Merge pull request #3495 from masatake/fix-broken-promise-filter

main,Fypp: stop applying the line filter if there is no more input

Author: masatake

Units/parser-fypp.r/run-alt-guest.d/expected.tags

@@ -7,7 +7,7 @@ F	input.fy	/^	F(1), \/\/ should be 'e', not missing$/;"	e	enum:e	file:	end:22
 debug_code	input.fy	/^#:def debug_code(code)$/;"	m	end:8
 e	input.fy	/^        e() {}$/;"	m	class:e	end:32
 e	input.fy	/^        e(int x) {}$/;"	m	class:e	end:31
-e	input.fy	/^public enum e {$/;"	g
+e	input.fy	/^public enum e {$/;"	g	end:51
 getShape	input.fy	/^	public final Shape getShape() {$/;"	m	class:e	end:46
 getString	input.fy	/^	public String getString() {$/;"	m	class:e	end:41
 repeat_code	input.fy	/^#:def repeat_code(code, repeat)$/;"	m	end:29

Units/parser-fypp.r/segv.d/README

@@ -0,0 +1 @@
+This is a crash test.

Units/parser-fypp.r/segv.d/args.ctags

@@ -0,0 +1 @@
+--extras=+g

Units/parser-fypp.r/segv.d/input.fy

@@ -0,0 +1,2 @@
+
+#:

docs/internal.rst

@@ -494,10 +494,15 @@ Available parser names can be listed by running ctags with
 the 3rd and 5th argument. They are byte offsets of the start and the end of the
 C language area from the beginning of the line which is 0 in this case. In
 general, the guest language's section does not have to start at the beginning of
-the line in which case the two offsets have to be provided. Compilers reading
+the line in which case the two offsets have to be provided. Parsers reading
 the input character by character can obtain the current offset by calling
 ``getInputLineOffset()``.
 
+In some cases, you may want to specifying the offset of the end of
+line (EOL).  A macro ``EOL_CHAR_OFFSET`` defined in ``main/promise.h``
+can be used for specying EOL in abstracted way; you don't have to find
+the real offset for the EOL.
+
 Internal design
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 

main/promise.c

@@ -71,7 +71,7 @@ int  makePromise   (const char *parser,
 	int r;
 	langType lang = LANG_IGNORE;
 
-	verbose("makePromise: %s start(line: %lu, offset: %lu, srcline: %lu), end(line: %lu, offset: %lu)\n",
+	verbose("makePromise: %s start(line: %lu, offset: %ld, srcline: %lu), end(line: %lu, offset: %ld)\n",
 			parser? parser: "*", startLine, startCharOffset, sourceLineOffset,
 			endLine, endCharOffset);
 
@@ -197,7 +197,7 @@ int getLastPromise (void)
 	return promise_count - 1;
 }
 
-static unsigned char* fill_or_skip (unsigned char *input, unsigned char *input_end, bool filling)
+static unsigned char* fill_or_skip (unsigned char *input, const unsigned char *const input_end, const bool filling)
 {
 	if ( !(input < input_end))
 		return NULL;
@@ -221,53 +221,59 @@ static unsigned char* fill_or_skip (unsigned char *input, unsigned char *input_e
 	}
 }
 
-static void line_filler (unsigned char *input, size_t size,
-						 unsigned long startLine, long startCharOffset,
-						 unsigned long endLine, long endCharOffset,
+static void line_filler (unsigned char *input, size_t const size,
+						 unsigned long const startLine, long const startCharOffset,
+						 unsigned long const endLine, long const endCharOffset,
 						 void *data)
 {
-	ulongArray *lines = data;
+	const ulongArray *lines = data;
+	const size_t count = ulongArrayCount (lines);
 	unsigned int start_index, end_index;
 	unsigned int i;
 
-	for (i = 0; i < ulongArrayCount (lines); i++)
+	for (i = 0; i < count; i++)
 	{
-		unsigned long line = ulongArrayItem (lines, i);
+		const unsigned long line = ulongArrayItem (lines, i);
 		if (line >= startLine)
+		{
+			if (line > endLine)
+				return;			/* Not over-wrapping */
 			break;
+		}
 	}
-	if (i == ulongArrayCount (lines))
-		return;
-	if (i > endLine)
-		return;
+	if (i == count)
+		return;					/* Not over-wrapping */
 	start_index = i;
 
-	for (; i < ulongArrayCount (lines); i++)
+	for (; i < count; i++)
 	{
-		unsigned long line = ulongArrayItem (lines, i);
+		const unsigned long line = ulongArrayItem (lines, i);
 		if (line > endLine)
 			break;
 	}
 	end_index = i;
 
 	unsigned long input_line = startLine;
+	const unsigned char *const input_end = input + size;
 	for (i = start_index; i < end_index; i++)
 	{
-		unsigned long line = ulongArrayItem (lines, i);
+		const unsigned long line = ulongArrayItem (lines, i);
 
 		while (1)
 		{
 			if (input_line == line)
 			{
-				input = fill_or_skip (input, input + size, true);
+				input = fill_or_skip (input, input_end, true);
 				input_line++;
 				break;
 			}
 			else
 			{
-				input = fill_or_skip (input, input + size, false);
+				input = fill_or_skip (input, input_end, false);
 				input_line++;
 			}
+			if (input == NULL)
+				return;
 		}
 	}
 }

main/promise.h

@@ -17,6 +17,8 @@
 #include "parse.h"
 #include "numarray.h"
 
+#define EOL_CHAR_OFFSET -1
+
 /* parser can be NULL; give a name with promiseUpdateLanguage()
  * when the name can be determined. */
 int  makePromise   (const char *parser,

main/read.c

@@ -26,6 +26,7 @@
 #include "routines_p.h"
 #include "options_p.h"
 #include "parse_p.h"
+#include "promise.h"
 #include "promise_p.h"
 #include "stats_p.h"
 #include "trace.h"
@@ -1133,7 +1134,17 @@ extern void   pushNarrowedInputStream (
 
 	tmp = getInputFilePositionForLine (endLine);
 	mio_setpos (File.mio, &tmp);
-	mio_seek (File.mio, endCharOffset, SEEK_CUR);
+	if (endCharOffset == EOL_CHAR_OFFSET)
+	{
+		long line_start = mio_tell (File.mio);
+		vString *tmpstr = vStringNew ();
+		readLine (tmpstr, File.mio);
+		endCharOffset = mio_tell (File.mio) - line_start;
+		vStringDelete (tmpstr);
+		Assert (endCharOffset >= 0);
+	}
+	else
+		mio_seek (File.mio, endCharOffset, SEEK_CUR);
 	q = mio_tell (File.mio);
 
 	mio_setpos (File.mio, &original);

parsers/fypp.c

@@ -373,7 +373,7 @@ static void findFyppTags (void)
 	{
 		promise = makePromise (vStringValue(fyppGuestParser),
 							   1, 0,
-							   getInputLineNumber(), 0,
+							   getInputLineNumber(), EOL_CHAR_OFFSET,
 							   0);
 		if (promise >= 0)
 			promiseAttachLineFiller (promise, parseCtx.fypp_lines);