Discussion:
[Jabref-devel] [jabref:discussion] Problem compiling JabRef (missing antlr package)
Oliver Kopp
2015-07-31 18:20:06 UTC
Permalink
Dear François,
I have a checkout of Jabref (current upstream/master) in my Eclipse Luna
workspace (as Gradle project) following the instructions at
https://github.com/JabRef/jabref/wiki/Guidelines-for-setting-up-a-local-workspace.
During the last three days, we did a mayor rewrite of the build system
and began to restructure the code. Therefore, you were middle in the
process.

The next days should be less noisy and things should get more stable.
Nevertheless, we still apply IntelliJs code refactorings to modernize
the code base.
Good to know that there is a Mac OS X developer. We did our best, but
could not really test our Mac OS X builds, maybe you can support us?
:)
1) The steps "./gradlew generateSource" and "./gradlew eclipse" went fine.
But in the next step when I ran "./gradlew assemble" I get an error during
This also happens when switching branches dev_2.11 and master back and
forth. `gradlew clean` does not remove `src/main/gen`. This must be
done manually when switching branches.

I assume that you started at a frech checkout. I would suggest to pull
the master branch again and retry to get the thing up and running.
I assume that Gradle should normally manage that antlr compile dependency
correctly.
Yes, it does.
Any idea what causes this and how I can circumvent it?
I think, the issue resolves itself if you pull and remove `src/main/gen`.
2) I had another problem when trying to run JabRef from Eclipse (via
Exception in thread "main" java.lang.NoClassDefFoundError: AboutEvent
This is strange because this class is natively provided by JDK8 on Mac via
com.apple.eawt.AppEvent (in rt.jar) ...
This is interesting. In the master branch, we included MacAdapter.java
in the direct source. This was made possible by including a JAR
providing these classes on Linux/Windows. Maybe this resolves the
issue, too or prevents JabRef from running at all. Maybe you can test?
BTW the reason I am doing this is to investigate a bug in the latest 2.11
beta version that was reported by a research team I am associated with, and
I thought I could perhaps fix it and offer you a patch.
That would be great. Then, I would recommend you to stop working on
the master branch and switch to the `dev_2.11`. We will try to port
the fixes to the main branch!
I first wanted to check whether it can still be reproduced with the latest development
version.
The bad thing would be if it happens on the dev_2.11 branch, but does
not happen on the master branch. Let's see :)

Hope, that helps and please keep me updated :)

Cheers,

Oliver

------------------------------------------------------------------------------
FC
2015-08-01 09:06:26 UTC
Permalink
Dear Oliver,

This is to confirm that you were right on all counts. After syncing again
with upstream (and removing src/main/gen) I was able to compile both the
master and the dev_2.11 branches successfully.

However I had 16 failing unit tests (on dev_2.11), but surely you are aware
of these via Travis.

And I still have the same issue with AppEvent on Mac when trying to run
JabRefMain from Eclipse. (I can live with that for the moment, but if I can
be of any help to investigate this, please tell me.)

The good news is that I could run the generated distribution (after
untarring it and running the shell script), even though JabRef complains
about a missing MacAdapter:

Exception (class java.lang.ClassNotFoundException):
osx.macadapter.MacAdapter

Probably this is specifically added to the Mac-only version, which is not
in the generic distribution?

I could also still reproduce the bug I mentioned in my original mail.
Apparently it also exists in 2.10:
see https://sourceforge.net/p/jabref/bugs/1243/ (with several duplicates).
This happens on MacOSX generally and also on Windows under certain
circumstances.
I will probably come back to you off-list about this very soon.

Regards,
François
Dear François,
I have a checkout of Jabref (current upstream/master) in my Eclipse Luna
workspace (as Gradle project) following the instructions at
https://github.com/JabRef/jabref/wiki/Guidelines-for-setting-up-a-local-workspace
.
During the last three days, we did a mayor rewrite of the build system
and began to restructure the code. Therefore, you were middle in the
process.
The next days should be less noisy and things should get more stable.
Nevertheless, we still apply IntelliJs code refactorings to modernize
the code base.
Now I am trying to compile it on Mac OS X but I ran into two
Good to know that there is a Mac OS X developer. We did our best, but
could not really test our Mac OS X builds, maybe you can support us?
:)
1) The steps "./gradlew generateSource" and "./gradlew eclipse" went
fine.
But in the next step when I ran "./gradlew assemble" I get an error
during
This also happens when switching branches dev_2.11 and master back and
forth. `gradlew clean` does not remove `src/main/gen`. This must be
done manually when switching branches.
I assume that you started at a frech checkout. I would suggest to pull
the master branch again and retry to get the thing up and running.
I assume that Gradle should normally manage that antlr compile dependency
correctly.
Yes, it does.
Any idea what causes this and how I can circumvent it?
I think, the issue resolves itself if you pull and remove `src/main/gen`.
2) I had another problem when trying to run JabRef from Eclipse (via
Exception in thread "main" java.lang.NoClassDefFoundError: AboutEvent
This is strange because this class is natively provided by JDK8 on Mac
via
com.apple.eawt.AppEvent (in rt.jar) ...
This is interesting. In the master branch, we included MacAdapter.java
in the direct source. This was made possible by including a JAR
providing these classes on Linux/Windows. Maybe this resolves the
issue, too or prevents JabRef from running at all. Maybe you can test?
BTW the reason I am doing this is to investigate a bug in the latest 2.11
beta version that was reported by a research team I am associated with,
and
I thought I could perhaps fix it and offer you a patch.
That would be great. Then, I would recommend you to stop working on
the master branch and switch to the `dev_2.11`. We will try to port
the fixes to the main branch!
I first wanted to check whether it can still be reproduced with the
latest development
version.
The bad thing would be if it happens on the dev_2.11 branch, but does
not happen on the master branch. Let's see :)
Hope, that helps and please keep me updated :)
Cheers,
Oliver
Loading...