what is the difference between _("<string>") and _$("<string>")

[AnnieRuru 957]

What is the difference between _("<string>") and _$("<string>") meant for HULD

I found a few clues in script.c

		else if( strcmp(buildin->name, "mes") == 0 ) script->buildin_mes_offset = script->buildin_count;
		else if( strcmp(buildin->name, "mesf") == 0 ) script->buildin_mesf_offset = script->buildin_count;
		else if( strcmp(buildin->name, "select") == 0 ) script->buildin_select_offset = script->buildin_count;
		else if( strcmp(buildin->name, "_") == 0 ) script->buildin_lang_macro_offset = script->buildin_count;
		else if( strcmp(buildin->name, "_$") == 0 ) script->buildin_lang_macro_fmtstring_offset = script->buildin_count;

so its script->buildin_lang_macro_offset and script->buildin_lang_macro_fmtstring_offset

 

next, in generate-translations.c

	if (!duplicate) {
		if (script->syntax.last_func == script->buildin_mes_offset
		 || script->syntax.last_func == script->buildin_select_offset
		 || script->syntax.lang_macro_active
		 ) {
			is_translatable_string = true;
		} else if (script->syntax.last_func == script->buildin_mesf_offset
				|| script->syntax.lang_macro_fmtstring_active
				) {
			is_translatable_fmtstring = true;
		}
	}

now this make sense, _("<string>") is somehow group together with mes and select
_$("<string>") is somehow group together with mesf

 

so this is my guess ...

_("<string>") is for those without %d %s ...
_$("<string>") is for those with %d %s ... like sprintf ... right ?

 

so ... my next question is ... there are 3 types

<none>
c format
no c format

Spoiler

-	script	asdf	FAKE_NPC,{
	.@a$ = _("test");
	.@b$ = _$("nothing");
	end;
OnInit:
	bindatcmd "mobcount", strnpcinfo(NPC_NAME)+"::Onaaa";
	end;
Onaaa:
	if ( !.@atcmd_numparameters ) {
		dispbottom "input a mob ID";
		end;
	}
	.@mobid = atoi( .@atcmd_parameters$ );
	if ( getmonsterinfo( .@mobid, MOB_LV ) == -1 ) {
		dispbottom "invalid monster ID";
		end;
	}
	.@size = getunits( BL_MOB, .@bl, false, strcharinfo(PC_MAP) );
	for ( .@i = 0; .@i < .@size; ++.@i )
		if ( getunitdata( .@bl[.@i], UDT_CLASS ) == .@mobid )
			++.@c;
	dispbottom sprintf(_("There are %1$dx %2$s in %3$s."), .@c, getmonsterinfo( .@mobid, MOB_NAME ), strcharinfo(PC_MAP) );
	end;
}

generate the file as

#: npc/zzz.txt
# .@a$ = _("test");
msgctxt "asdf"
msgid "test"
msgstr ""

#: npc/zzz.txt
# .@b$ = _$("nothing");
#, c-format
msgctxt "asdf"
msgid "nothing"
msgstr ""

#: npc/zzz.txt
# dispbottom sprintf(_("There are %1$dx %2$s in %3$s."), .@c, getmonsterinfo( .@mobid, MOB_NAME ), strcharinfo(PC_MAP) );
#, no-c-format
msgctxt "asdf"
msgid "There are %1$dx %2$s in %3$s."
msgstr ""

what are those actually means then ?

if translate by hand, it probably doesn't mean anything ...

line 187 in generate-translations.c

is_translatable_fmtstring ? "#, c-format\n" : (has_percent_sign ? "#, no-c-format\n" : ""),

 

@Haru @Dastgir

 

[Dastgir 1246]

This can give some overview: https://github.com/HerculesWS/Hercules/issues/2233

 

_$ is used if there's some string formatting (%s, %d or something like that), mesf is basically 3 commands in one, i.e mes+sprintf+_$

 

_() is used to enclose strings which aren't in commands that are translated(only few commands are detected like mes/select), others needs to be enclosed in _(), also raw string needs to be enclosed.

 

c-format is comment in pot file which says formatting is present.

[bWolfie 138]

this is beyond my thinking

why sprintf(_("Text %s", .@var));

not sprintf("Test %s", .@var);

[AnnieRuru 957]

that means _$() or _() is equally detect by HULD, the only difference is it generate a comment in the pot file
means nothing big differences
 ... I actually thought _() unable to parse %s or something ... seems I was wrong

 

@Myriad, soon we will roll out a new version of HULD, it already discuss on the staff level
old design -> http://herc.ws/board/topic/8687-hercules-ultimate-localization-design/
once the new design roll out, I will force recommend the community to adapt this new scripting style

and for your question, because sprintf("Test %s", .@var); isn't detect by HULD
and its sprintf(_("Text %s"), .@var); , only enclose the string

 

by the way, on a side note, I seriously think we should change all the ^000000 into F_MesColor(C_BLACK)

some machine translation can break
^000000 into ^ 000000 <-- can be solve by F_MesColor
%s into % s <-- nothing can be done with this, other than manually fix it

Edited February 13, 2019 by AnnieRuru