WxScintilla BlitzMax lexer

BlitzMax Forums/Brucey's Modules/WxScintilla BlitzMax lexer

Derron

(Posted 2013) [#1]

I just saw you added the proposal of UNZ for REM-Folding.

What about having other parts foldable too?


	if (
		(!strcmp(token, "function") && !abstract) ||
		!strcmp(token, "type")     ||
		(!strcmp(token, "method") && !abstract)   ||
		!strcmp(token, "repeat")   ||
		!strcmp(token, "while")    ||
		!strcmp(token, "for")      ||
		!strcmp(token, "select")      ||
		(!strcmp(token, "if") && !singleLineIf) ||
		!strcmp(token, "rem")      ||
		!strcmp(token, "try")
	    ) {
		level |= SC_FOLDLEVELHEADERFLAG;
		return 1;
	}
	if (
		!strcmp(token, "endfunction") || !strcmp(token, "end function") ||
		!strcmp(token, "endtype") || !strcmp(token, "end type") ||
		!strcmp(token, "endmethod") || !strcmp(token, "end method") ||
		!strcmp(token, "until")       ||
		!strcmp(token, "forever")     ||
		!strcmp(token, "wend")        ||
		!strcmp(token, "endwhile")    ||
		!strcmp(token, "next")        ||
		!strcmp(token, "endselect") || !strcmp(token, "end select") ||
		!strcmp(token, "endif") || !strcmp(token, "end if") ||
		!strcmp(token, "endrem") || !strcmp(token, "end rem") ||
		!strcmp(token, "endtry") || !strcmp(token, "end try")
	    ) {
		return -1;
	}

So in short: all "loops" should be collapseable.

I know this code (from the old blitz-lexer-file extended by some keywords) has a small flaw: the "singleLineIf"-check is not done properly as it ignores ".."-concat lines (if bla .. \n then)

Another thing for potential extension: more keywords:

Just an example how i broke up my keywords:

syntaxkeywords= method function endmethod endfunction type endtype field global final framework local onend module moduleinfo
modulekeywords= ... addhook ... bindsocket ... changedir ... drawoval ...
logicalkeywords= mod new next not object global field and or if then else assert endif elseif eachin for next step continue select case default endselect exit false forever true extends try wend while extern end const goto import incbin include import private public strict super superstrict to self
typekeywords= byte double float int long ptr short string var varptr null
customtypekeywords= floattypeid inttypeid objecttypeid shorttypeid ttypeid taudiosample taudiosampleloader tbank ... tthread tthreaddata

in the function Colourise I currently have:

				int kstates[11] = {
					SCE_B_KEYWORD,
					SCE_B_SYNTAXKEYWORD,
					SCE_B_MODULEKEYWORD,
					SCE_B_LOGICALKEYWORD,
					SCE_B_CUSTOMKEYWORD,
					SCE_B_TYPEKEYWORD,
					SCE_B_CUSTOMTYPEKEYWORD,
					SCE_B_KEYWORD2,
					SCE_B_KEYWORD3,
					SCE_B_KEYWORD4,
					SCE_B_KEYWORD5,
				};

Which should be enough to make it really configurable.

In scintilla/include/SciLexer.h some lines will have to get added:

#define SCE_B_COMMENT_MULTILINE 19
#define SCE_B_SYNTAXKEYWORD 20
#define SCE_B_MODULEKEYWORD 21
#define SCE_B_LOGICALKEYWORD 22
#define SCE_B_CUSTOMKEYWORD 23
#define SCE_B_TYPEKEYWORD 24
#define SCE_B_CUSTOMTYPEKEYWORD 25
#define SCE_B_KEYWORD5 26

in scintilla/include/Scintilla.lFace add:

val SCE_B_COMMENT_MULTILINE=19
val SCE_B_SYNTAXKEYWORD=20
val SCE_B_MODULEKEYWORD=21
val SCE_B_LOGICALKEYWORD=22
val SCE_B_CUSTOMKEYWORD=23
val SCE_B_TYPEKEYWORD=24
val SCE_B_CUSTOMTYPEKEYWORD=25
val SCE_B_KEYWORD5=26

Above is just a suggestion. Like already mentioned in the response to UNZ's editor this MIGHT add clutter as it adds more options. But this is just the case if one uses this entries. More options is always better (think this is the premise of Brucey).

bye
Ron