126686 - Sun's javadoc tool generates invalid HTML

Status: RESOLVED FIXED

(Web Compatibility :: Site Reports, defect)

Description

[Steve Chapel]

The javadoc tool that comes with J2SDK 1.4.0 produces invalid HTML output. The
invalid HTML is generally easy to fix: Use the frameset DTD instead of the
transitional DTD on frame documents, put JavaScript code in the head instead of
between the head and body, and so on.

There are two ways to see the invalid HTML:
1. Go to the above URL to see the javadoc output for the classes in J2SDK 1.4.0.
2. Download the J2SDK 1.4.0 from http://java.sun.com/j2se/1.4/download.html and
run javadoc on any Java source code.

This tool produces thousands if not millions of web pages on the internet, all
with invalid HTML.

Comment 1

[Bob Clary [:bc] (inactive)]

confirmed and contacted with details. Follow up in the summer about the next
point release.

Comment 2

[Daniel Jensen]

1.4.1 Beta is out- I don't know whether this is fixed in it or not, but javadoc
appears to have had a fair number of bug fixes.

Comment 3

[Steve Chapel]

No, this is not fixed with version 1.4.1. I tested the API docs on Sun's
website, and also using javadoc on my own computer. Both javadoc-generated sites
still have tons of validation problems. See also bugs 4275630, 4418115, and
4415270 in the Java Bug Parade at http://developer.java.sun.com/developer/bugParade/

Comment 4

[Bob Clary [:bc] (inactive)]

tech evang june 2003 reorg

Comment 5

[Bob Clary [:bc] (inactive)]

future

Comment 6

[Daniel Jensen]

Sun claims to have fixed this in 1.4.2:
http://java.sun.com/j2se/1.4.2/docs/tooldocs/javadoc/whatsnew-1.4.2.html#invalidhtml

Comment 7

[Steve Chapel]

The javadoc tool that comes with JDK 1.4.2 certainly is far, far better than
previous versions. All the problems with wrong DOCTYPES and nearly all serious
structural problems were fixed. Even the CSS validates and looks good.

However, as is, none of the HTML pages generated by javadoc validate. They're
missing charset information, and the W3C validator complains "I was not able to
extract a character encoding labeling from any of the valid sources for such
information. Without encoding information it is impossible to validate the
document."

After adding the tag
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=ISO-8859-1">
in the head of each of the HTML files, many of them do validate now. I'm setting
the URL for this bug to one of the many pages that still don't validate even
after specifying the charset. I'm also reducing severity of this bug to normal
because now the errors are not so serious.

Comment 8

[Steve Chapel]

Here are some Java Bug Database Ids for the remaining problems:
4663254  invalid characters allowed in named anchors
4753019  missing character encodings
4753126  various structural problems
4964455  various structural problems

Comment 9

[Steve Chapel]

It looks like none of the bugs in comment #8 have been fixed yet. I'm updating the URL to one used in Java 1.6.

Comment 10

[Hallvord R. M. Steen [:hallvors]]

http://bugs.java.com/bugdatabase/view_bug.do?bug_id=4663254 - RESOLVED
http://bugs.java.com/bugdatabase/view_bug.do?bug_id=4753126 - CLOSED
http://bugs.java.com/bugdatabase/view_bug.do?bug_id=4753019 - RESOLVED
http://bugs.java.com/bugdatabase/view_bug.do?bug_id=4964455 - CLOSED

So at least according to the state of the bugs in the system it's been taken care of.