Issue 202 in robotframework-swinglibrary: log4j errors written on standard error

[robotframewor...@googlecode.com]

Status: New
Owner: ----
Labels: Type-Defect Priority-Medium

New issue 202 by geoff.ba...@gmail.com: log4j errors written on standard
error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

In 1.2 I get a load of log4j errors from SwingLibrary when using it on my
real application. These were not present in 1.1.3 or in a snapshot received
one day before the 1.2 release. I presume these were caused by the fix for
issue 200.

log4j:ERROR A "org.robotframework.org.apache.log4j.varia.NullAppender"
object is not assignable to a "org.apache.log4j.Appender" variable.
log4j:ERROR The class "org.apache.log4j.Appender" was loaded by
log4j:ERROR [sun.misc.Launcher$AppClassLoader@6d6f0472] whereas object of
type
log4j:ERROR "org.robotframework.org.apache.log4j.varia.NullAppender"
was loaded by [sun.misc.Launcher$AppClassLoader@6d6f0472].
log4j:ERROR Could not instantiate appender named "A".

I also presume this is caused by some kind of log4j configuration inside
the app, though I don't know the app code well enough to really be sure.

[robotframewor...@googlecode.com]

Updates:
Status: Accepted

Comment #1 on issue 202 by janne.t.harkonen: log4j errors written on
standard error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

I suggest that we remove the log4j configuration that I added in
https://github.com/robotframework/SwingLibrary/commit/e6de63b3c3b96560b361f9400c2169c0391d9581,
thus going back to the original status before issue 200 was "fixed".

Trying to fix this for real is most likely an exercise in frustration.

[robotframewor...@googlecode.com]

Comment #2 on issue 202 by geoff.ba...@gmail.com: log4j errors written on
standard error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

That's OK by me. The warnings are a bit annoying but easy to ignore. The
main thing is that SwingLibrary doesn't screw up log4j configurations for
systems under test.

[robotframewor...@googlecode.com]

Comment #3 on issue 202 by geoff.ba...@gmail.com: log4j errors written on
standard error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

Any chance of a quick 1.2.1 release with this fix? I feel uncomfortable
rolling out tests that spew scary-looking error messages at people.

[robotframewor...@googlecode.com]

Comment #4 on issue 202 by janne.t.harkonen: log4j errors written on
standard error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

I removed the offending file and a snapshot build should be available soon.

I am not sure whether this fix only warrants an official release, though.

[robotframewor...@googlecode.com]

Comment #5 on issue 202 by geoff.ba...@gmail.com: log4j errors written on
standard error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

I would really appreciate an official release.

As far as I can see this bug could mean that log4j logs cannot easily be
enabled in the SUT in SwingLibrary tests, which could be a very serious bug
(it isn't for our immediate situation, but I have to tell *everyone* to
ignore the messages).

We're not only using SwingLibrary but building a product on top of it
(PyUseCase) and would like to make a release ourselves in a week or two. I
would like to depend on an official release of SwingLibrary.

[robotframewor...@googlecode.com]

Updates:
Status: Done
Owner: janne.t....@gmail.com
Labels: Target-1.2.1

Comment #6 on issue 202 by janne.t....@gmail.com: log4j errors written

on standard error
http://code.google.com/p/robotframework-swinglibrary/issues/detail?id=202

Another try in fixing this in
https://github.com/robotframework/SwingLibrary/commit/3647bdf912de7fa4273782e7d6edf935a6367cc4