Compiling releases with debugging on

[Mike Gratton <mi...@vee.net>]

Guys,

I'd suggest that it would be worthwhile compiling releases with 
debugging symbols turned on. This will make the compiled classes (only 
slightly?) larger, but should not have an impact on perfomance. 
Obviously, this buys us useable stack traces when people report 
problems, instead of having "(Unknown Source)" reported for the line 
number of a stack frame, which is pretty useless. If anyone really does 
not want debugging on, they can always recompile it with this off.

The patch enclosed turns debugging support on by default, but allows 
this to be overridden as needed. I've also included a few other 
tuneables as well, which all default to off for now: optimization, 
nowarn, deprecation, verbose.

To change these tuneables at compile time, invoke ant with a 
"-D[propertyname]={on|off}" argument. For example, to turn debugging 
off, and invoke the "clean" and "compile" targets, you'd do:

   ./bin/ant -Dcompile.debug=off clean compile

I've got a much more convenient, and persistant mechanism for setting 
Ant tuneables on a per-developer basis. I'll post the patch after this 
gets comitted. I'll also write up some docs on this as well.

Mike.

-- 
Mike Gratton <mi...@vee.net>, <http://web.vee.net/>
"Every motive escalate."

Re: Compiling releases with debugging on

[Kimbro Staken <ks...@xmldatabases.org>]

On Friday, March 8, 2002, at 04:01 AM, Mike Gratton wrote:

>
> I haven't really heard anything back about this, what's the verdict guys?
>  I still think it would be a mistake to make a release without Xindice 
> classes compiled with debugging on, and given other ASF do so (Xalan at 
> least, Xerces IIRC), I can't think of any reason not to.
>

James said he was going to commit it along with your other patch that he 
was working on. I've been kind of out things for over a week but I don't 
remember seeing the commit actually happen. James are you still planning 
to commit this?

>
> Mike Gratton wrote:
>> I'd suggest that it would be worthwhile compiling releases with 
>> debugging symbols turned on.
>
> [snip]
>
>> The patch enclosed turns debugging support on by default, but allows 
>> this to be overridden as needed.
>
> [snip]
>
> -- Mike Gratton <mi...@vee.net>, <http://web.vee.net/>
> "Every motive escalate."
>
>
Kimbro Staken - http://www.kstaken.org - http://www.xmldatabases.org
Apache Xindice native XML database http://xml.apache.org/xindice
XML:DB Initiative http://www.xmldb.org
Senior Technologist (Your company name here)

RE: Compiling releases with debugging on

[Vadim Gritsenko <va...@verizon.net>]

I don't have a vote and just lurking here, but +10 from me. Cocoon also
uses debugging ON.

Vadim

> -----Original Message-----
> From: Mike Gratton [mailto:mike@vee.net]
> Sent: Friday, March 08, 2002 6:02 AM
> To: xindice-dev
> Subject: Re: Compiling releases with debugging on
> 
> 
> I haven't really heard anything back about this, what's the verdict
> guys? I still think it would be a mistake to make a release without
> Xindice classes compiled with debugging on, and given other ASF do so
> (Xalan at least, Xerces IIRC), I can't think of any reason not to.
> 
> 
> Mike Gratton wrote:
> >
> > I'd suggest that it would be worthwhile compiling releases with
> > debugging symbols turned on.
> 
> [snip]
> 
> > The patch enclosed turns debugging support on by default, but allows
> > this to be overridden as needed.
> 
> [snip]
> 
> --
> Mike Gratton <mi...@vee.net>, <http://web.vee.net/>
> "Every motive escalate."

Re: Compiling releases with debugging on

[Mike Gratton <mi...@vee.net>]

I haven't really heard anything back about this, what's the verdict 
guys? I still think it would be a mistake to make a release without 
Xindice classes compiled with debugging on, and given other ASF do so 
(Xalan at least, Xerces IIRC), I can't think of any reason not to.


Mike Gratton wrote:
> 
> I'd suggest that it would be worthwhile compiling releases with 
> debugging symbols turned on.

[snip]

> The patch enclosed turns debugging support on by default, but allows 
> this to be overridden as needed.

[snip]

-- 
Mike Gratton <mi...@vee.net>, <http://web.vee.net/>
"Every motive escalate."