RE: [Squirrel-sql-users] Difference between 1.1 and 2.1 with Teradata
A Java SQL client for any JDBC compliant database
Brought to you by:
colbell,
gerdwagner
From: Matthew D. <mda...@su...> - 2006-03-22 01:22:12
|
My apologies if this goes through twice. But it seems that I sent the message to squirrel-sql-users-admin rather than squirrel-sql-users when I sent it yesterday. -Matt -----Original Message----- From: Matthew Dahlman [mailto:mda...@su...] Sent: Tuesday, March 21, 2006 7:02 PM To: 'Gerd Wagner'; 'squ...@li...' Subject: RE: [Squirrel-sql-users] Difference between 1.1 and 2.1 with Teradata Hi Gerd (and the rest), Victory! I have found the problem! Well, to more precise: I muddled around with Eclipse, and I have a slightly better understanding of one part of the problem. I created a lame temporary workaround that allows me to see the data (woo hoo!), but it does this at the expense of ignoring metadata (doh!). The Teradata driver was not happy with the ResultSetMetaData.isCurrency(idx) call. And then SQuirreL was not graceful about accepting this reply. (Sorry, my tech vocabulary for describing the problem isn't very good.) So I removed the call to isCurrency from net.sourceforge.squirrel_sql.fw.datasetviewer.ResultSetDataSet.java as shown here starting at line 284: isSigned, // Matt Dahlman: this causes problems with the JDBC driver delivered with Teradata V2R05.00.00.11 // md.isCurrency(idx) // Use hard-coded false instead false ); After making this change I still have errors in my jdbcdebug.log. But, more importantly, I can correctly see the resultset from my query. I am hoping that you can direct my next step. I fear that writing a proper fix to this issue would be well beyond my Java skills. Can you tell from my description above if there is a simple fix to this problem? Regards, Matt PS. Thanks for the directions on using Eclipse. I've never used Eclipse (or any other IDE) before. So it took me quite few hours to understand what was going on. But without your steps listed below couldn't have managed it at all. -----Original Message----- From: Gerd Wagner [mailto:bir...@t-...] Sent: Friday, March 10, 2006 5:57 AM To: Matthew Dahlman Cc: squ...@li... Subject: Re: [Squirrel-sql-users] Difference between 1.1 and 2.1 with Teradata Matthew Dahlman wrote: > Hi Gerd, > > I installed the weekly build > (squirrel-sql-Snapshot-20060226_0137-install.jar) and can confirm that > it has the same problem as 2.1 final. > > I'm not much of a Java programmer. But looking at line 266 I can agree > with you that it's surprising to see the exception there since it > seems to be catching all exceptions at that point. ;-) > > What do I need in order to be able to debug SquirreL? I attempted to > use jdb.exe in place of javaw.exe to run and debug SQuirreL, but I > wasn't successful. I'm not familiar with jdb, so it was just a guess. > If it's possible to explain to me what to do, I'm happy to try. But I > don't know how complicated a process it would be. Normally you use an IDE like Eclipse (http://www.eclipse.org/downloads/) to debug. Then you need SQuirreL's sources that you may either download from our homepage or check out from cvs using the command cvs -z3 -d:pserver:ano...@cv...:/cvsroot/squirrel-sql co -P sql12 Delete the sql12/plugins folder. You won't need plugins to do your debugging Create an Eclipse project with the sql12/ directory select in the second dialog of Eclipse's project wizard. In the project properties on the Libraries tab you need to add the jars in sql12/fw/lib and sql12/app/lib Build the project. Using the Eclipse Navigator open app/src/net/sourceforge/squirrel_sql/client/Main.java in the editor. Right click the editor and choose Debug as Java Application. You set breakpoints by double clicking the left border of the editor. Note, these steps are only a rough explanation of the most important steps. If you aren't used to IDE's it'll probably take you some experimenting. > > Should I report it as a bug so that you have a record of it? Since the > connection works in SQuirreL 1.1 but fails in 2.1, it seems like it > could be considered a regression of some sort. Since it looks like won't be able to reproduce your bug we will probably just close it after a time. So it won't be of much use. > > Regards, > Matt |