Yeah, I've been removing the locals also, but when reading through these classes I also realized that they could be useful for debugging. Will try to keep them in, though I usually heavily modify the generated byte code after the first batch I get. Most of the time I'd have to re-target the local var labels, but I think it could be worth the extra effort.
On 02 Oct 2007, at 18:37, Tim Eck wrote: > Second that. Any methods that are synthetic in nature shouldn't > really get line number tables. Seems like it can only confuse, > never help. > > I usually trim out the needless Labels that get generated, as well > as the local variable information. The extra labels don't hurt > anything (just take up space in the source). Now that I think about > though, leaving in locals might be useful actually (and might mean > some of the Labels might need to remain). > > > -----Original Message----- > > From: [EMAIL PROTECTED] [mailto:tc-dev- > > [EMAIL PROTECTED] On Behalf Of Alex Miller > > Sent: Tuesday, October 02, 2007 9:31 AM > > To: [email protected] > > Subject: Re: [tc-dev] Line numbers in generated code? > > > > I've been removing the line number instructions from code > generated by > > the bytecode plugin when I move it into our code. I think it's a > good > > idea as those numbers are not meaningful and actually confusing for > > debuggers, etc as you say. > > > > ----- Original Message ----- > > From: "Geert Bevin" <[EMAIL PROTECTED]> > > To: [email protected] > > Sent: Tuesday, October 2, 2007 11:10:10 AM (GMT-0600) America/ > Chicago > > Subject: [tc-dev] Line numbers in generated code? > > > > Hi everyone, > > > > while reading through the code for the ConcurrentHashMap support I > > ran into a lot of code that seems to be straight copy-pastes of Java > > code that was converted to ASM using the Eclipse plugin. For > > instance, look at JavaUtilConcurrentHashMapAdapter and for instance > > the method createWrapperMethod(). This is an instrumentation that is > > using to create a wrapper method around existing methods, yet it > uses > > line numbers. This means that the same line numbers are present in > > the byte code at different locations and for different method names. > > Almost all the other generated methods in this class to the same > thing. > > > > I think it's a good practice to remove the line numbers of generated > > byte code since I think that this could result into maybe unwanted > > behavior with debuggers / profilers / etc. > > > > What do you think? Should I removed the line numbers that are in > > those classes now? > > > > Geert > > > > -- > > Geert Bevin > > Terracotta - http://www.terracotta.org > > Uwyn "Use what you need" - http://uwyn.com > > RIFE Java application framework - http://rifers.org > > Music and words - http://gbevin.com > > > > _______________________________________________ > > tc-dev mailing list > > [email protected] > > http://lists.terracotta.org/mailman/listinfo/tc-dev > > > > _______________________________________________ > > tc-dev mailing list > > [email protected] > > http://lists.terracotta.org/mailman/listinfo/tc-dev > > _______________________________________________ > tc-dev mailing list > [email protected] > http://lists.terracotta.org/mailman/listinfo/tc-dev -- Geert Bevin Terracotta - http://www.terracotta.org Uwyn "Use what you need" - http://uwyn.com RIFE Java application framework - http://rifers.org Music and words - http://gbevin.com _______________________________________________ tc-dev mailing list [email protected] http://lists.terracotta.org/mailman/listinfo/tc-dev
