Pretty late reply, sorry. At Mon, 19 Feb 2007 15:23:24 +0100, Michael Haller <lolcy@xxxxxx> wrote: > > Hello, > > there arose problems when I tried to link cryptsetup-luks 1.0.4 against > dietlibc (http://www.fefe.de/dietlibc/) instead of glibc. I detected > some minor compile-time problems and one run-time problem with sscanf. > > I wrote the following patch to fix these (without digging into > autoconf/automake). I hope it proves helpful. Updating automake results in an equivalent update. The scanf change seems to be ok. > @@ -290,7 +290,7 @@ > done > cryptsetup$(EXEEXT): $(cryptsetup_OBJECTS) $(cryptsetup_DEPENDENCIES) > @rm -f cryptsetup$(EXEEXT) > - $(LINK) $(cryptsetup_LDFLAGS) $(cryptsetup_OBJECTS) $(cryptsetup_LDADD) $(LIBS) > + $(LINK) $(cryptsetup_LDFLAGS) $(cryptsetup_OBJECTS) $(cryptsetup_LDADD) $(LIBS) -lcompat > > mostlyclean-compile: > -rm -f *.$(OBJEXT) This can't go in, as we don't have a libcompat in glibc. Do you have an idea how we could conditionalize this on dietlibc? Presumably an autoconf test. What would be the best way for me to bootstrap a dietlibc environment including libuuid and libpopt? So I can test the changes. -- Fruhwirth Clemens - http://clemens.endorphin.org --------------------------------------------------------------------- dm-crypt mailing list - http://www.saout.de/misc/dm-crypt/ To unsubscribe, e-mail: dm-crypt-unsubscribe@xxxxxxxx For additional commands, e-mail: dm-crypt-help@xxxxxxxx
