1 An approximate method to merge from upstream is:
3 # Fetch latest from upstream (we also include some compat stuff)
4 $ CVS_RSH=ssh; export CVS_RSH
5 $ CVSROOT="anoncvs@anoncvs.netbsd.org:/cvsroot"
6 $ cvs co -d libedit -P src/lib/libedit
8 $ for f in src/common/lib/libc/string/strlcat.c \
9 > src/common/lib/libc/string/strlcpy.c \
11 > src/lib/libc/gen/unvis.c \
12 > src/lib/libc/gen/vis.c \
13 > src/tools/compat/fgetln.c
21 # Remove files we don't need/use
22 $ rm -rf CVS TEST Makefile shlib_version *.[0-9]
23 $ (cd readline; rm -rf CVS Makefile)
25 # Rename files to match our naming
26 $ mv makelist makelist.sh
27 $ mv terminal.h el_terminal.h
29 # Remove NetBSD-specific bits
30 $ for file in $(find . -type f)
32 > cp ${file} ${file}.orig
33 > sed -e 's/#include "terminal.h"/#include "el_terminal.h"/g' \
34 > -e 's/sig_handler/el_sig_handler/g' \
35 > -e 's/isprint/el_isprint/g' \
37 > ${file}.orig >${file}
41 then merge remaining bits by hand. All MySQL-specific changes should be
42 marked with XXXMYSQL to make them easier to identify and merge. To generate
43 a 'clean' diff against upstream you can use the above commands but use
45 cvs co -D "2011/10/23 17:37:55" [..]
47 to fetch the baseline of most recent merge.
49 Please feed any fixes to Jonathan Perkin <jonathan.perkin@oracle.com> who will
50 endeavour to merge them upstream and keep diffs minimal.