Win32::GUI

I've added timers to the testcases so that the Dialog loop actually runs
and the windows gets displayed for one second until the timer kills the
window.
This is a poor-mans gui test, not the required SendMessage tests yet.

Note that the docs for Timer say that Timer->Kill does not destroyl
the timer and it can be re-enabled with ->Interval().
This is not true as my tests show.
Killed timers can be re-enabled by changing the Interval > 0.
Or this there another trick?

> I've added more tests and added a Timer reflection -window to ensure
> that the parent window -timers and -name gets properly deleted on
> KillTimer().
>
> Same should be done for Cursor and Icon, but maybe the window
> can be easily reconstructed from the handle?
>
> I've also removed most of the ending whitespace.
-- 
Reini Urban
http://phpwiki.org/   http://xarch.tu-graz.ac.at/home/rurban/

Hi,

I'm sure if you ask Aldo for permissions to commit changes, he'll add you 
(or, indeed anyone else).

Cheers (and good work BTW).

jez.

Jeremy White sagte:
> I'm sure if you ask Aldo for permissions to commit changes, he'll add y=
ou
> (or, indeed anyone else).

Attached are todays patches:

fix and enhance docs for Timer.
fix Kill() do destroy the Timer, not as described before.
change default interval to 10 [ms] since 1 is invalid and set to 10 on
most NT's.
allow circular reference to parent for perl > 5.6 (Timer, NotifyIcon).
do Kill on Interval(0) as described in Tutorial/Part4.

Note that the default AUTOLOAD method should be enhanced to catch any
undef $control and call DESTROY, and help in circular refs to the parent.
See
http://www.perl.com/pub/a/2002/08/07/proxyobject.html

Timers will print something like
t/05_Menu................ok 3/3 (in cleanup) Can't call method "FETCH" on
an undefined value at ../Win32/GUI.pm line 2700 during global destruction=
.
because the DESTROY method is not called. This is mainly beautification.

One more Timer test enhancement: test NEM Timers

I'm not sure if destroying the parent should kill the children also,
but I would like to fix Win32::GUI to have this "feature", esp.
for Timers which are costly. Petzold talks about max 16 timers
per system at once.

But not for 1.02.
I would say 1.02 is ready. The missing testcases and "features" can
be added later.

Reini Urban sagte:
> Attached are todays patches:
>
> fix and enhance docs for Timer.
> fix Kill() do destroy the Timer, not as described before.
> change default interval to 10 [ms] since 1 is invalid and set to 10 on
> most NT's.
> allow circular reference to parent for perl > 5.6 (Timer, NotifyIcon).
> do Kill on Interval(0) as described in Tutorial/Part4.

Reini,

OK, I've not had a chance to keep up with your tests this week, and 
trying to catch up now find myself mightily confused.   Some of the 
patches seem to be against the source that I called 1.01_03, some seem 
to be incremental (probably on the previous set of patches?).  To add to 
my confusion, my local source has changes beyond the 1.01_03 source that 
you have.

As far as I did get I was failing one test, but I can't tell if that's 
because I didn't apply your patches correctly.  However, what you are 
doing is great, and I want to encourage you to continue, but I am not 
going to get the time in the next couple of weeks  to do much more than 
just keep up with the mail on this list.

I do want to get a V1.02 release out though.  Here's what I propose:

(1) I've put a copy of the current CVS HEAD up at 
http://www.robmay.me.uk/win32gui/. It's identified as 1.01_04.  Reini, 
I'd very much appreciate it if you can confirm whether this code still 
compiles for you under both your Cygwin and MSVC environments.   The 
other MSVC6 users on the list appear to have no issues.

(2) Everyone, a quick show of hands as to who would be happy for the 
current CVS HEAD to be used for the 1.02 release.  Assuming we have no 
negative feedback by the end of the week (when I am next in front of my 
computer), then I will update the version to 1.02, tag the current HEAD 
as release 1.02.

(3) Glenn L (I hope I got the right Glenn) - you offered to do the build 
under MSVC6 for both Perl 5.6 and 5.8.  Are you still happy to do this?  
If you are, then I will mail you once I have done (2).  If you can then 
let me have the PPMs, Aldo has given me access to the sourceforge file 
release system and I'll see if I can drive it and get the files up for 
everyone to use.  Suggestions as to where announcements should be made 
welcome - is there a 'usual' set of places used by Aldo, Laurent et al.?

