Commit ca62bd8bc703147314a94e20a4b4ea7c5a7b2d37

Ivan Zakharyaschev
Thu Oct 16 01:54:55 CEST 2008

Tree SHA1: 7424a8a
Parent SHA1: f67e7ee (Rm excessive regexp (already covered by previous check).)
raw diff | raw patch

Improved the well-f. check to ignore filenames
(by implementing a more generic mechanism for checks).

Now the checks are assumed destructive, and the hold space is used in
the implementation.

Diff rendering mode:
inline
side by side


42	42	##############################################
43	43	# N:
44	44	s\|N\|ŋ\|g # velar or nasalization
	45	h # Save before the destructive checks (copy pattern space to hold space).
45	46	##########################################
46		# Well-formedness of the input:
	47	# Well-formedness of the input (is checked destructively):
47	48	# - no ^'s should remain;
48	49	# - other capitals should not have been used (no S^, no C^ and no plain capitals);
49	50	# - more?
50		/\$Date.*\$$/ n # skip the 'preamble'
	51	s\|\$Date.*\$\|\| # ignore the 'preamble'
	52	s\|\\url{.*}\|\| # ignore filenames, which have been marked up like this: \url{Komi_rus_slovar1.pdf}
51	53	/\^/ {
52	54	w /dev/stderr
53	55	q 56;} # (mostly GNU sed extensions)
54	56	/$^\\|[^\\]$[[:upper:]]/ { # [^\\] -- a workaround for grammatical terms,
55	57	# to escape these restrictions on Latin characters: write them like \Pres, \Gen etc.
56		# FIXME: I also use this workaround for filenames, e.g.: \Komi_rus_slovar1.pdf, but
57		# this should be replaced for the sake of universality with smth like
58		# \url{Komi_rus_slovar1.pdf}.
59	58	w /dev/stderr
60	59	q 57;} # (mostly GNU sed extensions)
	60	g # Restore after the destructive checks (copy/append hold space to pattern space).
61	61	"
62	62
63	63	# FIXME: A "keep-going" mode: detect all the non-well-formednesses, and accumulate all

Commit ca62bd8bc703147314a94e20a4b4ea7c5a7b2d37

Gitorious