Open GoogleCodeExporter opened 9 years ago
In an attempt to port over the jcatapult implementation I'm still getting a null
pointer at:
com.google.inject.internal.ProviderMethodsModule on line 60
Original comment by a.rugg...@gmail.com
on 27 Nov 2008 at 8:28
If you could provide a test case, that'll make it possible to fix this.
Original comment by limpbizkit
on 27 Nov 2008 at 9:47
I'll see if I can work something up.
Original comment by a.rugg...@gmail.com
on 28 Nov 2008 at 5:07
After further investigation this is not a guice bug. I can't tell if the issue
is in
Struts or XWork. Basically what's happening is the object factory is getting
initialized and then the buildBean is called for the type converters before the
configuration settings are set (like setModule). So the user specified module
will
not be set when the first buildBean is called.
I do not understand why this affects the Guice plugin since it initializes
lazily.
This does not affect spring, jcatpult since they do not get their configuration
from
the struts constants.
Original comment by a.rugg...@gmail.com
on 28 Nov 2008 at 7:03
Original comment by limpbizkit
on 28 Nov 2008 at 7:35
Can anyone provide a link to the original bug of the causing project (struts or
xwork), which may provide a patch. This bug made me abandon Guice from struts
project.
Original comment by bahri.ge...@gmail.com
on 10 Feb 2010 at 12:58
http://issues.apache.org/jira/browse/WW-2898
Original comment by a.rugg...@gmail.com
on 10 Feb 2010 at 3:30
Issue 611 has been merged into this issue.
Original comment by sberlin
on 5 Mar 2011 at 2:50
Since this issue is marked as "Invalid" and the Struts2 issue that you have
linked to is marked as "Wont Fix", perhaps you guys could add some
documentation to the Struts2 plug-in indicating that the use of custom type
converters causes an issue.
Original comment by steven.b...@gmail.com
on 5 Mar 2011 at 6:47
Issue 619 has been merged into this issue.
Original comment by sberlin
on 22 Mar 2011 at 10:20
Here's a little patch that solves the issue in my environment. I might be
missing something, but it works for me.
It seems like the Struts2Factory could be refactored so that it doesn't rely on
a particular order of method calls like it does now. If struts calls
buildInterceptor() before it calls buildBean() then this factory fails. That
seems like a very tight coupling to an undocumented and internal struts
behavior.
I might tackle a refactoring once my current project is over.
Original comment by jeffgbut...@gmail.com
on 23 Mar 2011 at 1:10
Attachments:
This was marked as 'Invalid' on 2008, 2.5 years later a patch appears...
perhaps reopening this issue could be explored?
Original comment by voytek.j...@gmail.com
on 11 Jan 2012 at 8:38
That patch seems to work. If nobody objects, could it be included in a future
release? Despite the issue being "invalid".
Original comment by daniel.b...@gmail.com
on 8 Jan 2013 at 2:59
Original issue reported on code.google.com by
a.rugg...@gmail.com
on 27 Nov 2008 at 8:23