"\s" does not match non-breaking space characters

[Supa]

In Find Files, when the Find text contains a regular expression with "\s" text with non-breakable space (A0) is NOT matched.

This makes searching for a string in Word documents more challenging.

Same result in Lister. "\s" does NOT match non-breakable space when specified in the Find dialog.

[Dalai]

You can use this:

Code: Select all

\x00A0

to find such characters.

Escape sequences:
[...]
\x{nnnn}
Unicode character with hexadecimal code nnnn. Note that Total Commander uses Unicode for file names, so you need to use this notation for characters not in your local codepage.

BTW, 0A is a line feed. A non-breaking space is 00A0.

[Supa]

You can use this:

Code: Select all

\x00A0

to find such characters.

The point is that "\s", which stands for whitespace, does not match non-breaking spaces. Looks like a bug.

[Supa]

I have found out that "\h" does match non-breaking space. So I will simply use that one.

Still, it would be nice to make "\s" match it as as well.

[ghisler(Author)]

Different regex handlers use different definitions of white space. I have checked the sources of the one I use. It matches:
\s : Space (20), Tab (09), CR (0D), LF (0A), Form Feed (0C)
\h : Space (20), Tab (09), non-breaking space (A0) and if Unicode: 1680, 2000 .. 200A, 202F, 205F, 3000

Microsoft uses a different set for /s:
https://learn.microsoft.com/en-us/dotnet/standard/base-types/character-classes-in-regular-expressions#WhitespaceCharacter
I can't find A0 mentioned there.

GNU Grep seems to use \s as a synonym for [[:space:]]
https://www.gnu.org/software/grep/manual/html_node/Special-Backslash-Expressions.html
Which is defined here:
https://www.gnu.org/software/grep/manual/html_node/Character-Classes-and-Bracket-Expressions.html
‘[:space:]’
Space characters: in the ‘C’ locale, this is tab, newline, vertical tab, form feed, carriage return, and space.

[Dalai]

I have checked the sources of the one I use. It matches:
\s : Space (20), Tab (09), CR (0D), LF (0A), Form Feed (0C)
\h : Space (20), Tab (09), non-breaking space (A0) and if Unicode: 1680, 2000 .. 200A, 202F, 205F, 3000

Is it worth adding this information to the TC help? Currently \h isn't mentioned at all.

[ghisler(Author)]

Yes, I think it would be a good idea to add them.

[white]

I have checked the sources of the one I use. It matches:
\s : Space (20), Tab (09), CR (0D), LF (0A), Form Feed (0C)
\h : Space (20), Tab (09), non-breaking space (A0) and if Unicode: 1680, 2000 .. 200A, 202F, 205F, 3000

Is it worth adding this information to the TC help? Currently \h isn't mentioned at all.

Just to provide additional information: TC's help does mention the website https://regex.sorokin.engineer and it is described there.

[Dalai]

2white
Ah, thanks for pointing that out! I've missed that.

[white]

Ah, thanks for pointing that out! I've missed that.

Many people do. Perhaps it is wise to move the text:

Total Commander uses the free Delphi library TRegExpr by Andrey V. Sorokin, which is now available at https://regex.sorokin.engineer.
Some of the above explanations are from the help file for this library.

from the bottom to the top, just below the first paragraph. (and change "above explanations" to "explanations below")