read persian text file with using inetc plugin in nsis - nsis

Question

I have many text file with persian text: "سلام"
now I want to read the text file and write to the variable directly
i use from the inetc plugin sample this:
Unicode true
Section
;test_Unicode_BE_BOM.txt:
inetc::get /NOCANCEL /TOSTACK /SILENT "http://update.nodmarket.com/Persian_Text_Files/test_Unicode_BE_BOM.txt" /END
Pop $R0
Pop $R1
MessageBox MB_OK "$R1"
;test_Utf8_BOM.txt
inetc::get /NOCANCEL /TOSTACK /SILENT "http://update.nodmarket.com/Persian_Text_Files/test_Utf8_BOM.txt" /END
Pop $R0
Pop $R1
MessageBox MB_OK "$R1"
;test_Unicode_LE_BOM.txt
inetc::get /NOCANCEL /TOSTACK /SILENT "http://update.nodmarket.com/Persian_Text_Files/test_Unicode_LE_BOM.txt" /END
Pop $R0
Pop $R1
MessageBox MB_OK "$R1"
SectionEnd
but after show i see this text in messagebox:
inetc plugin and source page:
what is the problem?
why i see the broken persian text after read from inetc plugin?
when i use the code
Unicode true
Section
MessageBox MB_OK|MB_RIGHT|MB_RTLREADING "سلام"
SectionEnd
i see "سلام" in messagebox successful
but i think inetc plugin Ruins Persian words!

Answer 1

INetC does not really support Unicode conversion on the fly.
If you need full control you have to download the file to a temporary directory and then read it with NSIS file instructions.
InitPluginsDir
inetc::get /NOCANCEL /SILENT "http://update.nodmarket.com/Persian_Text_Files/test_Unicode_LE_BOM.txt" "$PluginsDir\test.txt" /END
Pop $0
${If} $0 == "OK"
FileOpen $1 "$PluginsDir\test.txt" r
FileReadUTF16LE $1 $2
FileClose $1
MessageBox mb_ok $2
${EndIf}
I uploaded a new version of INetC to the wiki for you with limited UTF-8 and UTF-16LE support (BOM required).
inetc::get /NOCANCEL /TOSTACKCONV /SILENT "http://update.nodmarket.com/Persian_Text_Files/test_Utf8_BOM.txt" /END ; v1.0.5.3+ required for /TOSTACKCONV to support BOM detection.
Pop $R0
Pop $R1
MessageBox MB_OK "$R0:$R1"
Putting Unicode text in a .nsi file works because the NSIS compiler supports Unicode but text returned by a plug-in has its text handling inside the plug-in and it might not support Unicode.