[XForms] New release

[Jens Thoms Toerring]

Hi,

sorry to bother you but has anyone of you given the
last pre-release a try? I haven't got any feedback and
I'm starting to wonder if the mailing list software may
have problems. So if you did tests it would be helpful
if you would just let me know that you did so - even if
you didn't encounter any new problems.

Best regards, Jens
--
\ Jens Thoms Toerring ________ j...@toerring.de
\_______________________________ http://toerring.de

[Paul]

Hi Jens, et al,

I've been using 1.0.94pre10 for some time (no problems)
but haven't had the chance to try 11.

Too many projects on the go here already - wish I had time
to get more involved with XForms.

Often I have wondered how regression tests are done on GUI
software. There must be a way to make a test harness - a
sequence of XForms library calls to build a form, simulated
button and mouse moves to exercise it, and something to check
that the window is correct at the pixel level at each step.

When testing a release here, all I can do is recompile my
applications and try them out, not really a thorough test
as only a few common library functions are called and none
of the new features are put to the test.

I'll give the 1.0.94pre11 a run later today.

--
Paul Nicholson
http://abelian.org
--

[Jens Thoms Toerring]

Hi Paul.

On Sun, May 13, 2012 at 06:07:52AM +0000, Paul wrote:
> I've been using 1.0.94pre10 for some time (no problems)
> but haven't had the chance to try 11.
>
> Too many projects on the go here already - wish I had time
> to get more involved with XForms.

No problem at all, if you tell when you experience
trouble with XForms that's already a big help.

> Often I have wondered how regression tests are done on GUI
> software. There must be a way to make a test harness - a
> sequence of XForms library calls to build a form, simulated
> button and mouse moves to exercise it, and something to check
> that the window is correct at the pixel level at each step.

There is, unfortunately, no systematic testing. I also
have no idea how one would write a test harness that could
automate testing of a GUI. So I try new versions with
my programs that use the library and run some of the demo
programs that I guess could be influenced by the changes
I made - doing all (nearly 100) of them would be too much
work... So I have to rely on feedback from others that
might be using stuff I didn't cover (or use it in ways I
never thought of).

> When testing a release here, all I can do is recompile my
> applications and try them out, not really a thorough test
> as only a few common library functions are called and none
> of the new features are put to the test.

Normally you just would have to only recompile the library
and not also your applications (as long as there eren't too
invasive changes like changing the layout of the FL_FORM
or FL_OBJECT structure). I try to avoid such changes and
when they become necessary I'll explicitely tell you. If you
should find that you have to recompile your applications to
get them to work again anyway then this is rather likely a
hint that there's something broken, i.e. I made a change
that shouldn't have been done that way.

> I'll give the 1.0.94pre11 a run later today.

Thank you very much, Jens

[Paul]

Out of habit I force a recompile of applications, just in
case the header has altered any of the interface definition.

Agreed, that is unnecessary for most changes and could actually
hide a defect.

Maybe Xnee could be put to use for regression testing,

http://itupw056.itu.chalmers.se/project-xnee/

I wish I had more time, perhaps someone on the mailing list
will test if the XForms demo program can be exercised by
using cnee to record and play back a sequence of XEvents.
If that worked, it could be a nice project for someone to
make an XForms test program that exercised all the objects
using cnee. I suppose at the simplest, it would just
run a sequence of X events into the form and then test
the final window against a reference image.

--
Paul Nicholson
--

[Jens Thoms Toerring]

Hi Paul,

On Sun, May 13, 2012 at 02:18:55PM +0000, Paul wrote:
> Maybe Xnee could be put to use for regression testing,
>
> http://itupw056.itu.chalmers.se/project-xnee/

Thanks for the tip, I wasn't aware of that one!

> I wish I had more time, perhaps someone on the mailing list
> will test if the XForms demo program can be exercised by
> using cnee to record and play back a sequence of XEvents.
> If that worked, it could be a nice project for someone to
> make an XForms test program that exercised all the objects
> using cnee. I suppose at the simplest, it would just
> run a sequence of X events into the form and then test
> the final window against a reference image.

Mmm, that's an idea. You probably would have to run it first
with the old version and then with the new one so that both
use the exact same fonts etc. Could be quite an interesting
project.

Though it would only test those things that have been put
into the scripts and quite a number of bug reports I got
are about combinations I never would have considered (like
"if you have an object of type A and one of type B that
hides part of A and then you do this then A isn't correct-
ly redrawn under certain circumstances"). But I guess that
something that only real world testing will be able to catch...

Anyone feeling like implementing some, however simple, test
suite is heartly invited to do so - I fear I won't have too
much time in the coming months.

Best regards, Jens