My NaturalSort unit

[BeniBela]

Or full Unicode support

[typo]

Floating point sort is bonus enough. Pure natural sort does not do it.

Perhaps I cut even this, because it conflicts with other sort types.

[JD]

Floating point sort is bonus enough. Pure natural sort does not do it.

Perhaps I cut even this, because it conflicts with other sort types.

Floating point sort is the main reason why I got interested.

[typo]

Well, the main thing that this floating point sort routine does is, instead of multiplying Result by 10 as it does with integers, it divides the increased amount by 10 after reaching a decimal separator. This way the function returns a double numerical value which can be compared with another.

For integer values:

     First char: '1'
Second char: '2'

So the number is 12. It is 1 * 10 + 2. (FirstChar * 10 + SecondChar).

But for floating values:

     First char: '1'
Second char: decimal separator
   Third char: '2'

So the number is 1.2. It is 1 + 2 / 10. (FirstChar + SecondChar / 10).

And Count is the distance between the current char and the decimal separator.

So the routine "translates" the char to its numerical value.

[typo]

New arrangement of the unit allows thousand separators and floats again.

http://sourceforge.net/projects/lazarusfiles/files/naturalsort.zip/download

[PeterX]

AFAIK, StrCmpLogicalW does the job on Windows and naturalstrcmp does it on Linux.

Looks like user typo was last active in 2015.

Does StrCmpLogicalW do the job ?
Has anybody worked with that, under Windows ?

[rvk]

AFAIK, StrCmpLogicalW does the job on Windows and naturalstrcmp does it on Linux.

Does StrCmpLogicalW do the job ?
Has anybody worked with that, under Windows ?

As far as I can remember... Yes it works.

Here is another topic related to this topic...
https://forum.lazarus.freepascal.org/index.php/topic,28310.0.html

[PeterX]

Thank You !

So I will give it a try and compare the results to ol' naturalsortunit