GRECLIPSE

Class expressions are problematic for inferencing

Details

  • Type: Bug Bug
  • Status: Resolved Resolved
  • Priority: Major Major
  • Resolution: Fixed
  • Affects Version/s: 2.0.0m2
  • Fix Version/s: 2.0.0RC1
  • Component/s: None
  • Labels:
    None
  • Number of attachments :
    0

Description

Playing around with code like:

GroovyClass.class.name
GroovyClass.getClass().name
GroovyClass.name
GroovyClass.getName()

All of this should be detected by the inferencing engine, but it seems like it is not. Should work on this.

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Andrew Eisenberg added a comment -

My mistake...these two should in fact be unknown references (unless of Course GroovyClass has a 'name' property):

GroovyClass.name
GroovyClass.getName()

The first two were not working, but I belive that I have fixed it. Needs some more work, though before I can commit.

Show
Andrew Eisenberg added a comment - My mistake...these two should in fact be unknown references (unless of Course GroovyClass has a 'name' property): GroovyClass.name GroovyClass.getName() The first two were not working, but I belive that I have fixed it. Needs some more work, though before I can commit.
Hide
Permalink
Andrew Eisenberg added a comment -

Arrrgh. From within the AST, there is no way to determine whether the ASTNode is 'GroovyClass' or 'GroovyClass.class'. In other words, I can't tell if the type of a selected expression should be the GroovyClass static context, or the Class type.

Show
Andrew Eisenberg added a comment - Arrrgh. From within the AST, there is no way to determine whether the ASTNode is 'GroovyClass' or 'GroovyClass.class'. In other words, I can't tell if the type of a selected expression should be the GroovyClass static context, or the Class type.
Hide
Permalink
Andrew Eisenberg added a comment -

This is a bit crude, but I can compare the number of characters in the type name to the number of characters in the text. If the difference is ".class".length(), then it is likely that there this is a reference to the Class class.

Of course this will fail if there are spaces in the text.

Show
Andrew Eisenberg added a comment - This is a bit crude, but I can compare the number of characters in the type name to the number of characters in the text. If the difference is ".class".length(), then it is likely that there this is a reference to the Class class. Of course this will fail if there are spaces in the text.
Hide
Permalink
Andrew Eisenberg added a comment -

Fixed with tests.

Required a bit of trickery. In the end, to address comment 3, I am looking at the underlying text to see if the last part is '.class'. There will still be a problem if there are spaces between '.' and 'class', but good enough for now. There is a FIXADE in the code.

Also, I had to explicitly add properties to the Class ClassNode so that things like 'cannonicalName' can be recognized as a property of 'getCanonicalName()'.

Show
Andrew Eisenberg added a comment - Fixed with tests. Required a bit of trickery. In the end, to address comment 3, I am looking at the underlying text to see if the last part is '.class'. There will still be a problem if there are spaces between '.' and 'class', but good enough for now. There is a FIXADE in the code. Also, I had to explicitly add properties to the Class ClassNode so that things like 'cannonicalName' can be recognized as a property of 'getCanonicalName()'.

People

  • Assignee:
    Andrew Eisenberg
    Reporter:
    Andrew Eisenberg
Vote (0)
Watch (0)

Dates

  • Created:
    Updated:
    Resolved: