opentracker-develop — OpenTracker Developer Discussions

[Opentracker-develop]Crash at startup

[Rainer R. <ml...@na...>]

Hi,

I have a problem with OpenTracker (the latest cvs version) crashing 
sometimes at startup. Perhaps someone has a idea:

loading symbols
segment violation occurred
BList::CountItems(void) const:
CountItems__C5BList:
+0006  ec27c97e:   *        08408b    movl    0x00000008(%eax), %eax
AutomountInitialScan:sc
   frame         retaddr
fd12dcb4   ec28c219  TypedList<Session *>::CountItems(void) const + 00000019
fd12dcc4   ec28bd12  EachPartitionIterator<EachMountedPartitionAdaptor, 
Partition *(*)(Partition *, void *), Partition *, void 
*>::EachPartition(DeviceList *, Partition *(*)(Partition *, void *), void 
*) + 00000052
fd12dcf4   ec25fbe9  DeviceList::EachMountedPartition(Partition 
*(*)(Partition *, void *), void *) + 00000021
fd12dd0c   ec4dd66e  BPrivate::AutoMounter::InitialRescanBinder(void *) + 
0000004a
fd12dd3c   ec05650d  thread_start + 00000039


The problem could be related to a hardware/software (UDMA-100/IDE 
replacement driver from BeBits) problem on my machine that causes other 
than the boot partitions to take some minutes until they are mounted 
sometimes. But that doesn't always seem to crash Tracker.

Perhaps someone with deeper Tracker knowledge does know what happens here

-- 
Bye,
   Rainer

Re: [Opentracker-develop] Switcher Keystrokes Missing

[Siarzhuk Z. <im...@gm...>]

Good evening,

06.12.2005 03:48 you wrote:

>Siarzhuk Zharski wrote on Tue, 29 Nov 2005 10:58:20 +0100:
>  
>
>>BTW latest CVS version of the Tracker have problems with keymap Switcher 
>>by Stas Maximov (http://bebits.com/app/233, 
>>http://sourceforge.net/projects/switcher). To tell you truth this 
>>software is must have for every Russian-speaking user. :-)
>>I have tried latest Tracker on two systems and get the same result: 
>>Switcher's input filter doesn't react on key  combination,  Settings 
>>window cannot save and locks.
>>    
>>
>
>So you're saying that Switcher doesn't get activated when you press its hot-keys?  How does Tracker handle keystrokes for add-ins like Switcher?  Or did Switcher wedge itself into the input server event chain to intercept keystrokes (in which case Tracker doesn't have anything to do with it).
>
>I don't know how Switcher gets its keystrokes - that would be the first thing to look into.
>
>As for Switcher's settings window locking, it sounds like Switcher is buggy, uses really old code APIs or something is wrong with the file requestor (which Tracker does supply).
>
>Unfortunately I'm busy with too many other things to look into it.  Perhaps you can check into it or find someone who has time.
>  
>
I have contacted with Switcher's Author about this problem. Hope this 
problem will be fixed ... :-)

-- 
Kind regards,
           S.Zharski

[Opentracker-develop] Switcher Keystrokes Missing

[Alexander G. M. S. <agm...@ro...>]

Siarzhuk Zharski wrote on Tue, 29 Nov 2005 10:58:20 +0100:
> BTW latest CVS version of the Tracker have problems with keymap Switcher 
> by Stas Maximov (http://bebits.com/app/233, 
> http://sourceforge.net/projects/switcher). To tell you truth this 
> software is must have for every Russian-speaking user. :-)
> I have tried latest Tracker on two systems and get the same result: 
> Switcher's input filter doesn't react on key  combination,  Settings 
> window cannot save and locks.

So you're saying that Switcher doesn't get activated when you press its hot-keys?  How does Tracker handle keystrokes for add-ins like Switcher?  Or did Switcher wedge itself into the input server event chain to intercept keystrokes (in which case Tracker doesn't have anything to do with it).

I don't know how Switcher gets its keystrokes - that would be the first thing to look into.

As for Switcher's settings window locking, it sounds like Switcher is buggy, uses really old code APIs or something is wrong with the file requestor (which Tracker does supply).

Unfortunately I'm busy with too many other things to look into it.  Perhaps you can check into it or find someone who has time.

- Alex

Re: [Opentracker-develop]Re: Time for an OpenTracker Release?

[Siarzhuk Z. <im...@gm...>]

Good evening,

28.11.2005 12:14 you wrote:

>"Alexander G. M. Smith" <agm...@ro...> wrote:
>  
>
>>The PowerPC version does compile, with a few warnings.  I'll include 
>>the
>>error messages so that someone with CVS access can fix them (mostly
>>changing 1.0 to 1.0F).  Seems to run OK as-is.  Only worrysome one is
>>the bitwise OR, should be a logical OR.
>>    
>>
>
>Doesn't really matter for boolean values, though :-)
>Thanks for the note, I think I've fixed all issues.
>  
>
BTW latest CVS version of the Tracker have problems with keymap Switcher 
by Stas Maximov (http://bebits.com/app/233, 
http://sourceforge.net/projects/switcher). To tell you truth this 
software is must have for every Russian-speaking user. :-)
I have tried latest Tracker on two systems and get the same result: 
Switcher's input filter doesn't react on key  combination,  Settings 
window cannot save and locks. Currently I use the Tracker's version that 
correspond to time of "my" fix with loading lock. (see my e-mail 
28.05.05 for details)  It have no such problem. Do you have any idea 
about sources of this problem? In some days I plan to look into sources 
of the Switcher to investigate this problem and let you know about results.

