RegExp - question about [:print:] and Chr(160)

[mLipok]

HelpFile for RegExp says:

Quote

[:print:]

ASCII printing characters, including space (same as Chr(32) ... Chr(126)).

Why this usage is not catching NBS I mean Chr(160) ?

 

[jchd]

Here, ASCII means ASCII, i.e. [0x00..0x7F] and not "Extended ASCII".  What the StrngRegexp doc doesn't say is that in UCP mode [:print:] cover the whole range of printable codepoints of UCS2 charset (= Unicode BMP).

Local $s
For $i = 0x00 To 0x3FF      ; put some low limit
    $s &= ChrW($i)
Next
$s = StringRegExpReplace($s, "(*UCP)[[:^print:]]", "")
Local $a = StringToASCIIArray($s)
For $c In $a
    $cw(Hex($c, 4) & @TAB & ">" & ChrW($c) & "<")
Next

Edit1: Other POSIX classes change when UCP is enabled, something which is not explicitely mentionned in our doc:
[:space:]
[:graph:]
[:punct:]

Edit2: It's noticeable that [:cntrl:] even in UCP mode doesn't cover the range of C1 control characters [0x80..0x9F] ∪ [0xAD].  That may have changed with more recent versions of PCRE1 and all PCRE2 versions.

Edit3: ticket #3725 posted in Trac

Edited July 6, 2019 by jchd
Expand to other classes

[jchd]

Just playing with POSIX classes a bit, I discovered that:

55045 printable codepoints
55027 graphic codepoints

1 graphic non-printable codepoints
0x0000

18 printable non-graphic codepoints (various spacing codepoints)
0x0020
0x00A0
0x1680
0x180E
0x2000
0x2001
0x2002
0x2003
0x2004
0x2005
0x2006
0x2007
0x2008
0x2009
0x200A
0x202F
0x205F
0x3000

Now back to work!

[jchd]

That's correct and it's the right site to use to test/debug regexp in AutoIt sauce.