Fix the extraction of tagged comments.

10 files changed, 69 insertions, 20 deletions

diff --git a/gettext-tools/src/ChangeLog b/gettext-tools/src/ChangeLog
index 60ab037..8f60594 100644
--- a/gettext-tools/src/ChangeLog
+++ b/gettext-tools/src/ChangeLog
@@ -1,3 +1,20 @@
+2003-11-22  Bruno Haible  <bruno@clisp.org>
+
+	* x-c.c (phase4_getc): Skip leading whitespace in C++ style comments
+	as well.
+	* x-sh.c (read_word): Skip leading whitespace in comments.
+	* x-lisp.c (read_object): Skip leading whitespace in single-line
+	comments as well.
+	* x-elisp.c (read_object): Skip leading whitespace in comments.
+	* x-librep.c (read_object): Skip leading whitespace in single-line
+	comments as well.
+	* x-java.c (phase4_getc): Skip leading whitespace in C++ style comments
+	as well.
+	* x-awk.c (phase2_getc): Skip leading whitespace in comments.
+	* x-ycp.c (phase2_getc): Skip leading whitespace in single-line
+	comments as well.
+	* x-tcl.c (read_command): Skip leading whitespace in comments.
+
 2003-11-24  Bruno Haible  <bruno@clisp.org>
 
