4 files changed, 57 insertions, 30 deletions

diff --git a/editors/diff.c b/editors/diff.c
index a892cfdf2..ff269360f 100644
--- a/editors/diff.c
+++ b/editors/diff.c
@@ -125,6 +125,7 @@
 //usage:     "\n        -w      Ignore all whitespace"
 
 #include "libbb.h"
+#include "common_bufsiz.h"
 
 #if 0
 # define dbg_error_msg(...) bb_error_msg(__VA_ARGS__)
@@ -748,6 +749,7 @@ static int diffreg(char *file[2])
                 fp[i] = fdopen(fd, "r");
         }
 
+        setup_common_bufsiz();
         while (1) {
                 const size_t sz = COMMON_BUFSIZE / 2;
                 char *const buf0 = bb_common_bufsiz1;
diff --git a/editors/ed.c b/editors/ed.c
index a4c419099..c028b78cb 100644
--- a/editors/ed.c
+++ b/editors/ed.c
@@ -23,6 +23,7 @@
 //usage:#define ed_full_usage ""
 
 #include "libbb.h"
+#include "common_bufsiz.h"
 
 typedef struct LINE {
         struct LINE *next;
@@ -35,8 +36,8 @@ typedef struct LINE {
 #define searchString bb_common_bufsiz1
 
 enum {
-        USERSIZE = sizeof(searchString) > 1024 ? 1024
-                 : sizeof(searchString) - 1, /* max line length typed in by user */
+        USERSIZE = COMMON_BUFSIZE > 1024 ? 1024
+                 : COMMON_BUFSIZE - 1, /* max line length typed in by user */
         INITBUF_SIZE = 1024, /* initial buffer size */
 };
 
@@ -66,6 +67,7 @@ struct globals {
 #define lines              (G.lines             )
 #define marks              (G.marks             )
 #define INIT_G() do { \
+        setup_common_bufsiz(); \
         SET_PTR_TO_GLOBALS(xzalloc(sizeof(G))); \
 } while (0)
 
diff --git a/editors/sed.c b/editors/sed.c
index a0c713f58..b629e060a 100644
--- a/editors/sed.c
+++ b/editors/sed.c
@@ -86,6 +86,7 @@
 //usage:       "bar\n"
 
 #include "libbb.h"
+#include "common_bufsiz.h"
 #include "xregex.h"
 
 #if 0
@@ -161,8 +162,9 @@ struct globals {
                 int len;    /* Space allocated */
         } pipeline;
 } FIX_ALIASING;
-#define G (*(struct globals*)&bb_common_bufsiz1)
+#define G (*(struct globals*)bb_common_bufsiz1)
 #define INIT_G() do { \
+        setup_common_bufsiz(); \
         BUILD_BUG_ON(sizeof(G) > COMMON_BUFSIZE); \
         G.sed_cmd_tail = &G.sed_cmd_head; \
 } while (0)
@@ -216,23 +218,33 @@ static void cleanup_outname(void)
 
 /* strcpy, replacing "\from" with 'to'. If to is NUL, replacing "\any" with 'any' */
 
-static void parse_escapes(char *dest, const char *string, int len, char from, char to)
+static unsigned parse_escapes(char *dest, const char *string, int len, char from, char to)
 {
+        char *d = dest;
         int i = 0;
 
+        if (len == -1)
+                len = strlen(string);
+
         while (i < len) {
                 if (string[i] == '\\') {
                         if (!to || string[i+1] == from) {
-                                *dest++ = to ? to : string[i+1];
+                                if ((*d = to ? to : string[i+1]) == '\0')
+                                        return d - dest;
                                 i += 2;
+                                d++;
                                 continue;
                         }
-                        *dest++ = string[i++];
+                        i++; /* skip backslash in string[] */
+                        *d++ = '\\';
+                        /* fall through: copy next char verbatim */
                 }
-                /* TODO: is it safe wrt a string with trailing '\\' ? */
-                *dest++ = string[i++];
+                if ((*d = string[i++]) == '\0')
+                        return d - dest;
+                d++;
         }
-        *dest = '\0';
+        *d = '\0';
+        return d - dest;
 }
 
