diff options
| author | rofl0r <retnyg@gmx.net> | 2012-12-25 06:03:24 +0100 |
|---|---|---|
| committer | rofl0r <retnyg@gmx.net> | 2012-12-25 06:03:48 +0100 |
| commit | 92e039f526ef7988a9b8cf500fd45a3a5b045ced (patch) | |
| tree | eb6f9945eec57d64c0eaf0c2cc885e5acf4311c2 /Makefile | |
| parent | ddf6dfea27f4a7d57ed73a4f5a146b6b704e9103 (diff) | |
| download | gettext-tiny-92e039f526ef7988a9b8cf500fd45a3a5b045ced.tar.gz | |
fix assertion error on huge string found in gnumericv0.0.2
the culprit is:
gnumeric-1.10.17/po-functions/dz.po
basically the code was wrongly checking for the end of the convert
buffer, instead of the line buffer.
this is no problem per se because the line buffer gets reset more
often, but when the convertbuffer is bigger (like after this change)
it wouldnt work correctly.
i put the convert_buf functions into a separate translation unit
and renamed them to escape and unescape to make it more clear
what they're doing. also new: size checks for the escape functions.
currently it will silently truncate the strings when the buffer
runs out of space. this is sufficient for no-op style, but should
be fixed at some point.
Diffstat (limited to 'Makefile')
| -rw-r--r-- | Makefile | 5 |
1 files changed, 3 insertions, 2 deletions
@@ -7,6 +7,7 @@ sysconfdir=$(prefix)/etc LIBSRC = $(sort $(wildcard libintl/*.c)) PROGSRC = $(sort $(wildcard src/*.c)) +PARSEROBJS = src/poparser.o src/StringEscape.o PROGOBJS = $(PROGSRC:.c=.o) LIBOBJS = $(LIBSRC:.c=.o) OBJS = $(PROGOBJS) $(LIBOBJS) @@ -46,10 +47,10 @@ libintl.a: $(LIBOBJS) $(RANLIB) $@ msgmerge: $(OBJS) - $(CC) $(LDFLAGS) -static -o $@ src/msgmerge.o src/poparser.o + $(CC) $(LDFLAGS) -static -o $@ src/msgmerge.o $(PARSEROBJS) msgfmt: $(OBJS) - $(CC) $(LDFLAGS) -static -o $@ src/msgfmt.o src/poparser.o + $(CC) $(LDFLAGS) -static -o $@ src/msgfmt.o $(PARSEROBJS) xgettext: cp src/xgettext.sh ./xgettext |