diff options
Diffstat (limited to 'doc/expert/det_lexi')
| -rw-r--r-- | doc/expert/det_lexi | 253 |
1 files changed, 253 insertions, 0 deletions
diff --git a/doc/expert/det_lexi b/doc/expert/det_lexi new file mode 100644 index 0000000..e0ad9f1 --- /dev/null +++ b/doc/expert/det_lexi @@ -0,0 +1,253 @@ +@Section + @Tag { lexical } + @Title { Lexical structure (words, spaces, symbols) and macros } +@Begin +@PP +The input to Lout consists of a sequence of @I {textual units}, +textual.unit @Index {Textual unit } +which may be +either {@I{white spaces}}, +@I identifiers, +@I delimiters, +or +@I {literal words}. Each +is a sequence of @I characters chosen from: +letter @Index { Letter character } +other @Index { Other character } +quote @Index { Quote character } +escape @Index { Escape character } +comment.char @Index { Comment character } +underscore.char @Index { Underscore character } +@ID @Tab + vmargin { 0.5vx } + @Fmta { @Col A ! @Col B } +{ +@Rowa A { letter } B { @Code "@ab-zAB-Z_" } +@Rowa A { white space } B { @I { space formfeed tab newline } } +@Rowa A { quote } B { @Code "\"" } +@Rowa A { escape } B { @Code "\\" } +@Rowa A { comment } B { @Code "#" } +@Rowa A { other } B { @Code "!$%&'()*+,-./0123456789:;<=>?[]^`{|}~" } +} +Notice that @Code "@" and @Code "_" are classed as letters. Basser +Lout accepts the accented letters of the ISO-LATIN-1 character set +(depending on how it is installed), and these are also classed as +letters. The ten digits are classed as `other' characters, and in +fact the `other' class contains all 8-bit characters (except octal 0) +not assigned to previous classes. +@PP +A @I {white space} is a sequence of one or more white space characters. +white.space @Index { White space } +formfeed @Index { Formfeed } +space.f @Index { Space } + Lout treats the formfeed character exactly like the space character; +it is useful for getting page breaks when printing Lout source code. +@PP +A @I delimiter is a sequence of one or more `other' characters which +delimiter @Index { Delimiter } +is the name of a symbol. For example, @Code "{" and @Code "//" are +delimiters. When defining a delimiter, the name must be enclosed +in quotes: +@ID @Code { +"def \"^\" { {} ^& {} }" +} +but quotes are not used when the delimiter is invoked. A delimiter may +have delimiters and any other characters adjacent, whereas identifiers +may not be adjacent to letters or other identifiers. The complete list +of predefined delimiters is +@ID @OneRow @Code { +{ + "/" + @JL "//" + @JL "^/" + @JL "^//" +} |2.2cx { + "|" + @JL "||" + @JL "^|" + @JL "^||" +} |2.2cx { + "&" + @JL "^&" +} |2.2cx { + "&&" + @JL "{" + @JL "}" +} +} +A longer delimiter like @Code "<=" will be recognised in +preference to a shorter one like {@Code "<"}. +@PP +An @I identifier is a sequence of one or more letters which is the name of a +identifier @Index { Identifier } +symbol. It is conventional but not essential to begin identifiers with +{@Code "@"}; Basser Lout will print a warning message if it finds an +unquoted literal word (see below) beginning with {@Code "@"}, since such +words are usually misspelt identifiers. The ten digits are not letters +and may not appear in identifiers; and although the underscore character +is a letter and may be used in identifiers, it is not conventional to +do so. The complete list of predefined identifiers is +@ID @OneRow @Code { +{ "@BackEnd" + @JL "@Background" + @JL "@Begin" + @JL "@Break" + @JL "@Case" + @JL "@Common" + @JL "@Char" + @JL "@CurrFace" + @JL "@CurrFamily" + @JL "@CurrLang" + @JL "@Database" + @JL "@End" + @JL "@Enclose" + @JL "@Filter" + @JL "@FilterErr" + @JL "@FilterIn" + @JL "@FilterOut" + @JL "@Font" + @JL "@ForceGalley" + @JL "@Galley" + @JL "@Graphic" + @JL "@HAdjust" + @JL "@HContract" + @JL "@HCover" + @JL "@HExpand" + @JL "@High" + @JL "@HLimited" + @JL "@HScale" + @JL "@HShift" + @JL "@HSpan" +} |4.4cx { + "@Include" + @JL "@IncludeGraphic" + @JL "@Insert" + @JL "@KernShrink" + @JL "@Key" + @JL "@Language" + @JL "@LClos" + @JL "@LEnv" + @JL "@LInput" + @JL "@LVis" + @JL "@LUse" + @JL "@Meld" + @JL "@Merge" + @JL "@Minus" + @JL "@Moment" + @JL "@Next" + @JL "@NotRevealed" + @JL "@Null" + @JL "@OneCol" + @JL "@OneOf" + @JL "@OneRow" + @JL "@Open" + @JL "@Optimize" + @JL "@PAdjust" + @JL "@PageLabel" + @JL "@PlainGraphic" + @JL "@Plus" + @JL "@PrependGraphic" + @JL "@RawVerbatim" + @JL "@Rotate" +} |4.4cx { + "@Rump" + @JL "@Scale" + @JL "@SetColor" + @JL "@SetColour" + @JL "@Space" + @JL "@StartHSpan" + @JL "@StartHVSpan" + @JL "@StartVSpan" + @JL "@SysDatabase" + @JL "@SysInclude" + @JL "@SysIncludeGraphic" + @JL "@SysPrependGraphic" + @JL "@Tag" + @JL "@Tagged" + @JL "@Target" + @JL "@Underline" + @JL "@Use" + @JL "@VAdjust" + @JL "@VContract" + @JL "@VCover" + @JL "@Verbatim" + @JL "@VExpand" + @JL "@VLimited" + @JL "@VScale" + @JL "@VShift" + @JL "@VSpan" + @JL "@Wide" + @JL "@Yield" + @JL "@YUnit" + @JL "@ZUnit" +} +} +plus the names of the parameters of @@Moment. The symbols @@LClos, @@LEnv, +lclos @Index { @@LClos symbol } +lenv @Index { @@LEnv symbol } +linput @Index { @@LInput symbol } +lvis @Index { @@LVis symbol } +luse @Index { @@LUse symbol } +@@LInput, @@LVis and @@LUse appear in cross reference databases generated +by Lout and are not for use elsewhere. +@PP +A sequence of characters which is neither a white space, an identifier, nor a +delimiter, is by default a @I {literal word}, which means that it will +word @Index { Word } +literal.word @Index { Literal word } +quoted.word @Index { Quoted word } +pass through Lout unchanged. An arbitrary sequence of characters +enclosed in double quotes, for example @Code "\"{ }\"", is also a +literal word. Space characters may be included, but not tabs or +newlines. There are special character sequences, used only between +quotes, for obtaining otherwise inaccessible characters: +@ID @Tab + vmargin { 0.5vx } + @Fmta { @Col A ! @Col B } +{ +@Rowa A { @Code "\\\"" } B { produces @Code "\"" } +@Rowa A { @Code "\\\\" } B { "\\" } +@Rowa A { @Code "\\ddd" } B { the character whose ASCII code is } +@Rowa A { } B { the up to three digit octal number {@Code ddd} } +} +So, for example, @Code "\"\\\"@PP\\\"\"" produces {@Code "\"@PP\""}. +@PP +When the comment character +comment @Index { Comment } +@Code "#" is encountered, everything from +that point to the end of the line is ignored. This is useful for +including reminders to oneself, like this: +@ID @OneRow @Code { +"# Lout user manual" +"# J. Kingston, June 1989" +} +for temporarily deleting parts of the document, and so on. +@PP +@I Macros +macro @Index { Macro } +provide a means of defining symbols which stand for a +sequence of textual units rather than an object. For example, the macro +definition +@ID @Code { +"macro @PP { //1.3vx 2.0f @Wide &0i }" +} +makes Lout replace the symbol @Code "@PP" by the given textual units +before assembling its input into objects. A similar macro to this +one is used to separate the paragraphs of the present document. The +enclosing braces and any spaces adjacent to them are dropped, which can +be a problem: @Code "@PP2i" has result {@Code "//1.3vx 2.0f @Wide &0i2i"} +which is erroneous. +@PP +The meaning of symbols used within the body of a macro is determined by +where the macro is defined, not by where it is used. Due to implementation +problems, @@Open symbols will not work within macros. Named and body +parameters will work if the symbol that they are parameters of is also +present. There is no way to get a left or right brace into the body of +a macro without the matching brace. +@PP +Macros may be nested within other definitions and exported, but they may +not be parameters. They may not have parameters or nested definitions +of their own, and consequently a preceding @Code export clause (Section +{@NumberOf visibility}) would be pointless; however, an @Code import +clause is permitted. +@End @Section |