Re: [xsl] Escaping special characters for *nix file path

[Liam R E Quin <liam@xxxxxx>]

On Sun, 2012-07-29 at 06:30 +0200, Lighton Phiri wrote:

> Yes they are allowed, but special characters need to be escaped for
> one to access a file path.
[..]
> phiri@PHRLIG001:~$ touch ../data/Sites/3218AD Eland\'s
> Bay/Bobbejaansberg/BB15/testFile.txt
> touch: cannot touch `Bay/Bobbejaansberg/BB15/testFile.txt': No such
> file or directory

This is because the shell splits arguments at spaces, so you actually
gave the touch command two filenames. If you used quotes
$ touch "../data/Sites/3218AD Eland's
Bay/Bobbejaansberg/BB15/testFile.txt"

then it would work fine.

It is not necessary to escape characters from the shell inside an XSLT
stylsheet, because the shell isn't involved.

[...]

> >replace($filename, "[ '\\`&;]", "\\&")
> >will probably do what you want.
> >Or use \s instead of space if there might be newlines.
> >You might also need to replace " with \"
> 
> I am getting an error when I try what you suggested.
> 
> <xsl:value-of select="{replace($filename, "[ '\\`&;]", "\\&")}" />

You don't want the curly braces there - value-of is expecting an
expression, not a string.

[...]

> My file paths have special characters in them and those characters
> include 'square brackets', 'parentheses', 'ampersands', etc. all of
> which are interpreted differently by the regular expression given by
> second argument of replace function.

That doesn't matter.

Actually what I tend to do myself is something like
   replace($filename, "[^a-zA-Z0-9]+", "-")
to turn any sequence of characters other than letters or digits into a
hyphen. If it's a path rather than a filename, include / after the 9
there. But if the data is untrusted you should not normally allow /
inside it.

Finally, as Mike Kay implied in a separate message, you probably want
\$0 as the replacement, not \& (I use too many regular expression
libraries, sorry)

Liam

-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/
Pictures from old books: http://fromoldbooks.org/
Ankh: irc.sorcery.net irc.gnome.org freenode/#xml