Re: Code donations for RPGUI initiative

[Aaron Bartell <aaronbartell@xxxxxxxxx>]

Cynical works :-)  I am just trying to call it like I see it and it is
possible I called it too soon and you are only being cynical with this
situation.

Maybe you could expound on why you would be experienced as it relates to
this situation of not contributing to open source.  Have you been burned by
open source?

Thanks for the tips - I do value them.  It is easy to lose site of the
dangers of public code, though one would hope that with source comparison
and multipel versions, it would be easier to find malicious code.  I would
guess (emphasis on guess) the issues from malicious code in open source and
the oversights in private code, would be about equal - though I have no
numbers to back them up.

I might need to start another thread to see what people use for automated
unit testing their RPG code.  I have an approach I use with RPG-XML Suite
that works, but think that there must be something better out there.
Basically what I do is I have a main RPG program up on the AS400 that when
invoked will subsequently invoke a particular pre-setup test.  A pre-setup
test is simply a sub routine that performs a single function against one of
my API's in a given scenario, and if I get the expected response I return
"sucess" up the call stack, if the response is not expected I return
"false".  I have all of this process being invoked from JUnit in eclipse on
my PC so it is easy to see the reported errors returned back from the unit
test.  Hmmm.. I might need to write an article about it - that brings people
out of the woodwork to give opinions on how else I could have done it (free
critiques is what I call it :-)

Even though we have ventured into a tougher conversation here, I think we
are still doing good things.

Aaron Bartell
www.SoftwareSavesLives.com


On Fri, Jan 8, 2010 at 11:50 AM, Nathan Andelin <nandelin@xxxxxxxxx> wrote:

> Please don't label me as "bitter".  A bit cynical, perhaps.  Experienced,
> would be preferable.  The appeal of being part of such widely reaching
> endeavors is not for me.  I'll offer one more tip. API's that engage
> programmers to dynamically allocate and deallocate memory are quite prone to
> unsustainable memory leaks, which are arguably the most difficult bugs to
> resolve in Web applications after the fact.  Actually, here's another tip.
>  When managing repositories of public code, beware of disgruntled Java
> programmers (a metaphor, only) embedding memory leaks deep in meaty RPG code
> donated to public repositories.  If you'd normally spend 10 hours testing
> your own code, spend 100 hours testing that which you might extract from
> others.
>
> Nathan.