Xbase++ forum

[DenGhostYY]

As far as I know, strings in Xbase++ do not have a special setting that is responsible for the case sensitivity of strings when comparing. You can select a collation table for characters using SET COLLATION. However, in Alaska there is no corresponding #define for the Russian language, so we use LocaleConfigure to recreate the Russian character table in 866 encoding.

The X# documentation for the SetCollation says that if the #Clipper mode is set, then the comparison of Latin characters is carried out in this order: A < B < C < ... < Z < a < b < c < ... < z

So after seeing this function in the documentation, I decided that the string comparison problem could be solved in a simpler way. Can I rely on these functions?

Den,

Is Alaska always ignoring the case when doing string comparisons, or is this behavior based on a setting?

[DenGhostYY]

Doesn't the Russian clipper/vo collation table correspond to the code page 866?
In X# the function Asc() returns the correct result for Latin 'b', 'C'. Imho, this means something is wrong with the comparison settings.

Den,
The sort order is controlled by the russian collation. I would have to check what the contents of that collation is.
We extracted that sort order from the clipper / vo collation table

Robert

[robert]

Guys,
The problem is different from what you would expect:
there is a SetCollation() function in both XSharp.Core and XSharp.XPP
I am not sure why we did this, but SetCollation() inside XSharp.XPP expects a numeric parameter from Collat.ch
This activates one of the Xbase++ collation tables.

In this case, you would need to call the function inside XSharp.Core like this:

Code: Select all

XSharp.Core.Functions.SetCollation("CLIPPER")

This would also have worked

SetCollation(COLLAT_AMERICAN)


I will change the version inside XPP to forward string (and symbol) parameters to the function inside XSharp.Core

Robert

[DenGhostYY]

Yes, clarifying the namespace helped. I decided to try with Cyrillic characters, but for some reason the function Asc() returns codes based on code page 1251, although the source was in 866 encoding and there was a compilation flag -codepage:866

Code: Select all

procedure Main()
#ifdef __XSHARP__
    SetInternational(#CLIPPER)
    XSharp.Core.Functions.SetCollation("CLIPPER")
    SetNatDll("RUSSIAN.DLL")
#endif
    ? asc("Г"), asc("б"), "Г" < "б"
    wait
    return

Alaska outputs:
67 98 Y
131 161 Y

X# outputs:
67 98 .T.
195 225 .T.

[robert]

Den,
The Asc() function translates the characters from unicode to ansi at runtime.
What is the value of
RuntimeState.DosCodePage
RuntimeState.WinCodePage

Robert

[DenGhostYY]

866, 1251

[DenGhostYY]

Hello. Is there an analogue of the function SymbolInfo in XSharp?

[robert]

The function SymbolInfo() does not exist yet.
But we could create / emulate it.
However, the result may be a HUGE array, depending on the information requested.
Especially when asking for SYMBOL_BUILTIN_CLASSFUNC
Should that include the classes in the X# runtime or also all the classes inside the .Net runtime?
And should SYMBOL_BUILTIN_FUNCTION only return functions declared in the X# runtime or also static methods in the .Net runtime?
In other words:
what exactly are you looking for?

Robert

[DenGhostYY]

SYMBOL_FUNCTION

[DenGhostYY]

SYMBOL_FUNCTION

The return value is a one-dimensional array holding the symbolic names of FUNCTIONs and PROCEDUREs as character strings.