Regexp - matching text not containing certain count of certain characters

[Jiri Fogl]

I'm trying to build a regular expression (POSIX extended). One part of
it has to match any text not containing a slash except standalone two
dots (..), but I can't find the way to say something like "match any
count of any characters except string ..".

For example, it should match
...anything..
aaa..
....

but must not match
anything/containing/a/slash
...

Thanks for any advice.

Jiri Fogl

[John Dunlop]

Jiri Fogl:

I'm trying to build a regular expression (POSIX extended). One part of
it has to match any text not containing a slash except standalone two
dots (..), but I can't find the way to say something like "match any
count of any characters except string ..".

For example, it should match
..anything..
aaa..
...

but must not match
anything/containing/a/slash
..

(UNTESTED)

!ereg('^\.\.$|/',$subject)

--
Jock

[Jiri Fogl]

Thanks for advice, but this is not exactly what I need. I'll try to
explain it better:

Regular expression I need has to search through path or URL and find
every reference to a parent directory, including this parent directory.

For example if I have URL like this:
http://www.example.com/mydir/../myotherdir
RE is supposed to match /mydir/../ so I can replace it by a single slash
and get "definite" URL.

I have built a regexp to achieve this:
ereg_replace('/[^/]+/\.\./', '/', $url);

Problem of this RE appears in case of URL like this:
http://www.example.com/level1/level2..../../../target

This URL is pointing to http://www.example.com/target, but ereg_replace
above will give me
http://www.example.com/level1/level2/level3/../target. After looped
ereg_replace I get URL http://www.example.com/level1/level2/target which
is wrong.

I need to modify this part of RE: /[^/]+/
Now it says "match any string if it doesn't contain a slash" and I need
it to say "match any string if it doesn't contain a slash and it is not
..." and I need it to be a part of a RE above.

John Dunlop wrote:

Jiri Fogl:

I'm trying to build a regular expression (POSIX extended). One part of
it has to match any text not containing a slash except standalone two
dots (..), but I can't find the way to say something like "match any
count of any characters except string ..".

For example, it should match
..anything..
aaa..
...

but must not match
anything/containing/a/slash
..

(UNTESTED)

!ereg('^\.\.$|/',$subject)

[John Dunlop]

Jiri Fogl:

Regular expression I need has to search through path or URL and find
every reference to a parent directory, including this parent directory.

For example if I have URL like this:
http://www.example.com/mydir/../myotherdir
RE is supposed to match /mydir/../ so I can replace it by a single slash
and get "definite" URL.

Sorry, I don't know how you'd do that with a regular expression, even
with a PCRE.

Does it need to be done with a single RE? If not, RFC3986 gives an
example algorithm for removing dot segments (sec. 5.2.4), and the whole
algorithm of section 5.2 is definitive, so I would refer to it anyway.

http://www.ietf.org/rfc/rfc3986.txt

--
Jock

[Jiri Fogl]

John Dunlop wrote:

Jiri Fogl:

Regular expression I need has to search through path or URL and find
every reference to a parent directory, including this parent directory.

For example if I have URL like this:
http://www.example.com/mydir/../myotherdir
RE is supposed to match /mydir/../ so I can replace it by a single slash
and get "definite" URL.

Sorry, I don't know how you'd do that with a regular expression, even
with a PCRE.

Does it need to be done with a single RE? If not, RFC3986 gives an
example algorithm for removing dot segments (sec. 5.2.4), and the whole
algorithm of section 5.2 is definitive, so I would refer to it anyway.

http://www.ietf.org/rfc/rfc3986.txt

Thank you for this link very much. I didn't know about this RFC, this is
definitely what I've been searching for (even I didn't know that :-) ).

Thanks again.