[Tcl-bugs] [ tktoolkit-Bugs-854178 ] textWind-7.* fail

[SourceForge.net <no...@so...>]

Bugs item #854178, was opened at 2003-12-04 17:26
Message generated for change (Comment added) made by vincentdarley
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112997&aid=854178&group_id=12997

Category: 18. [text]
Group: develop: 8.5a0
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Don Porter (dgp)
Assigned to: Vince Darley (vincentdarley)
Summary: textWind-7.* fail

Initial Comment:

Using 2003-12-04 HEAD on Solaris 9/CDE:


==== textWind-7.1 EmbWinLostSlaveProc procedure FAILED
==== Contents of test case:

    .t delete 1.0 end
    .t insert 1.0 "Some sample text"
    frame .f -width 10 -height 20 -bg $color
    .t window create 1.2 -window .f
    update
    place .f -in .t -x 100 -y 50
    update
    list [winfo geom .f] [.t bbox 1.2]

---- Result was:
10x20+105+55 {19 11 0 0}
---- Result should have been (exact matching):
10x20+104+54 {19 11 0 0}
==== textWind-7.1 FAILED


==== textWind-7.2 EmbWinLostSlaveProc procedure FAILED
==== Contents of test case:

    .t delete 1.0 end
    .t insert 1.0 "Some sample text"
    frame .t.f -width 10 -height 20 -bg $color
    .t window create 1.2 -window .t.f
    update
    place .t.f -x 100 -y 50
    update
    list [winfo geom .t.f] [.t bbox 1.2]

---- Result was:
10x20+105+55 {19 11 0 0}
---- Result should have been (exact matching):
10x20+104+54 {19 11 0 0}
==== textWind-7.2 FAILED


----------------------------------------------------------------------

>Comment By: Vince Darley (vincentdarley)
Date: 2003-12-10 12:15

Message:
Logged In: YES 
user_id=32170

A comparison with the behaviour of 'place' inside a frame
(adjusting -padx, -pady) shows that indeed the text widget
should do what it currently does, and therefore the old test
and old behaviour was wrong.

Text widget is now consistent with the behaviour of a frame
wrt embedded windows and overall widget padding.


----------------------------------------------------------------------

Comment By: Vince Darley (vincentdarley)
Date: 2003-12-10 10:05

Message:
Logged In: YES 
user_id=32170

I think you'll need to set -padx 0 -pady 0 as well, and then
all will be fine.  I believe therefore the new result is
correct, and the test is in error.


----------------------------------------------------------------------

Comment By: Don Porter (dgp)
Date: 2003-12-09 21:50

Message:
Logged In: YES 
user_id=80530


one more data point.  If I set
-bd 0 -highlightthickness 0

and [place ... -x 0 -y 0] then
the result is

10x20+1+1 {15 7 0 0}

Note the +1+1 shift instead
of +0+0 .  I don't know where
this extra pixel offset comes
from, but it appears to be
consistently present.

----------------------------------------------------------------------

Comment By: Don Porter (dgp)
Date: 2003-12-09 21:40

Message:
Logged In: YES 
user_id=80530


If both the -bd and -highlightthickness
options of the [text] widget are set to 0,
then the result of both tests is:

10x20+101+51 {15 7 0 0}

As each pixel is added to either of
those options, a pixel is added to the
offsets and the bbox results, which
seems ok to me.

If I set the options of the [place] command
to -x 99 -y 49 then the tests pass.

This is true on both Solaris and Linux/Alpha.

----------------------------------------------------------------------

Comment By: Vince Darley (vincentdarley)
Date: 2003-12-09 16:36

Message:
Logged In: YES 
user_id=32170

I meant (thinking more clearly now): Can you count the
pixels to tell whether +55 or +54 is the correct answer?



----------------------------------------------------------------------

Comment By: Don Porter (dgp)
Date: 2003-12-09 16:23

Message:
Logged In: YES 
user_id=80530

well, when I adjust -bd and -highlightthickness
of the [text .t] widget, I shift the values after the +'s,
but I also change the 1st and 2nd values in the [bbox]
result.

I'm not sure what it would mean
to "look correct".


----------------------------------------------------------------------

Comment By: Vince Darley (vincentdarley)
Date: 2003-12-09 09:19

Message:
Logged In: YES 
user_id=32170

Could you verify whether these test results look correct to
you or not?  (Taking account of the text widget's border
width, highlightthickness and padx/pady --- which you might
wish to increase temporarily for testing purposes?).

I think it might be that these tests are now in error...


----------------------------------------------------------------------

Comment By: Don Porter (dgp)
Date: 2003-12-04 17:40

Message:
Logged In: YES 
user_id=80530


Same tests fail the same way
on Linux/Alpha/GNOME

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112997&aid=854178&group_id=12997