#2150 frame test 2.5 failure

[Larry W. Virden]

built cvs head tk 8.5 on sparc solaris 2.9 using sun c
compiler; display for test suite was exceed 10.0.0.12
running on windows xp.

frame.test

==== frame-2.5 toplevel configuration options FAILED
==== Contents of test case:

catch {destroy .t}
toplevel .t -width 200 -height 100
wm geometry .t +0+0
list [catch {.t configure -use 0x44022} msg] $msg
[.t configure -use]

---- Result was:
1 {couldn't create child of window "0x44022"} {-use use
Use {} {}}
---- Result should have been (exact matching):
1 {can't modify -use option after widget is created}
{-use use Use {} {}}
==== frame-2.5 FAILED

[Peter Spjuth]

Logged In: YES
user_id=98900
Originator: NO

This was caused by rev 1.19-1.20 of tkFrame.c,
made by Chengye. Reassigning.

[Nobody/Anonymous]

Logged In: NO

I tested it at my Windows/XP and could not duplicate the bug.

[Larry W. Virden]

Logged In: YES
user_id=15949
Originator: YES

Note that the bug is reported to be occuring when running on SPARC Solaris displaying back to a Windows XP using Exceed 10. Were you trying that combo when you failed to duplicated the bug?
Because it's been happening in this combination for about 2 yrs now.

[Chengye Mao]

Logged In: YES
user_id=191079
Originator: NO

I added a #ifdef SUPPORT_CONFIG_EMBEDDED to conditionaly compile tkFrame.c. By default, it keeps the original error string if the platform does not support the configuration of an embedded window.

Currently only Windows platform supports configuration of -use option. A compilation flag -DSUPPORT_CONFIG_EMBEDDED is required to enable this freature and to pass frame2-5 test.

Please give it a try and let me know the result (I don't have a Linux or Solaris for a test).

[Mark Janssen]

Logged In: YES
user_id=1463011
Originator: NO

Same error occurs when testing Tk 85b1 RC2 On windows XP (built with MinGW/gcc)

[Michael Schlenker]

Logged In: YES
user_id=302287
Originator: NO

Test runs fine on Windows XP SP 2, Tk 8.5 HEAD (post 8.5b1) when configuring with CFLAGS=-DSUPPORT_CONFIG_EMBEDDED, as mentioned by chengye, it fails without the define (naturally as, the check for the windows isn't executed in that case).

Maybe add the define as a default on windows?

[Larry W. Virden]

Logged In: YES
user_id=15949
Originator: YES

I am torn about making this define. As I mentioned in my original report, my Tk doesn't _run_ on Windows. It runs on Suns. Sometimes it is displaying to a Sun SPARC X server. For other users, this solaris code is displaying back to a Windows Exceed X server.

If I make this define, with both types of users be able to run the same Sun Tk extension? If not, then that's not a compile time option I could choose.