[Emacs-diffs] /srv/bzr/emacs/trunk r111700: * src/lread.c (skip_dyn_bytes): New function.

[Stefan Monnier]

------------------------------------------------------------
revno: 111700
fixes bug: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=12598
committer: Stefan Monnier <address@hidden>
branch nick: trunk
timestamp: Fri 2013-02-08 12:42:09 -0500
message:
  * src/lread.c (skip_dyn_bytes): New function.
  (read1): Use it.  Use getc instead of READCHAR to read bytes.
  (load_each_byte): Remove.  Update users.
modified:
  src/ChangeLog
  src/doc.c
  src/lread.c

=== modified file 'src/ChangeLog'
--- a/src/ChangeLog     2013-02-08 14:44:53 +0000
+++ b/src/ChangeLog     2013-02-08 17:42:09 +0000
@@ -1,3 +1,9 @@
+2013-02-08  Stefan Monnier  <address@hidden>
+
+       * lread.c (skip_dyn_bytes): New function (bug#12598).
+       (read1): Use it.  Use getc instead of READCHAR to read bytes.
+       (load_each_byte): Remove.  Update users.
+
 2013-02-08  Dmitry Antipov  <address@hidden>
 
        * search.c (scan_buffer): Calculate end byte position just once.

=== modified file 'src/doc.c'
--- a/src/doc.c 2013-01-02 16:13:04 +0000
+++ b/src/doc.c 2013-02-08 17:42:09 +0000
@@ -176,9 +176,9 @@
       if (space_left <= 0)
        {
          ptrdiff_t in_buffer = p - get_doc_string_buffer;
-         get_doc_string_buffer =
-           xpalloc (get_doc_string_buffer, &get_doc_string_buffer_size,
-                    16 * 1024, -1, 1);
+         get_doc_string_buffer
+           = xpalloc (get_doc_string_buffer, &get_doc_string_buffer_size,
+                      16 * 1024, -1, 1);
          p = get_doc_string_buffer + in_buffer;
          space_left = (get_doc_string_buffer_size - 1
                        - (p - get_doc_string_buffer));
@@ -279,10 +279,10 @@
   else
     {
       /* The data determines whether the string is multibyte.  */
-      ptrdiff_t nchars =
-       multibyte_chars_in_text (((unsigned char *) get_doc_string_buffer
-                                 + offset),
-                                to - (get_doc_string_buffer + offset));
+      ptrdiff_t nchars
+       = multibyte_chars_in_text (((unsigned char *) get_doc_string_buffer
+                                   + offset),
+                                  to - (get_doc_string_buffer + offset));
       return make_string_from_bytes (get_doc_string_buffer + offset,
                                     nchars,
                                     to - (get_doc_string_buffer + offset));

=== modified file 'src/lread.c'
--- a/src/lread.c       2013-01-17 06:29:40 +0000
+++ b/src/lread.c       2013-02-08 17:42:09 +0000
@@ -96,11 +96,6 @@
    It must be set to nil before all top-level calls to read0.  */
 static Lisp_Object read_objects;
 
-/* True means READCHAR should read bytes one by one (not character)
-   when READCHARFUN is Qget_file_char or Qget_emacs_mule_file_char.
-   This is set by read1 temporarily while handling address@hidden  */
-static bool load_each_byte;
-
 /* List of descriptors now open for Fload.  */
 static Lisp_Object load_descriptor_list;
 
@@ -328,7 +323,7 @@
       return c;
     }
   c = (*readbyte) (-1, readcharfun);
-  if (c < 0 || load_each_byte)
+  if (c < 0)
     return c;
   if (multibyte)
     *multibyte = 1;
@@ -353,6 +348,30 @@
   return STRING_CHAR (buf);
 }
 
+static void
+skip_dyn_bytes (Lisp_Object readcharfun, ptrdiff_t n)
+{
+  if (EQ (readcharfun, Qget_file_char)
+      || EQ (readcharfun, Qget_emacs_mule_file_char))
+    {
+      block_input ();          /* FIXME: Not sure if it's needed.  */
+      fseek (instream, n, SEEK_CUR);
+      unblock_input ();
+    }
+  else
+    { /* We're not reading directly from a file.  In that case, it's difficult
+        to reliably count bytes, since these are usually meant for the file's
+        encoding, whereas we're now typically in the internal encoding.
+        But luckily, skip_dyn_bytes is used to skip over a single
+        dynamic-docstring (or dynamic byte-code) which is always quoted such
+        that \037 is the final char.  */
+      int c;
+      do {
+       c = READCHAR;
+      } while (c >= 0 && c != '\037');
+    }
+}
+
 /* Unread the character C in the way appropriate for the stream READCHARFUN.
    If the stream is a user function, call it with the char as argument.  */
 
@@ -407,14 +426,7 @@
   else if (EQ (readcharfun, Qget_file_char)
           || EQ (readcharfun, Qget_emacs_mule_file_char))
     {
-      if (load_each_byte)
-       {
-         block_input ();
-         ungetc (c, instream);
-         unblock_input ();
-       }
-      else
-       unread_char = c;
+      unread_char = c;
     }
   else
     call1 (readcharfun, make_number (c));
@@ -2388,7 +2400,6 @@
   bool multibyte;
 
   *pch = 0;
-  load_each_byte = 0;
 
  retry:
 
@@ -2598,7 +2609,7 @@
          return tmp;
        }
 
-      /* address@hidden is used to skip NUMBER following characters.
+      /* address@hidden is used to skip NUMBER following bytes.
         That's used in .elc files to skip over doc strings
         and function definitions.  */
       if (c == '@')
@@ -2606,7 +2617,6 @@
          enum { extra = 100 };
          ptrdiff_t i, nskip = 0;
 
-         load_each_byte = 1;
          /* Read a decimal integer.  */
          while ((c = READCHAR) >= 0
                 && c >= '0' && c <= '9')
@@ -2616,8 +2626,15 @@
              nskip *= 10;
              nskip += c - '0';
            }
-         UNREAD (c);
-
+         if (nskip > 0)
+           /* We can't use UNREAD here, because in the code below we side-step
+               READCHAR.  Instead, assume the first char after address@hidden 
occupies
+               a single byte, which is the case normally since it's just
+               a space.  */
+           nskip--;
+         else
+           UNREAD (c);
+           
          if (load_force_doc_strings
              && (EQ (readcharfun, Qget_file_char)
                  || EQ (readcharfun, Qget_emacs_mule_file_char)))
@@ -2659,19 +2676,17 @@
              saved_doc_string_position = file_tell (instream);
 
              /* Copy that many characters into saved_doc_string.  */
+             block_input ();
              for (i = 0; i < nskip && c >= 0; i++)
-               saved_doc_string[i] = c = READCHAR;
+               saved_doc_string[i] = c = getc (instream);
+             unblock_input ();
 
              saved_doc_string_length = i;
            }
          else
-           {
-             /* Skip that many characters.  */
-             for (i = 0; i < nskip && c >= 0; i++)
-               c = READCHAR;
-           }
+           /* Skip that many bytes.  */
+           skip_dyn_bytes (readcharfun, nskip);
 
-         load_each_byte = 0;
          goto retry;
        }
       if (c == '!')