Comparing strings

[dgimeno]

I'm modifying an old VB6 program and discovered something surprising.

When comparing strings, VB seems to compare always according Unicode values and not Ascii ones.

For example:

  If Chr$(128) > Chr$(146) Then
    MsgBox "128 is greater than 146!"
  End If

When in fact it's comparing &H20AC (128 translated to Unicode) and &H2019 (146 translated to Unicode).

Is there anything to do other than creating a function to compare Ansi?

Note: I don't need this function, I have already written one. My question is just curiousity wheter there's any other way to do it.

[dilettante]

Chr$() says "take this long expression as an ANSI character code, make an ANSI String out of that, then convert it to a Unicode String and return it."

ChrW$() is normally a more rational and faster choice, tough it wants a Unicode character code value.

Maybe ChrB$() is what you really want here? I doubt it though. String values are normally Unicode, not ANSI.

[baka]

u can only use ascii 0-127 when comparing ascii

128-255 are extended ascii code and can be anything depending on your locale.
so those can be all messed up, as u notice 128 is above 146 but in another locale 146 can be above 128.
but another way is to "not" do any chr$ but use safearray and take the "byte" value of the string directly when u compare.
doing so, will not change the string into any locale, it will use the original locale when u created the string (if the string is located in a .dat file example)

so if u plan on using unicode, u will need to work with locale as well.
if u instead just want to use default letters A-Z, a-z those are within 0-127 are always the same in any locale and will always work.

[dgimeno]

Chr$() says "take this long expression as an ANSI character code, make an ANSI String out of that, then convert it to a Unicode String and return it."

ChrW$() is normally a more rational and faster choice, tough it wants a Unicode character code value.

Maybe ChrB$() is what you really want here? I doubt it though. String values are normally Unicode, not ANSI.

ChrB$ is what I need and didn't know it exists.

It's an old app, as I have said, and it isn't working with Unicode at all.

Thanks!

[dgimeno]

u can only use ascii 0-127 when comparing ascii

128-255 are extended ascii code and can be anything depending on your locale.
so those can be all messed up, as u notice 128 is above 146 but in another locale 146 can be above 128.
but another way is to "not" do any chr$ but use safearray and take the "byte" value of the string directly when u compare.
doing so, will not change the string into any locale, it will use the original locale when u created the string (if the string is located in a .dat file example)

so if u plan on using unicode, u will need to work with locale as well.
if u instead just want to use default letters A-Z, a-z those are within 0-127 are always the same in any locale and will always work.

Yes, I was taking the byte value in my function to binary safe compare. Anyway, changing Chr by ChrB is a cleaner solution. Thank you.