[multiple changes]

1999-07-01  Warren Levy  <warrenl@cygnus.com>
	* gnu/gcj/convert/BytesToUnicode.java (read): Changed outlength
	to count and revised comments to match.
	* gnu/gcj/convert/Input_EUCJIS.java (read): Same as Input_8859_1.java.
	* gnu/gcj/convert/Input_JavaSrc.java (read): ditto.
	* gnu/gcj/convert/Input_SJIS.java (read): ditto.
	* gnu/gcj/convert/Input_UTF8.java (read): ditto.
	* gnu/gcj/convert/natInput_EUCJIS.cc (read): ditto.
	* gnu/gcj/convert/natInput_SJIS.cc (read): ditto.
1999-07-01  John-Marc Chandonia  <jmc@cmpharm.ucsf.edu>
	* gnu/gcj/convert/Input_8859_1.java (read): Use 3rd parameter
	properly as count rather than outlength.
	* java/io/BufferedOutputStream.java (write(byte[],int,int): Flush
	output on overflow rather than buffer fill.
	* java/io/BufferedReader.java (fill): Don't clear out the buffer
	if markPos is 0 and there is still room in the buffer.

From-SVN: r27892
This commit is contained in:
Warren Levy 1999-07-01 17:01:00 +00:00
parent b537d12084
commit cb1902adbf
11 changed files with 47 additions and 22 deletions

View file

@ -197,7 +197,7 @@ public class BufferedReader extends Reader
}
}
/* Read more data into the buffer. Update pos and limit appropriatly.
/* Read more data into the buffer. Update pos and limit appropriately.
Assumes pos==limit initially. May invalidate the mark if read too much.
Return number of chars read (never 0), or -1 on eof. */
private int fill() throws IOException
@ -214,7 +214,7 @@ public class BufferedReader extends Reader
if (markPos >= 0 && limit == buffer.length)
markPos = -1;
if (markPos <= 0)
if (markPos < 0)
pos = limit = 0;
int count = in.read(buffer, limit, buffer.length - limit);
if (count > 0)