(4) Once I have a bit more time, I'll work with Reini to get the test 
suite updated.  I'm sure that doing this will turn up more issues than 
those already found by Reini, and that should drive a following release.

Reini, I hope that leaving your tests out of this release is OK with 
you, but I simply don't have time to deal with it now, and I had set 
expectations of a release this week..  (If anyone else has time, and we 
can get some confidence that the tests pass in the various different 
environments by the end of the week, then I'd be happy to take them.)  I 
do agree with Jez, that going forward, if you would like to contribute 
the easiest way for everyone would be for you to mail Aldo and get your 
own CVS access.  At a minimum I think that you should get yourself 
anonymous CVS access so that you can keep a local source tree 
synchronised with the changes others are making - I can provide help 
with this if you need it.

Regards,
Rob.

Robert May schrieb:
> OK, I've not had a chance to keep up with your tests this week, and 
> trying to catch up now find myself mightily confused.   Some of the 
> patches seem to be against the source that I called 1.01_03, some seem 
> to be incremental (probably on the previous set of patches?).  To add to 
> my confusion, my local source has changes beyond the 1.01_03 source that 
> you have.

All of them are against 1.01_03, not incremental. To make it easier.

> As far as I did get I was failing one test, but I can't tell if that's 
> because I didn't apply your patches correctly.  However, what you are 
> doing is great, and I want to encourage you to continue, but I am not 
> going to get the time in the next couple of weeks  to do much more than 
> just keep up with the mail on this list.
> 
> I do want to get a V1.02 release out though.  Here's what I propose:
...
> (4) Once I have a bit more time, I'll work with Reini to get the test 
> suite updated.  I'm sure that doing this will turn up more issues than 
> those already found by Reini, and that should drive a following release.
> 
> Reini, I hope that leaving your tests out of this release is OK with 
> you, but I simply don't have time to deal with it now, and I had set 
> expectations of a release this week..  (If anyone else has time, and we 
> can get some confidence that the tests pass in the various different 
> environments by the end of the week, then I'd be happy to take them.)  I 
> do agree with Jez, that going forward, if you would like to contribute 
> the easiest way for everyone would be for you to mail Aldo and get your 
> own CVS access. 
> At a minimum I think that you should get yourself 
> anonymous CVS access so that you can keep a local source tree 
> synchronised with the changes others are making - I can provide help 
> with this if you need it.

Thanks, no help needed :)
I'm administrator of phpwiki at sf.net and maintainer of several
cygwin projects. If you have questions maybe I can help out :)

I'll ask Aldo then. Also for several Win32::API gcc patches needed.
I took over C::DynaLib recently and want to add gcc support (cygwin + 
mingw) for Win32::API. C::DynaLib callbacks at least works fine for 
Win32, in contrary to Win32::API.
-- 
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/
http://phpwiki.org/

I've re-added my GUI.pm patched and added the tests.
For me all tests work ok on cygwin and MSVC6, perl-5.8.6
With proper ChangeLog this time

Robert, which test exactly fails for you?

