diff: small optimizations; do not try to diff non-seekable stream

(currently we don't support that)
sort: fixes. testsuites fixed:
sort with non-default leading delim 1
sort with non-default leading delim 2
sort key doesn't strip leading blanks, disables fallback global sort

diff --git a/coreutils/diff.c b/coreutils/diff.c
index 923239ab17793764805e062e671fc628c8c94611..2ce681affd3df4adf645bd4948cee7cd4e41ac79 100644 (file)
--- a/coreutils/diff.c
+++ b/coreutils/diff.c
@@ -67,8 +67,12 @@
 
 /* XXX: FIXME: the following variables should be static, but gcc currently
  * creates a much bigger object if we do this. [which version of gcc? --vda] */
-int context, status;
-char *start, *label[2];
+/* This is the default number of lines of context. */
+int context = 3;
+int status;
+char *start;
+const char *label1;
+const char *label2;
 struct stat stb1, stb2;
 char **dl;
 USE_FEATURE_DIFF_DIR(static int dl_count;)
@@ -276,7 +280,8 @@ static void prepare(int i, FILE * fd, off_t filesize)
                sz = 100;
 
        p = xmalloc((sz + 3) * sizeof(struct line));
-       for (j = 0; (h = readhash(fd));) {
+       j = 0;
+       while ((h = readhash(fd))) {
                if (j == sz) {
                        sz = sz * 3 / 2;
                        p = xrealloc(p, (sz + 3) * sizeof(struct line));
@@ -735,14 +740,14 @@ static void dump_unified_vec(FILE * f1, FILE * f2)
 
 static void print_header(const char *file1, const char *file2)
 {
-       if (label[0] != NULL)
-               printf("%s %s\n", "---", label[0]);
+       if (label1)
+               printf("--- %s\n", label1);
        else
-               printf("%s %s\t%s", "---", file1, ctime(&stb1.st_mtime));
-       if (label[1] != NULL)
-               printf("%s %s\n", "+++", label[1]);
+               printf("--- %s\t%s", file1, ctime(&stb1.st_mtime));
+       if (label2)
+               printf("+++ %s\n", label2);
        else
-               printf("%s %s\t%s", "+++", file2, ctime(&stb2.st_mtime));
+               printf("+++ %s\t%s", file2, ctime(&stb2.st_mtime));
 }
 
 
@@ -771,8 +776,7 @@ static void change(char *file1, FILE * f1, char *file2, FILE * f2, int a,
 
                max_context <<= 1;
                context_vec_start = xrealloc(context_vec_start,
-                                                                        max_context *
-                                                                        sizeof(struct context_vec));
+                               max_context * sizeof(struct context_vec));
                context_vec_end = context_vec_start + max_context;
                context_vec_ptr = context_vec_start + offset;
        }
@@ -894,8 +898,8 @@ static int diffreg(char *ofile1, char *ofile2, int flags)
 {
        char *file1 = ofile1;
        char *file2 = ofile2;
-       FILE *f1 = NULL;
-       FILE *f2 = NULL;
+       FILE *f1;
+       FILE *f2;
        int rval = D_SAME;
        int i;
 
@@ -912,12 +916,18 @@ static int diffreg(char *ofile1, char *ofile2, int flags)
                f1 = xfopen(bb_dev_null, "r");
        else if (NOT_LONE_DASH(file1))
                f1 = xfopen(file1, "r");
-
        if (flags & D_EMPTY2)
                f2 = xfopen(bb_dev_null, "r");
        else if (NOT_LONE_DASH(file2))
                f2 = xfopen(file2, "r");
 
+/* We can't diff non-seekable stream - we use rewind(), fseek().
+ * This can be fixed (volunteers?).
+ * Meanwhile we should check it here by stat'ing input fds,
+ * but I am lazy and check that in main() instead.
+ * Check in main won't catch "diffing fifos buried in subdirectories"
+ * failure scenario - not very likely in real life... */
+
        i = files_differ(f1, f2, flags);
        if (i == 0)
                goto closem;
@@ -1169,48 +1179,32 @@ int diff_main(int argc, char **argv)
        char *f1, *f2;
        llist_t *L_arg = NULL;
 
-       opt_complementary = "L::";
+       /* exactly 2 params; collect multiple -L <label> */
+       opt_complementary = "=2:L::";
        getopt32(argc, argv, "abdiL:NqrsS:tTU:wu"
                        "p" /* ignored (for compatibility) */,
                        &L_arg, &start, &U_opt);
-
-       if (option_mask32 & FLAG_L) {
-               while (L_arg) {
-                       if (label[0] == NULL)
-                               label[0] = L_arg->data;
-                       else if (label[1] == NULL)
-                               label[1] = L_arg->data;
-                       else
-                               bb_show_usage();
-
-                       L_arg = L_arg->link;
-               }
-
-               /* If both label[0] and label[1] were set, they need to be swapped. */
-               if (label[0] && label[1]) {
-                       char *tmp;
-
-                       tmp = label[1];
-                       label[1] = label[0];
-                       label[0] = tmp;
+       /*argc -= optind;*/
+       argv += optind;
+       while (L_arg) {
+               if (label1 && label2)
+                       bb_show_usage();
+               if (!label1)
+                       label1 = L_arg->data;
+               else { /* then label2 is NULL */
+                       label2 = label1;
+                       label1 = L_arg->data;
                }
+               /* we leak L_arg here... */
+               L_arg = L_arg->link;
        }
-
-       context = 3;            /* This is the default number of lines of context. */
-       if (option_mask32 & FLAG_U) {
+       if (option_mask32 & FLAG_U)
                context = xatou_range(U_opt, 1, INT_MAX);
-       }
-       argc -= optind;
-       argv += optind;
 
