先前找 ports directory 的 script 效能太差，已用 C 重寫

3 files changed, 145 insertions, 29 deletions

diff --git a/Makefile b/Makefile
index a6af80c..d0837c2 100644
--- a/Makefile
+++ b/Makefile
@@ -4,7 +4,7 @@
 
 all: debian freebsd
 
-.pkg.list: packages.h
+.pkg.list: packages.h packages.sh
 	@echo "==> Generating list file $@"
 	./packages.sh `echo "$<" | cut -d . -f 1` | sort | uniq > "$@" || rm -f "$@"
 
@@ -13,9 +13,11 @@ debian.control: debian.list
 
 freebsd: freebsd.makefile
 freebsd.makefile: freebsd.list freebsd.ports
-freebsd.ports: freebsd.list
+freebsd.ports: freebsd.list freebsd.ports.sh freebsd.ports.find
 	@echo "==> Generating $@"
 	./freebsd.ports.sh freebsd.list > "$@" || rm -f "$@"
+freebsd.ports.find: freebsd.ports.find.c
+	c99 -DHASH_TABLE_SIZE=50000 "$<" -o "$@"
 
 clean:
 	rm -f *.control *.makefile *.ports *.list
diff --git a/freebsd.ports.find.c b/freebsd.ports.find.c
new file mode 100644
index 0000000..309ea89
--- /dev/null
+++ b/freebsd.ports.find.c
@@ -0,0 +1,139 @@
+#define _POSIX_C_SOURCE 200809L
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <search.h>
+
+#ifndef HASH_TABLE_SIZE
+# define HASH_TABLE_SIZE 50000
+#endif
+
+#define msg(...)     fprintf(stderr, __VA_ARGS__)
+#define msgstr(x)    fputs((x), stderr)
+
+static void write_maps (const char* list_file) {
+	FILE* fp = fopen (list_file, "r");
+	if (fp == NULL) {
+		msg ("Cannot open list file %s\n", list_file);
+		perror ("fopen");
+		exit (3);
+	}
+
+	char* line = NULL;
+	size_t len = 0;
+	unsigned n;
+
+	for (n = 0; getline (&line, &len, fp) >= 0; n++) {
+		strtok (line, "\n");
+
+		ENTRY hfind = {
+			.key = line,
+			.data = NULL
+		};
+		ENTRY* hresult = hsearch (hfind, FIND);
+		if (hresult == NULL) {
+			msg ("Cannot find origin for %s\n", line);
+			exit (4);
+		}
+
+		printf ("%-32s %s\n", line, hresult->data);
+		line = NULL;
+		len = 0;
+	}
+
+	if (!feof (fp)) {
+		msgstr ("Fail to read the whole list file\n");
+		perror ("getline");
+		exit (6);
+	}
+
+	msg ("=> %u packages entries processed\n", n);
+	fclose (fp);
+}
+
+static void read_entries (const char* index_file) {
+	FILE* fp = fopen (index_file, "r");
+	if (fp == NULL) {
+		msg ("Cannot open index file %s\n", index_file);
+		perror ("fopen");
+		exit (3);
+	}
+
+	char* line = NULL;
+	size_t len = 0;
+	unsigned n;
+
+	for (n = 0; getline (&line, &len, fp) >= 0; n++) {
+		strtok (line, "\n");
+
+		char* pkgname = strtok (line, "|");
+		if (pkgname == NULL) {
+			msgstr ("Malformed index file - no package name\n");
+			exit (4);
+		}
+
+		char* pkgname_end = strrchr (line, '-');
+		if (pkgname_end != NULL) {
+			*pkgname_end = '\0';
+		}
+
+		char* fullpath = strtok (NULL, "|");
+		if (fullpath == NULL) {
+			msgstr ("Malformed index file - no path\n");
+			exit (4);
+		}
+
+		bool dir_sp_got = false;
+		for (char* p = fullpath + strlen (fullpath); p >= fullpath; p--) {
+			if (*p == '/') {
+				if (dir_sp_got) {
+					fullpath = p + 1;
+					break;
+				} else {
+					dir_sp_got = true;
+				}
+			}
+		}
+
+		ENTRY hentry = {
+			.key = pkgname,
+			.data = fullpath
+		};
+		if (hsearch (hentry, ENTER) == NULL) {
+			msgstr ("The hash table is full!\n");
+			exit (5);
+		}
+
+		line = NULL;
+		len = 0;
+	}
+
+	if (!feof (fp)) {
+		msgstr ("Fail to read the whole index file\n");
+		perror ("getline");
+		exit (6);
+	}
+
+	msg ("=> %u ports entries read\n", n);
+	fclose (fp);
+}
+
+int main (int argc, char* argv[]) {
+	if (argc < 3) {
+		msg ("Usage: %s list_file index_file\n", argv[0]);
+		return 1;
+	}
+
+	if (!hcreate (HASH_TABLE_SIZE)) {
+		msg ("Cannot create a hash table with size %d\n", HASH_TABLE_SIZE);
+		perror ("hcreate");
+		return 2;
+	}
+
+	read_entries (argv[2]);
+	write_maps (argv[1]);
+
+	return 0;
+}
diff --git a/freebsd.ports.sh b/freebsd.ports.sh
index ed7b89d..1ef3dc7 100755
--- a/freebsd.ports.sh
+++ b/freebsd.ports.sh
@@ -4,20 +4,6 @@ msg () {
 	echo "$@" 1>&2
 }
 
-get_origin () {
-	grep "^$1" "${index}" | (
-		IFS="${IFS}|"
-		while read pkgnamever fullpath trash; do
-			pkgname="${pkgnamever%-*}"
-			if [ "${pkgname}" = "$1" ]; then
-				this_origin="`echo "${fullpath}" | sed 's|^.*/\(.*/.*\)$|\1|'`"
-				echo "${this_origin}"
-				break
-			fi
-		done
-	)
-}
-
 [ -z "$1" ] && msg "Usage: $0 list_file" && exit 1
 : ${FREEBSD_VERSION:="`uname -r | sed 's|^\([0-9]*\).*$|\1|'`"}
 : ${PORTSDIR:="/usr/ports"}
@@ -27,18 +13,7 @@ msg "==> FreeBSD version is ${FREEBSD_VERSION}"
 msg "==> FreeBSD ports tree is ${PORTSDIR}"
 msg "==> FreeBSD ports index file is ${index}"
 
-exec 3< "$1"       # list file
-
-while read pkgname 0<&3; do
-	msg "=> Processing package ${pkgname}"
-	origin="`get_origin ${pkgname}`"
-	[ -z "${origin}" ] && \
-		msg "==> Cannot find origin for ${pkgname} in your index file" && \
-		msg "==> Exit now!" && exit 1
-	msg "=> Processing package ${pkgname} - ${origin}"
-	printf "%-32s %s\n" "${pkgname}" "${origin}"
-done
-
-exec 3<&-
+msg "==> Running freebsd.ports.find"
+./freebsd.ports.find "$1" "${index}"
 
 msg "==> Done!"