> Robert May schrieb:
>> OK, I've not had a chance to keep up with your tests this week, and
>> trying to catch up now find myself mightily confused.   Some of the
>> patches seem to be against the source that I called 1.01_03, some seem
>>  to be incremental (probably on the previous set of patches?).  To add
>> to  my confusion, my local source has changes beyond the 1.01_03
>> source that  you have.
>
> All of them are against 1.01_03, not incremental. To make it easier.
>
>> As far as I did get I was failing one test, but I can't tell if that's
>>  because I didn't apply your patches correctly.  However, what you are
>>  doing is great, and I want to encourage you to continue, but I am not
>>  going to get the time in the next couple of weeks  to do much more
>> than  just keep up with the mail on this list.
>>
>> I do want to get a V1.02 release out though.  Here's what I propose:
> ...
>> (4) Once I have a bit more time, I'll work with Reini to get the test
>> suite updated.  I'm sure that doing this will turn up more issues than
>>  those already found by Reini, and that should drive a following
>> release.
>>
>> Reini, I hope that leaving your tests out of this release is OK with
>> you, but I simply don't have time to deal with it now, and I had set
>> expectations of a release this week..  (If anyone else has time, and
>> we  can get some confidence that the tests pass in the various
>> different  environments by the end of the week, then I'd be happy to
>> take them.)  I  do agree with Jez, that going forward, if you would
>> like to contribute  the easiest way for everyone would be for you to
>> mail Aldo and get your  own CVS access.
>> At a minimum I think that you should get yourself
>> anonymous CVS access so that you can keep a local source tree
>> synchronised with the changes others are making - I can provide help
>> with this if you need it.
>
> Thanks, no help needed :)
> I'm administrator of phpwiki at sf.net and maintainer of several
> cygwin projects. If you have questions maybe I can help out :)
>
> I'll ask Aldo then. Also for several Win32::API gcc patches needed. I
> took over C::DynaLib recently and want to add gcc support (cygwin +
> mingw) for Win32::API. C::DynaLib callbacks at least works fine for
> Win32, in contrary to Win32::API.
-- 
Reini Urban
http://phpwiki.org/   http://xarch.tu-graz.ac.at/home/rurban/

> I've re-added my GUI.pm patched and added the tests.
> For me all tests work ok on cygwin and MSVC6, perl-5.8.6
> With proper ChangeLog this time

Oops, I made a small mistake with two lines in GUI.pm,
needed for perl < 5.6
Attached is the patch against yesterdays patch.

-- 
Reini Urban
http://phpwiki.org/   http://xarch.tu-graz.ac.at/home/rurban/

Reini Urban sagte:
>> I've re-added my GUI.pm patched and added the tests.
>> For me all tests work ok on cygwin and MSVC6, perl-5.8.6
>> With proper ChangeLog this time
>
> Oops, I made a small mistake with two lines in GUI.pm,
> needed for perl < 5.6
> Attached is the patch against yesterdays patch.

Also verified ok with latest perl-5.8.7 for cygwin.
All tests pass, now even without the global destructor warning as before.

So tested with perl-5.8.7 and 5.8.6 with cygwin, and
ActivePerl-5.8.6.811 with MSVC6

All,

I just did a final quick tidy-up and tagged the HEAD revisions as 
'Win32-GUI-1_02'.

I hope we'll have a release on SourceForge very soon.

Regards,
Rob.

Robert May wrote:

> Reini,
>
> OK, I've not had a chance to keep up with your tests this week, and 
> trying to catch up now find myself mightily confused.   Some of the 
> patches seem to be against the source that I called 1.01_03, some seem 
> to be incremental (probably on the previous set of patches?).  To add 
> to my confusion, my local source has changes beyond the 1.01_03 source 
> that you have.
>
> As far as I did get I was failing one test, but I can't tell if that's 
> because I didn't apply your patches correctly.  However, what you are 
> doing is great, and I want to encourage you to continue, but I am not 
> going to get the time in the next couple of weeks  to do much more 
> than just keep up with the mail on this list.
>
> I do want to get a V1.02 release out though.  Here's what I propose:
>
> (1) I've put a copy of the current CVS HEAD up at 
> http://www.robmay.me.uk/win32gui/. It's identified as 1.01_04.  Reini, 
> I'd very much appreciate it if you can confirm whether this code still 
> compiles for you under both your Cygwin and MSVC environments.   The 
> other MSVC6 users on the list appear to have no issues.
>
> (2) Everyone, a quick show of hands as to who would be happy for the 
> current CVS HEAD to be used for the 1.02 release.  Assuming we have no 
> negative feedback by the end of the week (when I am next in front of 
> my computer), then I will update the version to 1.02, tag the current 
> HEAD as release 1.02.
>
> (3) Glenn L (I hope I got the right Glenn) - you offered to do the 
> build under MSVC6 for both Perl 5.6 and 5.8.  Are you still happy to 
> do this?  If you are, then I will mail you once I have done (2).  If 
> you can then let me have the PPMs, Aldo has given me access to the 
> sourceforge file release system and I'll see if I can drive it and get 
> the files up for everyone to use.  Suggestions as to where 
> announcements should be made welcome - is there a 'usual' set of 
> places used by Aldo, Laurent et al.?
>
> (4) Once I have a bit more time, I'll work with Reini to get the test 
> suite updated.  I'm sure that doing this will turn up more issues than 
> those already found by Reini, and that should drive a following release.
>
> Reini, I hope that leaving your tests out of this release is OK with 
> you, but I simply don't have time to deal with it now, and I had set 
> expectations of a release this week..  (If anyone else has time, and 
> we can get some confidence that the tests pass in the various 
> different environments by the end of the week, then I'd be happy to 
> take them.)  I do agree with Jez, that going forward, if you would 
> like to contribute the easiest way for everyone would be for you to 
> mail Aldo and get your own CVS access.  At a minimum I think that you 
> should get yourself anonymous CVS access so that you can keep a local 
> source tree synchronised with the changes others are making - I can 
> provide help with this if you need it.
>
> Regards,
> Rob.
>
>
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
> _______________________________________________
> Perl-Win32-GUI-Hackers mailing list
> Perl-Win32-GUI-Hackers@...
> https://lists.sourceforge.net/lists/listinfo/perl-win32-gui-hackers
>

