[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Is anyone working on a test suite?

Gmail isn't working well with Sourceforge at the moment so I am
sending this to both places.

I think a test suite will be very important.  There are, however, a
few challenges:
1)  I don't think SQL-Ledger ever underwent any sort of formal testing
(as evidenced by the COGS bug I found).  Therefore we can expect a
*lot* of corner-case bugs in the current code-base.  Indeed we have
been correcting bugs even involing seemingly obvious things like
rounding numbers....

2)  We are in the process of dramatically changing the application at
many levels in order to address serious shortcomings of the current
codebase.  A test suite may be a lot of work to tell us what we
already know (the current code is buggy).

My own suggestion is that a test suite may begin to be useful at
around 1.3 when certain parts of the application may become more
stable.  Then we can expect that the test suite may take as long to
develop as it takes us to re-engineer the entire application (maybe 6
months to a year).  I think that the test suite should cover the
portions of the application that are stable, and not those we hacen't
gotten to yet.  This will allow us to define the application behavior
as we develop the application.

Best Wishes,
Chris Travers

On 10/2/06, Hugh Esco <..hidden..> wrote:
Hey all:

I'm excited to see a more open process for the development of
the code base which Dieter created and particularly for the
development of documentation which might make it easier to
propogate this code more broadly.

I'm curious to know where this may all head and hope to perhaps
work with others to build on this code base, tools for managing
a membership base for member based organizations.

But it makes me nervous to start working with this code without
a test suite.  That's like coding without a net.  I never
saw a t/ directory in the tarballs I got from SL's site.
I'm wondering if there has been any thought given to building
a test suite for LSMB.

I don't have a lot of time to spare for this, but if I could
make some contribution to the project, it might be in the area
of building a test suite.  I'd take one module at a time and
build tests one method at a time.  It would be slow, it would
be tedious.  But over time I might be able to make a meaningful
contribution in that way.

But it would also be a waste of time unless the Project as a
whole were willing to use, maintain and build on it.

What do folks think.  In terms of project infrastructure, is
this something the existing project members would be willing
to commit to?

It would certainly make the job of those folks building new
functionality, or adapting / maintaining existing features
alot easier.

-- Hugh Esco

Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
Ledger-smb-devel mailing list