Mailing lists / fpc-pascal / 2019-05
7 messages · 5 people · –
Reply by email Subscribe to this list Your subscription
Read-only archive. Use j and k to move between messages.
I am making small unit https://github.com/alexey-t/aencoding
so I need to know:
a) will it work on FPC 3.0.4? do I need to use {$ifdef}?
b) my trick to set CP_UTF8 at the end of nnnnFromTo function is ok?
c) what encoding numbers (e.g. 936, 950) exist in FPC 3.3 but didn't exist in FPC 3.0.4?
It will probably (not) work the same under FPC 3.0.4 and trunk. No guarantees though.
b) my trick to set CP_UTF8 at the end of nnnnFromTo function is ok?
It's not ok. It will at leas break when trying to write such a string, or when assigning it to a unicodestring or widestring, or when assigning it to a typed ansistring with a different code page than UTF-8 or DefaultSystemCodePage.
c) what encoding numbers (e.g. 936, 950) exist in FPC 3.3 but didn't exist in FPC 3.0.4?
FPC generally uses system library functions for the conversion, so it does not depend on the FPC version. The only exception is if you use the fpwidestring unit. In that case, the supported code pages depend on which extra units from packages/rtl-unicode you include. I don't know which code page(s) it supports by default.
Jonas
FPC generally uses system library functions for the conversion, so it
does not depend on the FPC version.
But FPC has such files. Are they used in SetCodePage code?
ls cp*
cp895.pas cp932.pas cp936.pas cp949.pas cp950.pas cpbuildu.pp
and
ls
8859-10.txt 8859-2.txt 8859-9.txt cp1256.txt cp850.txt cp862.txt cp895.txt
8859-11.txt 8859-3.txt cp1250.txt cp1257.txt cp852.txt cp863.txt cp932.txt
8859-13.txt 8859-4.txt cp1251.txt cp1258.txt cp855.txt cp864.txt cp936.txt
8859-14.txt 8859-5.txt cp1252.txt cp3021.txt cp856.txt cp865.txt cp949.txt
8859-15.txt 8859-6.txt cp1253.txt cp437.txt cp857.txt cp866.txt cp950.txt
8859-16.txt 8859-7.txt cp1254.txt cp737.txt cp860.txt cp869.txt koi8-r.txt
8859-1.txt 8859-8.txt cp1255.txt cp775.txt cp861.txt cp874.txt koi8-u.txt
These files are not used, they predate the fpWidestring unit, they are for the charset unit, which can be used to manually convert codepages.
Michael.
Correction: if you use the fpwidestring unit, these charset units are used.
Michael.
Well, I believe that it's still a bit imprecise. ;-) My understanding is that a programmer deciding to use fpwidestring may decide on the set of character mappings to be supported out of the box in his program. That may be done in two ways. One is by including a unit for a particular character set (e.g. fpcsrc/rtl/charmaps/cp1252.pas or fpcsrc/packages/rtl-unicode/src/inc/cp932.pas) or a set of character sets (fpcsrc/rtl/charmaps/cpall.pas and/or
fpcsrc/packages/rtl-unicode/src/inc/cpbuildu.pas) in the uses clause of one of the project source files. The other is by loading a binary file (install/unicode/charset/*.bcm) dynamically at runtime (LoadBinaryUnicodeMapping in fpcsrc/rtl/inc/charset.pp). Both the units and the binary files are generated from the *.txt files using tools included with FPC.
Tomas
Yes. See
https://www.freepascal.org/docs-html/current/rtl/fpwidestring/index.html
Michael.