 static char *copy_parsing_escapes(const char *string, int len)
@@ -243,9 +255,8 @@ static char *copy_parsing_escapes(const char *string, int len)
         /* sed recognizes \n */
         /* GNU sed also recognizes \t and \r */
         for (s = "\nn\tt\rr"; *s; s += 2) {
-                parse_escapes(dest, string, len, s[1], s[0]);
+                len = parse_escapes(dest, string, len, s[1], s[0]);
                 string = dest;
-                len = strlen(dest);
         }
         return dest;
 }
@@ -471,7 +482,7 @@ static int parse_subst_cmd(sed_cmd_t *sed_cmd, const char *substr)
  */
 static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 {
-        static const char cmd_letters[] = "saicrw:btTydDgGhHlnNpPqx={}";
+        static const char cmd_letters[] ALIGN1 = "saicrw:btTydDgGhHlnNpPqx={}";
         enum {
                 IDX_s = 0,
                 IDX_a,
@@ -514,6 +525,8 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
         }
         /* handle edit cmds: (a)ppend, (i)nsert, and (c)hange */
         else if (idx <= IDX_c) { /* a,i,c */
+                unsigned len;
+
                 if (idx < IDX_c) { /* a,i */
                         if (sed_cmd->end_line || sed_cmd->end_match)
                                 bb_error_msg_and_die("command '%c' uses only one address", sed_cmd->cmd);
@@ -527,10 +540,11 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
                                 break;
                         cmdstr++;
                 }
-                sed_cmd->string = xstrdup(cmdstr);
+                len = strlen(cmdstr);
+                sed_cmd->string = copy_parsing_escapes(cmdstr, len);
+                cmdstr += len;
                 /* "\anychar" -> "anychar" */
-                parse_escapes(sed_cmd->string, sed_cmd->string, strlen(cmdstr), '\0', '\0');
-                cmdstr += strlen(cmdstr);
+                parse_escapes(sed_cmd->string, sed_cmd->string, -1, '\0', '\0');
         }
         /* handle file cmds: (r)ead */
         else if (idx <= IDX_w) { /* r,w */
@@ -562,8 +576,8 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 
                 cmdstr += parse_regex_delim(cmdstr, &match, &replace)+1;
                 /* \n already parsed, but \delimiter needs unescaping. */
-                parse_escapes(match, match, strlen(match), i, i);
-                parse_escapes(replace, replace, strlen(replace), i, i);
+                parse_escapes(match,   match,   -1, i, i);
+                parse_escapes(replace, replace, -1, i, i);
 
                 sed_cmd->string = xzalloc((strlen(match) + 1) * 2);
                 for (i = 0; match[i] && replace[i]; i++) {
@@ -942,13 +956,22 @@ static void puts_maybe_newline(char *s, FILE *file, char *last_puts_char, char l
         *last_puts_char = lpc;
 }
 
-static void flush_append(char *last_puts_char, char last_gets_char)
+static void flush_append(char *last_puts_char)
 {
         char *data;
 
         /* Output appended lines. */
-        while ((data = (char *)llist_pop(&G.append_head))) {
-                puts_maybe_newline(data, G.nonstdout, last_puts_char, last_gets_char);
+        while ((data = (char *)llist_pop(&G.append_head)) != NULL) {
+                /* Append command does not respect "nonterminated-ness"
+                 * of last line. Try this:
+                 * $ echo -n "woot" | sed -e '/woot/a woo' -
+                 * woot
+                 * woo
+                 * (both lines are terminated with \n)
+                 * Therefore we do not propagate "last_gets_char" here,
+                 * pass '\n' instead:
+                 */
+                puts_maybe_newline(data, G.nonstdout, last_puts_char, '\n');
                 free(data);
         }
 }
@@ -956,13 +979,13 @@ static void flush_append(char *last_puts_char, char last_gets_char)
 /* Get next line of input from G.input_file_list, flushing append buffer and
  * noting if we ran out of files without a newline on the last line we read.
  */
-static char *get_next_line(char *gets_char, char *last_puts_char, char last_gets_char)
+static char *get_next_line(char *gets_char, char *last_puts_char)
 {
         char *temp = NULL;
         int len;
         char gc;
 