        /*
         * Do sanity checks, fill in stb1 and stb2 and call the appropriate
         * driver routine.  Both drivers use the contents of stb1 and stb2.
         */
-       if (argc < 2) {
-               bb_error_msg("missing filename");
-               bb_show_usage();
-       }
 
        f1 = argv[0];
        f2 = argv[1];
@@ -1239,8 +1233,12 @@ int diff_main(int argc, char **argv)
                }
                if (S_ISDIR(stb2.st_mode)) {
                        f2 = concat_path_file(f2, f1);
-                       xstat(argv[1], &stb2);
+                       xstat(f2, &stb2);
                }
+/* We can't diff e.g. stdin supplied by a pipe - we use rewind(), fseek().
+ * This can be fixed (volunteers?) */
+               if (!S_ISREG(stb1.st_mode) || !S_ISREG(stb2.st_mode))
+                       bb_error_msg_and_die("can't diff non-seekable stream");
                print_status(diffreg(f1, f2, 0), f1, f2, NULL);
        }
        return status;

diff --git a/coreutils/sort.c b/coreutils/sort.c
index c7abc3355124db63c4cf1cad67e1be11426c145c..29cadf5f67dce164128dbd789f9611ea0c6195fa 100644 (file)
--- a/coreutils/sort.c
+++ b/coreutils/sort.c
@@ -75,14 +75,10 @@ static char *get_key(char *str, struct sort_key *key, int flags)
                        end = 0;
                        for (i = 1; i < key->range[2*j] + j; i++) {
                                if (key_separator) {
-                                       /* Skip first separator */
-                                       while (str[end] == key_separator)
-                                               end++;
-                                       /* Skip body of key */
+                                       /* Skip body of key and separator */
                                        while (str[end]) {
-                                               if (str[end] == key_separator)
+                                               if (str[end++] == key_separator)
                                                        break;
-                                               end++;
                                        }
                                } else {
                                        /* Skip leading blanks */
@@ -99,9 +95,6 @@ static char *get_key(char *str, struct sort_key *key, int flags)
                }
                if (!j) start = end;
        }
-       /* Key with explicit separator starts after separator */
-       if (key_separator && str[start] == key_separator)
-               start++;
        /* Strip leading whitespace if necessary */
 //XXX: skip_whitespace()
        if (flags & FLAG_b)

diff --git a/testsuite/sort.tests b/testsuite/sort.tests
index 5a4937b5818c36a8494c2f73740641a0bae17d2d..df5f7c7dd069c60e82c8bef3c4e8cf7a48707faf 100755 (executable)
--- a/testsuite/sort.tests
+++ b/testsuite/sort.tests
@@ -66,18 +66,38 @@ testing "sort key range with multiple options" "sort -k2,3rn input" \
 egg    1       2       papyrus
 " "$data" ""
 
+testing "sort with non-default leading delim 1" "sort -n -k2 -t/ input" "\
+/a/2
+/b/1
+" "\
+/a/2
+/b/1
+" ""
+
+testing "sort with non-default leading delim 2" "sort -n -k3 -t/ input" "\
+/b/1
+/a/2
+" "\
+/b/1
+/a/2
+" ""
+
+testing "sort with non-default leading delim 3" "sort -n -k3 -t/ input" "\
+//a/2
+//b/1
+" "\
+//a/2
+//b/1
+" ""
+
+testing "sort -u should consider field only when discarding" "sort -u -k2 input" "\
+a c
+" "\
+a c
+b c
+" ""
+
 testing "sort key doesn't strip leading blanks, disables fallback global sort" \
 "sort -n -k2 -t ' '" " a \n 1 \n 2 \n" "" " 2 \n 1 \n a \n"
 
-testing "sort key edge case with -t" "sort -n -k4 -t/" \
-"/usr/lib/finish-install.d/1
-/usr/lib/finish-install.d/4
-/usr/lib/prebaseconfig.d/2
-/usr/lib/prebaseconfig.d/6
-" "" "/usr/lib/finish-install.d/1
-/usr/lib/prebaseconfig.d/2
-/usr/lib/finish-install.d/4
-/usr/lib/prebaseconfig.d/6
-"
-
 exit $FAILCOUNT