Language support Digraphs and trigraphs

1 language support

1.1 algol
1.2 pascal
1.3 vim
1.4 gnu screen
1.5 j
1.6 c
1.7 c++

1.7.1 removal of trigraphs


1.8 rpl

language support

different systems define different sets of digraphs , trigraphs, described below.

algol

early versions of algol predated standardised ascii , ebcdic character sets, , typically implemented using manufacturer-specific six-bit character code. number of algol operations either lacked codepoints in available character set or not supported peripherals, leading number of substitutions including := ← (assignment) , >= ≥ (greater or equal).

pascal

pascal programming language supports digraphs (., .), (* , *) [, ], { , } respectively. unlike other cases mentioned here, (* , *) , still in wide use.

vim

vim text editor supports digraphs actual entry of text characters, following rfc 1345.

gnu screen

gnu screen has digraph command, bound ^a ^v default.

j

the j programming language descendant of apl using ascii character set rather traditional apl symbols. handle fact printable range of ascii smaller apl s specialised set of symbols, . (dot) , : (colon) characters used inflect ascii symbols, interpreting unigraphs, digraphs or trigraphs standalone symbols .

unlike use of digraphs , trigraphs in c , c++, there no single-character equivalents these in j.

c

the c preprocessor replaces occurrences of following 9 trigraph sequences single-character equivalents before other processing.

a programmer may want place 2 question marks yet not have compiler treat them introducing trigraph. c grammar not permit 2 consecutive ? tokens, places in c file 2 question marks in row may used in multi-character constants, string literals, , comments. particularly problem classic mac os, constant ???? may used file type or creator. safely place 2 consecutive question marks within string literal, programmer can use string concatenation ...? ?... or escape sequence ...?\?... .

??? not trigraph sequence, when followed character such - interpreted ? + ??-, in example below has 16 ?s before /.

the ??/ trigraph can used introduce escaped newline line splicing; must taken account correct , efficient handling of trigraphs within preprocessor. can cause surprises, particularly within comments. example:

// next line executed????????????????/
a++;

which single logical comment line (used in c++ , c99), and

/??/
* comment *??/
/

which correctly formed block comment.

in 1994, normative amendment c standard, included in c99, supplied digraphs more readable alternatives 5 of trigraphs. listed in table on right.

unlike trigraphs, digraphs handled during tokenization, , digraph must represent full token itself, or compose token %:%: replacing preprocessor concatenation token ##. if digraph sequence occurs inside token, example quoted string, or character constant, not replaced.

c++

c++ (through c++14, see below) behaves c, including c99 additions, additional tokens listed in table.

as note, %:%: treated single token, rather 2 occurrences of %:.

the c++ standard makes comment regards term digraph :

the term digraph (token consisting of 2 characters) not descriptive, since 1 of alternative preprocessing-tokens %:%: , of course several primary tokens contain 2 characters. nonetheless, alternative tokens aren’t lexical keywords colloquially known digraphs .

removal of trigraphs

trigraphs proposed deprecation in c++0x, released c++11. opposed ibm, speaking on behalf of , other users of c++, , result trigraphs retained in c++0x. trigraphs proposed again removal (not deprecation) in c++17. passed committee vote, , trigraphs removed c++17 despite opposition ibm. existing code uses trigraphs can supported translating source files (parsing trigraphs) basic source character set not include trigraphs.

rpl

hewlett-packard calculators supporting rpl language , input method provide support large number of trigraphs (also called tio codes) reliably transcribe non-seven-bit ascii characters of calculators extended character set on foreign platforms, , ease keyboard input without using chars application.