Processing Semicolon

[tiago]

In this case it is known, SammaySarkar.

I have several of those files and online locations to process thus the need of mastering the problem.
Have you tried to use readfile on the sample plus the posted test script?

[bdeshi]

You still haven't said what you want to do.

[tiago]

Eliminate the semicolon which is not being recognized by the described routine.

[bdeshi]

Aha! that semi-colon is actually the greek question mark character.

Try this script:

Code: Select all

regexreplace(<your text>,"(\(..\))(\u037e|;)", "$1");

The regex matches any TWO characters within braces. This pattern handles both a normal ; and that special character.

[tiago]

A-hah!!!


But... why ASC reported it as being 59? Why once pasted into the forum and copied from it it behaves as a regular semicolon? Is there a general formula which could prevent ANY breaking of regexmatches on any other unknown character like this one? Could you please elaborate a bit on this?

Thanks-a-bunch!

[highend]

Because it's not in the ascii table?

There is no general formula, regexes with unicode chars are completely different than regexes inside the ascii character set

[tiago]

And how can a user find the difference? I mean... that special character is matched when a simple semicolon is searched by CTRL+F on Notepad, but regexmatches didn't recognized it as a twin.

@SammaySarkar: how did you find the difference?

[bdeshi]

Well, I tried matching a ";" but that didn't succeed, so the only logical conclusion was that it wasn't in fact a semi-colon, leading me to look at the character code. You can find this code with a hex editor. I just used the HexSel plugin of AkelPad.

[tiago]

Similar question, I guess.

XYplorer sc NEW fails to create "d:\x2\WwwComBr\wwwcombr25_01\linux\apache_1.3.24.tar.gz" mirror (zero size) file.
Pasting source path into pspad or notepad++ reveals nothing special.
=> removing any single letter from 'WwwComBr\wwwcombr25_01\linux' out of the addy bar part allows the creation.

Try it:

Code: Select all

new ("d:\x2\WwwComBr\wwwcombr25_01\linux\apache_1.3.24.tar.gz", file); beep;

although I think that copy from browser into XY addy bar will remove the potential invisible char.

XYplorer hex view:

Code: Select all

00000000: 64 00 3A 00 5C 00 78 00 32 00 5C 00 57 00 77 00   d.:.\.x.2.\.W.w.
00000010: 77 00 43 00 6F 00 6D 00 42 00 72 00 5C 00 77 00   w.C.o.m.B.r.\.w.
00000020: 77 00 77 00 63 00 6F 00 6D 00 62 00 72 00 32 00   w.w.c.o.m.b.r.2.
00000030: 35 00 5F 00 30 00 31 00 5C 00 6C 00 69 00 6E 00   5._.0.1.\.l.i.n.
00000040: 75 00 78 00 5C 00 61 00 70 00 61 00 63 00 68 00   u.x.\.a.p.a.c.h.
00000050: 65 00 5F 00 31 00 2E 00 33 00 2E 00 32 00 34 00   e._.1...3...2.4.
00000060: 2E 00 74 00 61 00 72 00 2E 00 67 00 7A 00         ..t.a.r...g.z.  

[highend]

Code: Select all

new("d:\x2\WwwComBr\wwwcombr25_01\linux\apache_1.3.24.tar.gz", "file");

Works fine...

[tiago]

although I think that copy from browser into XY addy bar will remove the potential invisible char.

Here too.
I'm trying to strip down the source file and code so I can share a more reliable sample.

Thank you for trying.

[highend]

Just post the part that gets / build / whatever the url and transforms it into the real path?

[tiago]

I see the problem now.

Say you have a file
Linux (yes, no extension)
new("d:\x2\WwwComBr\wwwcombr25_01\Linux", "file");

then you try to create

new("d:\x2\WwwComBr\wwwcombr25_01\linux\apache_1.3.24.tar.gz", "file");

the existing file 'Linux' will prevent the creation of a folder 'linux' at the same directory.

Anything I could do about this apart from testing existence of each component of each file that'll be created...?

[highend]

Nope

[tiago]

Thanks!

//%$#@#$%%$##$%*&$!!@#$#$$*&%#@$$#@!@#$%% ¨¨