-        flush_append(last_puts_char, last_gets_char);
+        flush_append(last_puts_char);
 
         /* will be returned if last line in the file
          * doesn't end with either '\n' or '\0' */
@@ -1045,7 +1068,7 @@ static void process_files(void)
         int substituted;
 
         /* Prime the pump */
-        next_line = get_next_line(&next_gets_char, &last_puts_char, '\n' /*last_gets_char*/);
+        next_line = get_next_line(&next_gets_char, &last_puts_char);
 
         /* Go through every line in each file */
  again:
@@ -1059,7 +1082,7 @@ static void process_files(void)
 
         /* Read one line in advance so we can act on the last line,
          * the '$' address */
-        next_line = get_next_line(&next_gets_char, &last_puts_char, last_gets_char);
+        next_line = get_next_line(&next_gets_char, &last_puts_char);
         linenum++;
 
         /* For every line, go through all the commands */
@@ -1286,7 +1309,7 @@ static void process_files(void)
                                 free(pattern_space);
                                 pattern_space = next_line;
                                 last_gets_char = next_gets_char;
-                                next_line = get_next_line(&next_gets_char, &last_puts_char, last_gets_char);
+                                next_line = get_next_line(&next_gets_char, &last_puts_char);
                                 substituted = 0;
                                 linenum++;
                                 break;
@@ -1322,7 +1345,7 @@ static void process_files(void)
                         pattern_space[len] = '\n';
                         strcpy(pattern_space + len+1, next_line);
                         last_gets_char = next_gets_char;
-                        next_line = get_next_line(&next_gets_char, &last_puts_char, last_gets_char);
+                        next_line = get_next_line(&next_gets_char, &last_puts_char);
                         linenum++;
                         break;
                 }
@@ -1426,7 +1449,7 @@ static void process_files(void)
 
         /* Delete and such jump here. */
  discard_line:
-        flush_append(&last_puts_char, last_gets_char);
+        flush_append(&last_puts_char /*,last_gets_char*/);
         free(pattern_space);
 
         goto again;
@@ -1512,12 +1535,12 @@ int sed_main(int argc UNUSED_PARAM, char **argv)
         while (opt_f) { // -f
                 char *line;
                 FILE *cmdfile;
-                cmdfile = xfopen_for_read(llist_pop(&opt_f));
+                cmdfile = xfopen_stdin(llist_pop(&opt_f));
                 while ((line = xmalloc_fgetline(cmdfile)) != NULL) {
                         add_cmd(line);
                         free(line);
                 }
-                fclose(cmdfile);
+                fclose_if_not_stdin(cmdfile);
         }
         /* if we didn't get a pattern from -e or -f, use argv[0] */
         if (!(opt & 0x30)) {
diff --git a/editors/vi.c b/editors/vi.c
index 75ce8d396..243676d5f 100644
--- a/editors/vi.c
+++ b/editors/vi.c
@@ -251,7 +251,7 @@ enum {
 // cmds modifying text[]
 // vda: removed "aAiIs" as they switch us into insert mode
 // and remembering input for replay after them makes no sense
-static const char modifying_cmds[] = "cCdDJoOpPrRxX<>~";
+static const char modifying_cmds[] ALIGN1 = "cCdDJoOpPrRxX<>~";
 #endif
 
 enum {