QDox

QDox is confused by "<" or ">" in field initializers

Details

  • Type: Bug Bug
  • Status: Closed Closed
  • Priority: Blocker Blocker
  • Resolution: Fixed
  • Affects Version/s: 1.6
  • Fix Version/s: 1.6
  • Component/s: None
  • Labels:
    None
  • Number of attachments :
    0

Description

QDox-1.6-SNAPSHOT currently can't parse the following:

package foo;

public class C

{ boolean flag = (X < Y); }

I think this is due to my recent lexer changes to support generics.

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Joe Walnes added a comment -

I've changed the priority of this to blocker as it has introduced a regression failure, causing QDox to fail on 32 of the standard JDK 1.4 classes - some of which have been around since Java 1.0.

Show
Joe Walnes added a comment - I've changed the priority of this to blocker as it has introduced a regression failure, causing QDox to fail on 32 of the standard JDK 1.4 classes - some of which have been around since Java 1.0.
Hide
Permalink
Joe Walnes added a comment -

I ran some regression tests over the entire source tree for JDK 1.4.2 and 1.5.0. The test detemines how many files are successfully parsed:

JDK 1.4.2 - 99.18% (34 failed out of 4137)
JDK 1.5.0 - 98.73% (83 failed out of 6558)

I then disabled the generics support in the 'assignment' state of the lexer and reran the tests:

JDK 1.4.2 - 99.95% (2 failed out of 4137)
JDK 1.5.0 - 99.47% (35 failed out of 6558)

As expected, this gets rid of the regression failure recently introduced with generics support. What I didn't expect to see is that it actually reduced the number of failures in the 1.5 sources as well.

I guess this sorts out the dillema of which bug is deemed to be more critical. I have disabled the generics in field assignment support for the moment to fix this issue and have raised a new issue QDOX-77.

The changes are here: http://cvs.codehaus.org/changelog/qdox/?cs=MAIN:joe:20050103131703

cheers
-Joe

Show
Joe Walnes added a comment - I ran some regression tests over the entire source tree for JDK 1.4.2 and 1.5.0. The test detemines how many files are successfully parsed: JDK 1.4.2 - 99.18% (34 failed out of 4137) JDK 1.5.0 - 98.73% (83 failed out of 6558) I then disabled the generics support in the 'assignment' state of the lexer and reran the tests: JDK 1.4.2 - 99.95% (2 failed out of 4137) JDK 1.5.0 - 99.47% (35 failed out of 6558) As expected, this gets rid of the regression failure recently introduced with generics support. What I didn't expect to see is that it actually reduced the number of failures in the 1.5 sources as well. I guess this sorts out the dillema of which bug is deemed to be more critical. I have disabled the generics in field assignment support for the moment to fix this issue and have raised a new issue QDOX-77 . The changes are here: http://cvs.codehaus.org/changelog/qdox/?cs=MAIN:joe:20050103131703 cheers -Joe

People

  • Assignee:
    Mike Williams
    Reporter:
    Mike Williams
Vote (0)
Watch (1)

Dates

  • Created:
    Updated:
    Resolved: