From: Oleksandr G. <gav...@gm...> - 2012-12-14 17:16:32
|
On 2012-12-14, Amir Szekely wrote: > We want to remove the UNICODE=no option, so INSTALL probably needs to > be updated. The plan is to have makensis.exe always unicode capable, but > able to choose between ANSI and Unicode installers. That does mean "choose between ANSI and Unicode installers"? Does this mean only that .nsi input file to makensis interpreted as UNICODE or ANSI so this is effected: MessageBox MB_OK "Привет/გამარჯობა/नमस्ते" like commands? While current .nsh/.nlf files in UTF-8 this: Docs/src/usefulinfos.but:440: There are two macros that can help you write scripts that work for both Unicode and ANSI installers. To figure out if the script is being compiled to generate a Unicode installer, use !ifdef check for $\{NSIS_UNICODE\}. To see what the size of the default character is, use $\{NSIS_CHAR_SIZE\}. It will be 1 for ANSI and 2 for Unicode installers. say that internally all string converted to UCS-16-le (or UTF-16-le if full Unicode support provided). It is interesting to have two different coding system - one for source files another for variable values... I work of translation of Ukrainian.nsh (changes still has not been commited) and introduce use of French quotes - «». While cp1251 copdepage have appropriated codepoint I want to check that conversion of .nsh performed successfully (if they are performed at all for ANSI build...). -- Best regards! |