>>
>>>>>
>>>>> Email: raman@users.sf.net WWW:
>>>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>>>> tv.raman.tv@gmail.com PGP:
>>>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>>>
>>> -- 
>>> Best Regards, --raman
>>>
>>>
>>> Email: raman@users.sf.net WWW:
>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>> tv.raman.tv@gmail.com PGP:
>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>
> -- 
> Best Regards,
> --raman
>
>
> Email:  raman@users.sf.net
> WWW:    http://emacspeak.sf.net/raman/
> AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
> PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
> Google: tv+raman
> IRC:    irc://irc.freenode.net/#emacs

Re: export to html and relative links:

[Carsten Dominik]

[-- Attachment #1.1: Type: text/plain, Size: 4919 bytes --]


On Apr 13, 2008, at 8:48 PM, T. V. Raman wrote:

> two suggestions:
>
> 1)  accept paths like "./foo.html" as local links.


This is now allowed.

>
> 2) Augment C-c C-l to react to file: by providing filename
> completion relative to the working directory.

This is hard and therefore not yet implemented.  Use `C-u C-c C-l' to  
insert a file link with completion.

- Carsten


>
>
> The emacs binding to the w3m browser does this if you type file: in  
> the minibuffer
> when prompted for a URL.
>
>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl> writes:
>    Carsten> Hmmmm, this is *such* a good idea, that I will
>    Carsten> implement this retroactively into all versions since
>    Carsten> ... 2.0 or so?  Abracadabra! Done.
>    Carsten>
>    Carsten> In fact, it has been working just so for a very long
>    Carsten> time.
>    Carsten>
>    Carsten> file:foo.html -> href="foo.html" file:foo.org ->
>    Carsten> href="foo.html"
>    Carsten>
>    Carsten> The second line assumes that the org file foo.org
>    Carsten> will also be converted to an html file in the
>    Carsten> publishing process....
>    Carsten>
>    Carsten> Also, I was mistaken earlier:
>    Carsten>
>    Carsten> http:foo.html becomes href="http:foo.html" and not
>    Carsten> href="foo.html"
>    Carsten>
>    Carsten> so this actually cannot be used for a local link.
>    Carsten>
>    Carsten> So what remains from this exchange?  Should we
>    Carsten> accept links that look like absolute file names
>    Carsten> (including ./foo.html) as equivalent to
>    Carsten> file:./foo.html ?
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>    Carsten> On Apr 13, 2008, at 6:25 PM, T. V. Raman wrote:
>    Carsten>
>>>
>> I think writing http:foo.html is a bad idea because typing that
>>> string in other contexts is sort of meaningless as a URL.
>>> Until now, everything one types in org-mode sort of has
>>> meaning elsewhere. If you want it to look like a url in
>>> this case too then I'd suggest file:foo.html -- rther than
>>> http:file.html.
>>>
>>> I did not suggest that earlier because it's more chars to
>>> type, but that's not a huge concern since emacs does the
>>> typing, and we could always augment c-c c-l to enable
>>> authoring of relative links with file completion.
>>>
>>> so file:a/b/c.html would become href="a/b/c.html" in the
>>> generated html.
>>>
>>>>>>>> "Carsten" == Carsten Dominik <dominik@science.uva.nl>
>>>>>>>> writes:
>    Carsten> On Apr 13, 2008, at 5:33 PM, T. V. Raman wrote:
>    Carsten>
>>>>> I've not tried http:foo.html -- but I suggest doing that
>>>>> to author a relative URL is a bad idea.
>    Carsten>
>    Carsten> Can you explain why you think that this is a bad
>    Carsten> idea?  Educate me!  What is wrong with writing
>    Carsten> http:foo.html ??????
>    Carsten>
>    Carsten> - Carsten
>    Carsten>
>>>>>
>>>
>>> I tried ./foo.html and that didn't work either. One compromise
>>>>> would be to get ./foo.html to link to a relative url,
>>>>> while foo.html continues to link to a local anchor
>>>>>
>>>>>>>>>> "Carsten" == Carsten Dominik
>>>>>>>>>> <dominik@science.uva.nl> writes:
>    Carsten> On Apr 13, 2008, at 6:49 AM, T. V. Raman wrote:
>    Carsten>
>>>>>>> Hi,
>>>>>>>
>>>>>>> org-export turns links of the form [[foo][link to
>>>>>>> relative url foo]]
>>>>>>>
>>>>>>> ends up creating links of the form <a
>>>>>>> href="#foo">...</a>
>>>>>>>
>>>>>>> this means that it becomes impossible to write
>>>>>>> hyperlinks that are relative URLs.
>    Carsten>
>    Carsten> Hi Raman,
>    Carsten>
>    Carsten> the url goes into the first pair of square brackets,
>    Carsten> the description into the second.  If you put
>    Carsten> something into the first bracket pair that does not
>    Carsten> look lie a url, it will be treated as an internal
>    Carsten> link.
>    Carsten>
>    Carsten> Have you tried to use something like
>    Carsten>
>    Carsten> [[http:foo.html][link to relative url]]
>    Carsten>
>    Carsten> ?
>    Carsten>
>    Carsten> - Carsten
>>>>>
>>>>> -- 
>>>>> Best Regards, --raman
>>>>>
>>>>>
>>>>> Email: raman@users.sf.net WWW:
>>>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>>>> tv.raman.tv@gmail.com PGP:
>>>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>>>
>>> -- 
>>> Best Regards, --raman
>>>
>>>
>>> Email: raman@users.sf.net WWW:
>>> http://emacspeak.sf.net/raman/ AIM: emacspeak GTalk:
>>> tv.raman.tv@gmail.com PGP:
>>> http://emacspeak.sf.net/raman/raman-almaden.asc Google:
>>> tv+raman IRC: irc://irc.freenode.net/#emacs
>
> -- 
> Best Regards,
> --raman
>
>
> Email:  raman@users.sf.net
> WWW:    http://emacspeak.sf.net/raman/
> AIM:    emacspeak       GTalk: tv.raman.tv@gmail.com
> PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
> Google: tv+raman
> IRC:    irc://irc.freenode.net/#emacs


