Paolo Mantovani wrote:
Hello Stephan,
Alle 09:34, venerdì 22 luglio 2005, Stephan Bergmann ha scritto:
[...]
I must admit that I don't understand this part of your explanation
I'm sure that I'm missing something, may you detail a bit more this part?
The syntax for vnd.sun.star.expand URLs is
"vnd.sun.star.expand:" + <escaped macrofied URL>
So if the macrofied URL is
"$ORIGIN/foo%20bar"
its escaped representation is
"$ORIGIN/foo%2520bar"
(because "%" is a character that must be escaped as "%25" when inserted
into a URL). If you do not de-escape this before de-macrofication, the
result will be some URL like
"file:///the/origin/foo%2520bar"
which would correspond to a filepath
"/the/origin/foo%20bar"
instead of the intended
"/the/origin/foo bar"
Thank you very much for the clearful explanation!
IIRC, there currently is no functionality available from OOo Basic to do
the de-escaping. There is C++ functionality for that in rtl/uri.hxx,
and on CWS sb36 (targeted for OOo 3.0) I just added the UNO
com.sun.star.uri.UriSchemeParser_vndDOTsunDOTstarDOTexpand that includes
functionality to do the complete de-escape--and--de-macrofy step.
Speaking of OOBasic, it's true, there is no a direct way to de-escape URLs,
but In this specific context (macrofied url's) this workaround seems working
quite well:
sEscapedTwiceUrl = "$ORIGIN/foo%2520bar"
sEscapedSimpleUrl = convertFromUrl(sEscapedTwiceUrl)
print sEscapedSimpleUrl
What do you think about?
According to Andreas Bregas, convertFromUrl is intended to convert from
(file) URLs to platform-specific filepaths. One step it does is to
de-escape the URL, so it might work ok here by chance---but it could
also do harm to the URL.
As a further possibility, I'm looking in the sources of cUrl library.
http://curl.haxx.se/
The file lib/escape.c contains two functions that encodes/decodes Url's.
It shouldn't be very complicated to "translate" them in OOBasic, therefore, if
the first workaround wouldn't be accettable for any reasons, I could try this
second way
I did not look at the cUrl stuff. In theory, what you have to do here
is to de-escape sequences of "%xx"-sequences into byte sequences (each
"xx" is a hex digit pair that represents a byte), treat those byte
sequences as UTF-8 strings, and convert those UTF-8 strings into UTF-16
strings. In practice, the original input (the macrofied URL
"$ORIGIN/foo%20bar") should have been a valid macrofied URL, which
implies that it should only contain ASCII characters, which in turn
implies that the escaped macrofied URL should only contain escape
sequences %00--%7F. That simplifies matters: You can individually
de-escape such an escape sequence in the range %00--%7F into a single
Basic (UTF-16) character by building a Basic integer value from the two
hex digits and using $chr (or whatever that function is called in OOo
Basic) to turn it into a Basic character value.
-Stephan
ciao
Paolo
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
