This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.

Bug 176048 - Can't debug Rails apps with V3 b70
Summary: Can't debug Rails apps with V3 b70
Status: RESOLVED FIXED
Alias: None
Product: serverplugins
Classification: Unclassified
Component: GlassFish (show other bugs)
Version: 6.x
Hardware: All All
: P2 blocker (vote)
Assignee: Vince Kraemer
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-11-05 14:02 UTC by Erno Mononen
Modified: 2009-11-11 13:59 UTC (History)
1 user (show)

See Also:
Issue Type: DEFECT
Exception Reporter:


Attachments
messages.log (46.51 KB, text/plain)
2009-11-05 14:04 UTC, Erno Mononen
Details
server.log (147.70 KB, text/plain)
2009-11-05 14:07 UTC, Erno Mononen
Details
a hack of a patch... (4.67 KB, application/octet-stream)
2009-11-08 18:34 UTC, Vince Kraemer
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Erno Mononen 2009-11-05 14:02:42 UTC
Steps:
1. create a rails app with JRuby 1.4.0, Rails 2.3.4, V3 build 70
2. try to debug the app => the server doesn't start up

I'm attaching the IDE log and server log files.
Comment 1 Erno Mononen 2009-11-05 14:04:54 UTC
Created attachment 90515 [details]
messages.log
Comment 2 Erno Mononen 2009-11-05 14:07:33 UTC
Created attachment 90516 [details]
server.log
Comment 3 Vince Kraemer 2009-11-06 05:04:01 UTC
yes. the run-time code was changed.  the code that supported our debug support depended on was deleted at some point... in the last few months.

I will get a fix ready.. though it will most likely be in the glassfish-scripting project... 

I will leave this issue open to track the status of the whole mess.
Comment 4 Vince Kraemer 2009-11-08 18:34:20 UTC
Created attachment 90563 [details]
a hack of a patch...

Erno... please give me some feedback on this.  Is this patch headed in the wrong direction?
Comment 5 Erno Mononen 2009-11-09 01:32:37 UTC
So does this mean debugging is not going work with the "real" V3, but only with the v3 gem? Can't the approach used for prelude, i.e. just running part of the server stack, work for V3 FCS too? IMO that approach would be preferable over installing the V3 gem (which can fail for various reasons) when the user hits debug.
Comment 6 Vince Kraemer 2009-11-09 07:45:07 UTC
I agree... but there are a couple things that have conspired against that working.

There was a major rewrite of the glassfish-scripting modules... When that rewrite happened, it sounds like PCW had told them that they did not have to keep that code in sound working order... so it has fallen into disrepair:
   random npes and the like.

The code in GF scripting has a major code freeze today, so it is unlikely that I can resolve all the issues that I am encountering before midnight tonight.
Comment 7 Vince Kraemer 2009-11-10 09:13:59 UTC
I was not able to resolve all the various issues related to trying to replicate the 'prelude solution' for v3 before the code freeze.

Erno: did you have other comments, etc. concerning the change... or should I commit something along these lines...
Comment 8 Erno Mononen 2009-11-10 12:32:21 UTC
No other comments really - as I wrote I don't like the solution much, but given the current situation I'm not sure we have much choices here. Perhaps there should be a dialog informing the user what's going on (i.e. explaining that debugging on V3 doesn't work, need to install the GF gem) when he runs debug for the first time - thoughts?
Comment 9 Vince Kraemer 2009-11-10 13:01:13 UTC
I think we need to open I dialog to explain what is about to happen in the case of 'first debug'...  I hadn't done that in the patch, since I was looking for feedback on gotcha's I might not have encountered.

I will add an appropriate dialog to cover the situation.

One question:

If I ask the user "do you want to download the gem?" and they answer "no"... how would I stop the debug command from completing?  It looks like there is no catch between ensureRunning and the event loop, but I might have missed one.  It looks like returning true won't work... returning false doesn't seem to mean cancel the user action either.
Comment 10 Vince Kraemer 2009-11-11 13:59:09 UTC
http://hg.netbeans.org/web-main/rev/c326911fca0d