Moritz Wilhelmy dixit: >> The existence of a line consisting only of two dots is usually >> a good indicator for the presence of custom macros. A line >> consisting only of a dot, a backslash and a closing curly brace >> (b >Interesting. Would be nice to have it fall back to something >sensible in this case.
Indeed. Maybe there could be a grep or, in the case of compressed manpages (which weren’t used by default on BSDs last time I looked) some kind of buffering… Well, the following POSIX BREs show evidence of it being a roff document, not an mdocml compatible manpage: ^\.\.$ ^\.\\}$ ^\.[ ]*de[ ] ^\.i[ef][ ] >> (While here, could TNF please use \*(Gt instead of \*[Gt] so >> their manpages are nroff compatible, since two-character named >> references can be written either way?) > >As someone who doesn't really know much about roff, how do they >differ? \*[foo] is a GNU groff extension to have arbitrarily-long named strings, whereas \*X and \*(XY are the standard nroff/troff/ditroff string named references (string, macro, register, etc. names are either one or two characters in classic roff). bye, //mirabilos -- Sometimes they [people] care too much: pretty printers [and syntax highligh- ting, d.A.] mechanically produce pretty output that accentuates irrelevant detail in the program, which is as sensible as putting all the prepositions in English text in bold font. -- Rob Pike in "Notes on Programming in C"
