On Sat, Dec 13, 2014 at 2:25 AM, Alan Bateman <[email protected]> wrote: > On 12/12/2014 19:53, roger riggs wrote: >> >> Please review this fix for bad link in javadoc: >> >> jdk/src/java.base/share/classes/java/lang/ProcessBuilder.java >> @@ -748,11 +748,11 @@ >> * Process#getErrorStream()}. >> * If the destination is set to any other value, then >> * {@link Process#getErrorStream()} will return a >> * <a href="#redirect-output">null input stream</a>. >> * >> - * <p>If the {@link #redirectErrorStream redirectErrorStream} >> + * <p>If the {@link #redirectErrorStream() redirectErrorStream} >> * attribute has been set {@code true}, then the redirection set >> * by this method has no effect. >> * >> * @param destination the new standard error destination >> * @return this process builder > > This looks good to me, it always amazes me that we continue to find broken > links.
In this particular case I'm pretty sure I intentionally relied on the historic behavior of javadoc which generated a link to one of the overloads (poor judgment on my part - I, for one, welcome our new overload-choice-refusing javadoc). It always amazes me that we cannot get procedures in place so that docs are guaranteed dead-link-clean.
