Mercurial > hg > xemacs-beta
view lib-src/rcs-checkin @ 4549:68d1ca56cffa
First part of interactive checks that coding systems encode regions.
2008-01-21 Aidan Kehoe <kehoea@parhasard.net>
* coding.el (decode-coding-string):
(encode-coding-string): Accept GNU's NOCOPY argument for
these. Todo; write compiler macros to use it.
(query-coding-warning-face): New face, to show unencodable
characters.
(default-query-coding-region-safe-charset-skip-chars-map):
New variable, a cache used by #'default-query-coding-region.
(default-query-coding-region): Default implementation of
#'query-coding-region, using the safe-charsets and safe-chars
coding systemproperties.
(query-coding-region): New function; can a given coding system
encode a given region?
(query-coding-string): New function; can a given coding system
encode a given string?
(unencodable-char-position): Function API taken from GNU; return
the first unencodable position given a string and coding system.
(encode-coding-char): Function API taken from GNU; return CHAR
encoded using CODING-SYSTEM, or nil if CODING-SYSTEM would trash
CHAR.
((unless (featurep 'mule)): Override the default
query-coding-region implementation on non-Mule.
* mule/mule-coding.el (make-8-bit-generate-helper): Eliminate a
duplicate comment.
(make-8-bit-choose-category): Simplify implementation.
(8-bit-fixed-query-coding-region): Implementation of
#'query-coding-region for coding systems created with
#'make-8-bit-coding-system.
(make-8-bit-coding-system): Initialise the #'query-coding-region
implementation for these character sets.
(make-8-bit-coding-system): Ditto for the compiler macro version
of this function.
* unicode.el (unicode-query-coding-skip-chars-arg): New variable,
used by unicode-query-coding-region, initialised in
mule/general-late.el.
(unicode-query-coding-region): New function, the
#'query-coding-region implementation for Unicode coding systems.
Initialise the query-coding-function property for the Unicode
coding systems to #'unicode-query-coding-region.
* mule/mule-charset.el (charset-skip-chars-string): New
function. Return a #'skip-chars-forward argument that skips all
characters in CHARSET.
(map-charset-chars): Function synced from GNU, modified to work
with XEmacs. Map FUNC across the int value charset ranges of
CHARSET.
author | Aidan Kehoe <kehoea@parhasard.net> |
---|---|
date | Mon, 21 Jan 2008 22:51:21 +0100 |
parents | 3ecd8885ac67 |
children |
line wrap: on
line source
#! /bin/sh # This script accepts any number of file arguments and checks them into RCS. # # Arguments which are detectably either RCS masters (with names ending in ,v) # or Emacs version files (with names of the form foo.~<number>~) are ignored. # For each file foo, the script looks for Emacs version files related to it. # These files are checked in as deltas, oldest first, so that the contents of # the file itself becomes the latest revision in the master. # # The first line of each file is used as its description text. The file itself # is not deleted, as under VC with vc-keep-workfiles at its default of t, but # all the version files are. # # If an argument file is already version-controlled under RCS, any version # files are added to the list of deltas and deleted, and then the workfile # is checked in again as the latest version. This is probably not quite # what was wanted, and is the main reason VC doesn't simply call this to # do checkins. # # This script is intended to be used to convert files with an old-Emacs-style # version history for use with VC (the Emacs 19 version-control interface), # which likes to use RCS as its back end. It was written by Paul Eggert # and revised/documented for use with VC by Eric S. Raymond, Mar 19 1993. case $# in 0) echo "rcs-checkin: usage: rcs-checkin file ..." echo "rcs-checkin: function: checks file.~*~ and file into a new RCS file" echo "rcs-checkin: function: uses the file's first line for the description" esac # expr pattern to extract owner from ls -l output ls_owner_pattern='[^ ][^ ]* *[^ ][^ ]* *\([^ ][^ ]*\)' for file do # Make it easier to say `rcs-checkin *' # by ignoring file names that already contain `~', or end in `,v'. case $file in *~* | *,v) continue esac # Ignore non-files too. test -f "$file" || continue # Check that file is readable. test -r "$file" || exit # If the RCS file does not already exist, # initialize it with a description from $file's first line. rlog -R "$file" >/dev/null 2>&1 || rcs -i -q -t-"`sed 1q $file`" "$file" || exit # Get list of old files. oldfiles=` ls $file.~[0-9]*~ 2>/dev/null | sort -t~ -n +1 ` # Check that they are properly sorted by date. case $oldfiles in ?*) oldfiles_by_date=`ls -rt $file $oldfiles` test " $oldfiles $file" = " $oldfiles_by_date" || { echo >&2 "rcs-checkin: skipping $file, because its mod times are out of order. Sorted by mod time: $oldfiles_by_date Sorted by name: $oldfiles $file" continue } esac echo >&2 rcs-checkin: checking in: $oldfiles $file # Save $file as $file.~-~ temporarily. mv "$file" "$file.~-~" || exit # Rename each old file to $file, and check it in. for oldfile in $oldfiles do mv "$oldfile" "$file" || exit ls_l=`ls -l "$file"` || exit owner=-w`expr " $ls_l" : " $ls_owner_pattern"` || owner= echo "Formerly ${oldfile}" | ci -d -l -q $owner "$file" || exit done # Bring $file back from $file.~-~, and check it in. mv "$file.~-~" "$file" || exit ls_l=`ls -l "$file"` || exit owner=-w`expr " $ls_l" : " $ls_owner_pattern"` || owner= ci -d -q -u $owner -m"entered into RCS" "$file" || exit done