Bug 319979 - [1.5][compiler] Make optional warning from compiler error for generics clash bug 289247
Summary: [1.5][compiler] Make optional warning from compiler error for generics clash ...
Status: VERIFIED WONTFIX
Alias: None
Product: JDT
Classification: Eclipse Project
Component: Core (show other bugs)
Version: 3.6   Edit
Hardware: PC All
: P3 enhancement (vote)
Target Milestone: 3.7 M1   Edit
Assignee: Srikanth Sankaran CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-07-15 08:45 EDT by Mark Hoffmann CLA
Modified: 2011-01-25 12:35 EST (History)
3 users (show)

See Also:
  • 6182950


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Mark Hoffmann CLA 2010-07-15 08:45:25 EDT
Build Identifier: 201006171415

I refer to Bug 289247.

I work on big projects. We use all use Eclipse as IDE. With Helios we found a lot of compile errors in our projects. We will not switch to Java 1.7 in the next years and we can't get all the generics problems fixted soon. So we cant use Helios.

What about make the compile error from sun bug 6182950 an optional error in the preferences? So you can make a warning from it? You could provide it in the Helios SR1.

I agree, that the JDT works correct. But this is a really annoying situation for many Java users.

Regards,
Mark

Reproducible: Always
Comment 1 Olivier Thomann CLA 2010-07-15 11:11:11 EDT
I understand your problem, but if the fix is backported to a JDK6 you will have the same problem with javac.
Srikanth, I let you decide on this one. I don't like to introduce options to reproduce bugs.
Comment 2 Srikanth Sankaran CLA 2010-07-19 00:48:59 EDT
(In reply to comment #1)
> I understand your problem, but if the fix is backported to a JDK6 you will have
> the same problem with javac.
> Srikanth, I let you decide on this one. I don't like to introduce options to
> reproduce bugs.

My sentiments are along the same lines. I think the problem is better
addressed by adjusting the source code, even if that route is painful
currently (sorry), it is an investment that needs to be made anyways
at some point in future.
Comment 3 Mark Hoffmann CLA 2010-07-19 05:09:26 EDT
Thank you for your comments.
After some thinking this weekend, I the optinion that you're right.
So I take back this feature request.

But can you tell me how complex it is, to create a own patched version of the JDT? I would patch the code and could provide it in our company. But I don't know where I have to look for the right code position.

Thank you,
Mark
Comment 4 Srikanth Sankaran CLA 2010-07-19 05:25:39 EDT
> But can you tell me how complex it is, to create a own patched version of the
> JDT? I would patch the code and could provide it in our company. But I don't
> know where I have to look for the right code position.

You can start with taking a look at the source & test patches attached to
bug #289247. You will want to also test your changes : org.eclipse.jdt.core.tests.RunCompilerTests is the entry
point for bulk of compiler related tests. You should also run org.eclipse.jdt.core.tests.RunBuilderTests, org.eclipse.jdt.core.tests.RunDOMTests,org.eclipse.jdt.core.tests.RunFormatterTests, org.eclipse.jdt.core.tests.RunModelTests etc.
Comment 5 Dani Megert CLA 2010-07-19 07:05:08 EDT
Mark, patching the JDT is expensive and with each new version you have to do this again. I don't know how you ended up with lots of such errors (generated code) but it's definitely better to fix them earlier than later.
Comment 6 Olivier Thomann CLA 2011-01-25 12:35:20 EST
Verified.