the header file has size_t and that seems to make more sense. no bump in pkgname since there is no change except for on 64-bit architectures where it didn't built in the first place. comments?
Index: patches/patch-crypto_c =================================================================== RCS file: patches/patch-crypto_c diff -N patches/patch-crypto_c --- /dev/null 1 Jan 1970 00:00:00 -0000 +++ patches/patch-crypto_c 8 Sep 2005 00:24:03 -0000 @@ -0,0 +1,39 @@ +$OpenBSD$ +--- crypto.c.orig Wed Sep 7 18:22:23 2005 ++++ crypto.c Wed Sep 7 18:22:58 2005 +@@ -60,7 +60,7 @@ + * bits. + */ + +-void md4lite(char *message, unsigned int len) ++void md4lite(char *message, size_t len) + { + + #ifndef HAVE_MD4 +@@ -92,7 +92,7 @@ void md4lite(char *message, unsigned int + * bits. + */ + +-void md5lite(char *message, unsigned int len) ++void md5lite(char *message, size_t len) + { + + #ifndef HAVE_MD5 +@@ -124,7 +124,7 @@ void md5lite(char *message, unsigned int + * into 64 bits. + */ + +-void rmd160lite(char *message, unsigned int len) ++void rmd160lite(char *message, size_t len) + { + + #ifndef HAVE_RMD160 +@@ -175,7 +175,7 @@ void rmd160lite(char *message, unsigned + * into 64 bits. + */ + +-void sha1lite(char *message, unsigned int len) ++void sha1lite(char *message, size_t len) + { + + #ifndef HAVE_SHA1