Support more separator characters

[sebastiankuhnert]

• New separator "+" (and "&"?): Book/chapter/verse (range) before, same or higher level behind.
• New separator ".": Verse (range) before, verse (range) behind. (Complication: the parts before and behind might together form the name of a numbered book)
• Modified separator ",": depending on context either chapter/verse (range) before, same level or chapter behind (maybe also book?). or: chapter before, verse behind. In bibleref-parse, the behavior of "," can be switched using the following options:
  • comma=list: Always the former behavior
  • comma=chvsep: Always the latter behavior
  • comma=likeBR: The latter behavior if \BRchvsep contains a comma, the former otherwise
  • comma=preferchvsep: The letter behavior unless the specified passage contains a colon
  • comma=preferlikeBR (default in bibleref-parse): The latter behaviour if \BRchvsep contains a comma and the passage contains no colon, and the former otherwise.
• New separator newline: Act like semicolon or plus. Implementation might require catcode changes or clever uses of something like \expandafter\NextParserMacro\string#1.

[chadparry]

There is a mitigating factor that will make the "." parsing not so complicated. The book parsing is a separate command. It consumes as many characters as it can, as long as it still matches a valid book alias. For example, suppose there were a book named (inexplicably) "Jasher2", and a user input of "Jasher 2 5". The parser would try to match "Jasher25" against a book name, and it would fail, and then it would try to match "Jasher2" and it would succeed. It continues shortening the user input until it finds a match. (That's also why 1Corinthians would be considered a better match than 1Cor). So with regards to the period, the book name parser will consume it if at all possible. If a period is left in the input stream, then it can be considered a separator.

[sebastiankuhnert]

Ah - I see. This indeed makes "." parsing easier than the corresponding code in bibleref-parse.