At 13:57 2005-07-10, Robert May wrote:
>I just did a final quick tidy-up and tagged the HEAD revisions as 
>'Win32-GUI-1_02'.
>
>I hope we'll have a release on SourceForge very soon.

Great news, and great work everyone!


/J

Reini Urban wrote:

> Robert May schrieb:
>
>> OK, I've not had a chance to keep up with your tests this week, and 
>> trying to catch up now find myself mightily confused.   Some of the 
>> patches seem to be against the source that I called 1.01_03, some 
>> seem to be incremental (probably on the previous set of patches?).  
>> To add to my confusion, my local source has changes beyond the 
>> 1.01_03 source that you have.
>
>
> All of them are against 1.01_03, not incremental. To make it easier.

My apologies.   I couldn't get them to apply to my source without 
errors, and at a quick scan it didn't look like the later patches 
contained all the changes from the earlier ones.

I've done now Devel::Cover tests for the old and new tests.

$ perldoc Devel::Cover
$ HARNESS_PERL_SWITCHES=3D-MDevel::Cover make test
$ cover

what                   stmt   bran   cond   sub   pod    time   total
3 tests only:        11.4     6.3     9.5   11.1  16.7  100.0    9.9
with my tests      48.5    26.0   28.6   63.7  16.7  100.0  43.2

That means that with my included tests we get a coverage of 43.2%,
compared to 9.9% with the current tests.
I'm working on getting closer to 100%, but the pod coverage of 16.7% is poo=
r.

It's interesting to see the detailed reports, which statements or if-then-e=
lse
branches are never executed in the tests.

On 7/11/05, Robert May <rmay@...> wrote:
> >Reini Urban sagte:
> >>>I've re-added my GUI.pm patched and added the tests.
> >>>For me all tests work ok on cygwin and MSVC6, perl-5.8.6
> >>>With proper ChangeLog this time
> >
> >Also verified ok with latest perl-5.8.7 for cygwin.
> >All tests pass, now even without the global destructor warning as before=
.
> >
> >So tested with perl-5.8.7 and 5.8.6 with cygwin, and
> >ActivePerl-5.8.6.811 with MSVC6
> >
> Great.  I'll try to have a look at this properly later in the week.
> Rob.
--
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/

Reini Urban wrote:
> I've done now Devel::Cover tests for the old and new tests.
> 
> $ perldoc Devel::Cover
> $ HARNESS_PERL_SWITCHES=-MDevel::Cover make test
> $ cover
> 
> what                   stmt   bran   cond   sub   pod    time   total
> 3 tests only:        11.4     6.3     9.5   11.1  16.7  100.0    9.9
> with my tests      48.5    26.0   28.6   63.7  16.7  100.0  43.2
> 
> That means that with my included tests we get a coverage of 43.2%,
> compared to 9.9% with the current tests.
> I'm working on getting closer to 100%, but the pod coverage of 16.7% is poor.
> 
> It's interesting to see the detailed reports, which statements or if-then-else
> branches are never executed in the tests.

I'll have to install Devel::Cover and have a look.  Thanks for the 
heads-up.   Would I be right in assuming that the coverage figures are 
for the Perl code only, and not the XS code?

Regards,
Rob.