Backport the fixes made in Changeset 24063 to Scala 2.8.1/2

[scabug]

=== What steps will reproduce the problem (please be specific and use wikiformatting)? ===

Please see issue 4067. Scala generated class files cannot be used in Eclipse. 

I'm using Scala 2.9.0 and it works fine. I believe this is because it was fixed in Changeset 24063 

This is a fundamental and blocking issue and undermines project credibility.

=== What is the expected behavior? ===

That Scala generates correct class files.

=== What do you see instead? ===

=== Additional information ===
(for instance, a link to a relevant mailing list discussion)

=== What versions of the following are you using? ===

• Scala:
• Java:
• Operating system:

[scabug]

Imported From: https://issues.scala-lang.org/browse/SI-4259?orig=1
Reporter: bryan hunt (sentimental)

[scabug]

bryan hunt (sentimental) said:
[https://lampsvn.epfl.ch/trac/scala/ticket/4067#comment:40]

[scabug]

bryan hunt (sentimental) said:
Fixes were committed in Changeset 24063, the current release 2.8.1 is unusable in Eclipse workspace.

[scabug]

@SethTisue said:
I won't venture an opinion on whether there should be a 2.8.2. But if there is, I hope this fix will be included, as not only Eclipse, but also ENSIME is affected.

[scabug]

Marko Mikulicic (ithkuil) said:
Replying to [comment:7 SethTisue]:

I won't venture an opinion on whether there should be a 2.8.2. But if there is, I hope this fix will be included, as not only Eclipse, but also ENSIME is affected.

I didn't notice that in ensime yet, how can it be reproduced?

[scabug]

@SethTisue said:
It happens on my large codebase —�I haven't boiled it down to something small and self-contained.

[scabug]

@paulp said:
For the record, the patch for #4067 is not sufficient and the dimensions of the problem aren't 100% clear yet, so nothing is going to happen too imminently.

[scabug]

bryan hunt (sentimental) said:
Erm, excuse my impertinence - but that sounds just a bit lame.

How can the world take a project seriously that breaks binary compatibility with Java (or at least it's tooling (same difference unless you code that banking application in notepad.exe) and then brushes the problem under the carpet....

Net result, everyone that tries out Scala for the first time in the months since 2.8.1 was released finds that it breaks their IDE. So those people go and try something else instead.

I don't think this issue is being handled very professionally.

At least when I reported my last bug ((Some) Scala collections didn't implement Serializable) it made it into the next minor release (2.8.1).... oh yeah, that's the version that doesn't work with Eclipse.

So Scala is project usable, 7 years after project start, provided you run HEAD.

r u guyz seriouz?

[scabug]

@paulp said:
I'm sure this and the screenfuls of similar unhappiness you left in the other ticket are cathartic for you, but I'd appreciate your leaving the bug database free for the actual work. If your goal is to see the matter addressed expeditiously, rest assured I heard you the first time: further exposition of your opinions regarding how we're measuring up to your standards is unlikely to accomplish anything.

[scabug]

bryan hunt (sentimental) said:
extempore:

I quote the last commenter:

"For the record, the patch for #4067 is not sufficient and the dimensions of the problem aren't 100% clear yet, so nothing is going to happen too imminently."

extempore:

So are you just telling me to shut up or does your posting indicate that this issue is being taken seriously?

If you heard me the first time, was that indicated by the totality of your response, "owner changed from scala_meeting to scala_reviewer"?

I also don't think your comment,"further exposition of your opinions regarding how we're measuring up to your standards is unlikely to accomplish anything" - is particularly mature, constructive or professional.

So what is the Scala project doing about this blocking issue?

[scabug]

bryan hunt (sentimental) said:
"If your goal is to see the matter addressed expeditiously, rest assured I heard you the first time"

This problem has been known for 3 months, would you describe that as expeditious?

[scabug]

@soc said:
@sentimental:

I think Paul's point is that the bug tracker is not the right place for this sort of "political" discussion, especially if it doesn't contribute any knowledge or contain code on how to fix this bug better or faster.

The more time Paul has to spend discussing, the less time he can spend on actually fixing your bug.

I almost had to learn it the hard way which of those activities he prefers (and yes, I also have my unfixed pet-peeves, but I have understood that these discussions don't fix bugs).

[scabug]

bryan hunt (sentimental) said:
There's nothing "political" about it, and I wouldn't call it a discussion.

What do you even mean, "political"?????

This bug is three months old, I've spent 2 weeks trying to draw attention to it and find out about it's status.

If it breaks my workspace, it breaks a lot of other people's too.

It's not a pet peeve. If it were a clumsy interface or a method with a fence post error that would qualify as a pet peeve.

Breaking a users workspace because of generating invalid class files is another matter entirely.

So far the only response I've gotten back is has been obfuscation and no clear answers.

It's been fixed, in HEAD. Nobody knows how, nobody knows when, some people think it was fixed in patch #4067, extempore says no. No elaboration. Just no, and forget about it.

So how about some answers rather than just shooting the messenger?

[scabug]

@lindydonna said:
I appreciate that you are concerned about the issue, but rest assured that a fix is in the works. The reason for the delay is that the bug is not as simple as you believe, and Paul is working quite intensely on finding the root cause and then determining the correct solution. This is, in fact, the reason for the terse responses. I'm sure you can sympathize with a developer's heavy workload in the time leading up to a release.

I am sorry that you feel this is a shoot-the-messenger situation, but in fact we are simply suggesting that you post this discussion elsewhere (e.g., scala-user, scala-language, scala-debate). Comments on trac should be limited to information about the bug itself and possible solutions, rather than its importance or lack of attention.

[scabug]

bryan hunt (sentimental) said:
I totally understand, if someone is on it, that's cool - I won't distract him.
I was just concerned that this seemed to be slipping under the radar.

Bryan

[scabug]

@ingoem said:
Assigning to Paul so it doesn't get forgotten.

[scabug]

@soc said:
This seems to be in 2.8.2:

http://www.scala-lang.org/node/10496

Please reopen if the problem persists.