[-- Attachment #1.2: Type: text/html, Size: 15062 bytes --]

[-- Attachment #2: Type: text/plain, Size: 204 bytes --]

_______________________________________________
Emacs-orgmode mailing list
Remember: use `Reply All' to send replies to the list.
Emacs-orgmode@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: export to html and relative links:

[Paul R]

Carsten Dominik <carsten.dominik@gmail.com> writes:


>     2) Augment C-c C-l to react to file: by providing filename
>     completion relative to the working directory.
>
> This is hard and therefore not yet implemented.  Use `C-u C-c C-l' to insert a file link with
> completion.

Drop it if I'm off-topic, but isn't it what (file-relative-name path)
does ?

-- 
      Paul

Re: Re: export to html and relative links:

[Carsten Dominik]

On Jun 24, 2008, at 10:36 AM, Paul R wrote:

> Carsten Dominik <carsten.dominik@gmail.com> writes:
>
>
>>    2) Augment C-c C-l to react to file: by providing filename
>>    completion relative to the working directory.
>>
>> This is hard and therefore not yet implemented.  Use `C-u C-c C-l'  
>> to insert a file link with
>> completion.
>
> Drop it if I'm off-topic, but isn't it what (file-relative-name path)
> does ?

It is hard because I would have to write a special completion function  
which sees the "file:" types and the switches its completion tactics.

- Carsten

Re: Re: export to html and relative links:

[Paul R]

Carsten Dominik <carsten.dominik@gmail.com> writes:

>> Drop it if I'm off-topic, but isn't it what (file-relative-name path)
>> does ?
>
> It is hard because I would have to write a special completion function
> which sees the "file:" types and the switches its completion tactics.

So I was right ! Right to beleive I was off topic ;-)

-- 
      Paul