 	* format-lisp.c (check_params): Use ngettext for one of the messages.
diff --git a/gettext-tools/src/x-awk.c b/gettext-tools/src/x-awk.c
index 933daed..a902340 100644
--- a/gettext-tools/src/x-awk.c
+++ b/gettext-tools/src/x-awk.c
@@ -187,12 +187,16 @@ phase2_getc ()
 	  c = phase1_getc ();
 	  if (c == '\n' || c == EOF)
 	    break;
-	  if (buflen >= bufmax)
+	  /* We skip all leading white space, but not EOLs.  */
+	  if (!(buflen == 0 && (c == ' ' || c == '\t')))
 	    {
-	      bufmax = 2 * bufmax + 10;
-	      buffer = xrealloc (buffer, bufmax);
+	      if (buflen >= bufmax)
+		{
+		  bufmax = 2 * bufmax + 10;
+		  buffer = xrealloc (buffer, bufmax);
+		}
+	      buffer[buflen++] = c;
 	    }
-	  buffer[buflen++] = c;
 	}
       if (buflen >= bufmax)
 	{
diff --git a/gettext-tools/src/x-c.c b/gettext-tools/src/x-c.c
index cfff2d3..d38bc6a 100644
--- a/gettext-tools/src/x-c.c
+++ b/gettext-tools/src/x-c.c
@@ -662,7 +662,9 @@ phase4_getc ()
 	  c = phase3_getc ();
 	  if (c == '\n' || c == EOF)
 	    break;
-	  comment_add (c);
+	  /* We skip all leading white space, but not EOLs.  */
+	  if (!(buflen == 0 && (c == ' ' || c == '\t')))
+	    comment_add (c);
 	}
       comment_line_end (0);
       last_comment_line = newline_count;
diff --git a/gettext-tools/src/x-elisp.c b/gettext-tools/src/x-elisp.c
index d1ecabc..0325163 100644
--- a/gettext-tools/src/x-elisp.c
+++ b/gettext-tools/src/x-elisp.c
@@ -864,7 +864,11 @@ read_object (struct object *op, bool first_in_list, bool new_backquote_flag,
 		if (c != ';')
 		  all_semicolons = false;
 		if (!all_semicolons)
-		  comment_add (c);
+		  {
+		    /* We skip all leading white space, but not EOLs.  */
+		    if (!(buflen == 0 && (c == ' ' || c == '\t')))
+		      comment_add (c);
+		  }
 	      }
 	    comment_line_end (0);
 	    continue;
diff --git a/gettext-tools/src/x-java.c b/gettext-tools/src/x-java.c
index eb38a84..04a61c3 100644
--- a/gettext-tools/src/x-java.c
+++ b/gettext-tools/src/x-java.c
@@ -766,7 +766,9 @@ phase4_getc ()
 	  c = phase3_getc ();
 	  if (RED (c) == '\n' || c == P2_EOF)
 	    break;
-	  comment_add (c);
+	  /* We skip all leading white space, but not EOLs.  */
+	  if (!(comment_at_start () && (RED (c) == ' ' || RED (c) == '\t')))
+	    comment_add (c);
 	}
       phase3_ungetc (c); /* push back the newline, to decrement line_number */
       comment_line_end (0);
diff --git a/gettext-tools/src/x-librep.c b/gettext-tools/src/x-librep.c
index 106b808..afa99fc 100644
--- a/gettext-tools/src/x-librep.c
+++ b/gettext-tools/src/x-librep.c
@@ -800,7 +800,11 @@ read_object (struct object *op, flag_context_ty outer_context)
 		if (c != ';')
 		  all_semicolons = false;
 		if (!all_semicolons)
-		  comment_add (c);
+		  {
+		    /* We skip all leading white space, but not EOLs.  */
+		    if (!(buflen == 0 && (c == ' ' || c == '\t')))
+		      comment_add (c);
+		  }
 	      }
 	    comment_line_end (0);
 	    continue;
diff --git a/gettext-tools/src/x-lisp.c b/gettext-tools/src/x-lisp.c
index 0c687ee..c3c8a55 100644
--- a/gettext-tools/src/x-lisp.c
+++ b/gettext-tools/src/x-lisp.c
@@ -1174,7 +1174,11 @@ read_object (struct object *op, flag_context_ty outer_context)
 		    if (c != ';')
 		      all_semicolons = false;
 		    if (!all_semicolons)
-		      comment_add (c);
+		      {
+			/* We skip all leading white space, but not EOLs.  */
+			if (!(buflen == 0 && (c == ' ' || c == '\t')))
+			  comment_add (c);
+		      }
 		  }
 		comment_line_end (0);
 		continue;
diff --git a/gettext-tools/src/x-sh.c b/gettext-tools/src/x-sh.c
index 2d8afa7..de63406 100644
--- a/gettext-tools/src/x-sh.c
+++ b/gettext-tools/src/x-sh.c
@@ -714,7 +714,9 @@ read_word (struct word *wp, int looking_for, flag_context_ty context)
 	      c = phase1_getc ();
 	      if (c == EOF || c == '\n')
 		break;
-	      comment_add (c);
+	      /* We skip all leading white space, but not EOLs.  */
+	      if (!(buflen == 0 && (c == ' ' || c == '\t')))
+		comment_add (c);
 	    }
 	  comment_line_end ();
 	}
diff --git a/gettext-tools/src/x-tcl.c b/gettext-tools/src/x-tcl.c
index bfad460..f35143a 100644
--- a/gettext-tools/src/x-tcl.c
+++ b/gettext-tools/src/x-tcl.c
@@ -830,7 +830,9 @@ read_command (int looking_for, flag_context_ty outer_context)
 	      c = phase2_getc ();
 	      if (c == EOF || c == CL_BRACE || c == '\n')
 		break;
-	      comment_add (c);
+	      /* We skip all leading white space, but not EOLs.  */
+	      if (!(buflen == 0 && (c == ' ' || c == '\t')))
+		comment_add (c);
 	    }
 	  comment_line_end ();
 	  continue;
diff --git a/gettext-tools/src/x-ycp.c b/gettext-tools/src/x-ycp.c
index 80a8cfd..e13ed18 100644
--- a/gettext-tools/src/x-ycp.c
+++ b/gettext-tools/src/x-ycp.c
@@ -159,12 +159,16 @@ phase2_getc ()
 	      c = phase1_getc ();
 	      if (c == '\n' || c == EOF)
 		break;
-	      if (buflen >= bufmax)
+	      /* We skip all leading white space, but not EOLs.  */
+	      if (!(buflen == 0 && (c == ' ' || c == '\t')))
 		{
-		  bufmax = 2 * bufmax + 10;
-		  buffer = xrealloc (buffer, bufmax);
+		  if (buflen >= bufmax)
+		    {
+		      bufmax = 2 * bufmax + 10;
+		      buffer = xrealloc (buffer, bufmax);
+		    }
+		  buffer[buflen++] = c;
 		}
-	      buffer[buflen++] = c;
 	    }
 	  if (buflen >= bufmax)
 	    {
@@ -201,7 +205,7 @@ phase2_getc ()
 	      if (c == EOF)
 		break;
 	      /* We skip all leading white space, but not EOLs.  */
-	      if (buflen == 0 && isspace (c) && c != '\n')
+	      if (buflen == 0 && (c == ' ' || c == '\t'))
 		continue;
 	      if (buflen >= bufmax)
 		{
@@ -260,12 +264,16 @@ phase2_getc ()
 	      c = phase1_getc ();
 	      if (c == '\n' || c == EOF)
 		break;
-	      if (buflen >= bufmax)
+	      /* We skip all leading white space, but not EOLs.  */
+	      if (!(buflen == 0 && (c == ' ' || c == '\t')))
 		{
-		  bufmax = 2 * bufmax + 10;
-		  buffer = xrealloc (buffer, bufmax);
+		  if (buflen >= bufmax)
+		    {
+		      bufmax = 2 * bufmax + 10;
+		      buffer = xrealloc (buffer, bufmax);
+		    }
+		  buffer[buflen++] = c;
 		}
-	      buffer[buflen++] = c;
 	    }
 	  if (buflen >= bufmax)
 	    {