-- 
Kind regards,
           S.Zharski

Re: [Opentracker-develop]Re: Time for an OpenTracker Release?

[Alexander G. M. S. <agm...@ro...>]

Axel Dörfler wrote on Mon, 28 Nov 2005 12:14:01 +0100 CET:
> Thanks for the note, I think I've fixed all issues.

Got the latest version from CVS with your fixes, compiles with no warnings.  Great!

- Alex

Re: [Opentracker-develop]Time for an OpenTracker Release?

[Axel <ax...@pi...>]

Ingo Weinhold <bon...@cs...> wrote:
> On 2005-11-28 at 23:16:27 [+0100], Axel D=F6rfler wrote:
> > "Alexander G. M. Smith" <agm...@ro...> wrote:
> > > The last release was in 2003, and I think there are enough new
> > > features to make
> > > it worthwhile to do another one.  Any opinions=3F
> > Yes, it's definitely worthwhile. However, at least the draggable 
> > icon
> > feature in the upper right corner is not yet ready for prime time: 
> > it
> > has some drawing bugs, and also has some other minor problems.
> > I look into this again, maybe I can get it complete this week - it
> > shouldn't be too much work involved.
> Since you're using FFM you might not have noticed, that a mouse down 
> on the 
> icon causes the window to be activated for non-FFM, which is a bit 
> annoying, 
> if one wants to drag'n'drop it to the currently active window (e.g. a 
> file 
> panel).

Indeed, thanks for the tip!

Bye,
   Axel.

Re: [Opentracker-develop]Time for an OpenTracker Release?

[Ingo W. <bon...@cs...>]

On 2005-11-28 at 23:16:27 [+0100], Axel D=F6rfler wrote:
> "Alexander G. M. Smith" <agm...@ro...> wrote:
> > The last release was in 2003, and I think there are enough new
> > features to make
> > it worthwhile to do another one.  Any opinions?
>=20
> Yes, it's definitely worthwhile. However, at least the draggable icon
> feature in the upper right corner is not yet ready for prime time: it
> has some drawing bugs, and also has some other minor problems.
> I look into this again, maybe I can get it complete this week - it
> shouldn't be too much work involved.

Since you're using FFM you might not have noticed, that a mouse down on the=
=20
icon causes the window to be activated for non-FFM, which is a bit annoying=
,=20
if one wants to drag'n'drop it to the currently active window (e.g. a file=
=20
panel).

CU, Ingo

Re: [Opentracker-develop]Time for an OpenTracker Release?

[Axel <ax...@pi...>]

"Alexander G. M. Smith" <agm...@ro...> wrote:
> The last release was in 2003, and I think there are enough new 
> features to make
> it worthwhile to do another one.  Any opinions?

Yes, it's definitely worthwhile. However, at least the draggable icon 
feature in the upper right corner is not yet ready for prime time: it 
has some drawing bugs, and also has some other minor problems.
I look into this again, maybe I can get it complete this week - it 
shouldn't be too much work involved.

> If needed, I can compile the R5 and PPC versions, but not BONE 
> (though I could
> dust of my Zeta RC3 and figure out how to use it).  Guess I should 
> try the PPC
> first on my own, in case it no longer compiles.

I can compile the R5 version, but it would be great if you could do the 
PPC version. I don't think there is any need for a special BONE or Zeta 
release.

Bye,
   Axel.

[Opentracker-develop]Time for an OpenTracker Release?

[Alexander G. M. S. <agm...@ro...>]

The last release was in 2003, and I think there are enough new features to make it worthwhile to do another one.  Any opinions?

If needed, I can compile the R5 and PPC versions, but not BONE (though I could dust of my Zeta RC3 and figure out how to use it).  Guess I should try the PPC first on my own, in case it no longer compiles.

- Alex

Re: [Opentracker-develop]Re: Time for an OpenTracker Release?

[Axel <ax...@pi...>]

"Alexander G. M. Smith" <agm...@ro...> wrote:
> The PowerPC version does compile, with a few warnings.  I'll include 
> the
> error messages so that someone with CVS access can fix them (mostly
> changing 1.0 to 1.0F).  Seems to run OK as-is.  Only worrysome one is
> the bitwise OR, should be a logical OR.

Doesn't really matter for boolean values, though :-)
Thanks for the note, I think I've fixed all issues.

Bye,
   Axel.

[Opentracker-develop]Re: Time for an OpenTracker Release?

[Alexander G. M. S. <agm...@ro...>]

The PowerPC version does compile, with a few warnings.  I'll include the error messages so that someone with CVS access can fix them (mostly changing 1.0 to 1.0F).  Seems to run OK as-is.  Only worrysome one is the bitwise OR, should be a logical OR.

- Alex


mwcc -c PoseView.cpp  -I./  -I../shared -i-   -O7 -w all -w nonotinlined  -D_BUILDING_tracker=1 -DOPEN_TRACKER=1 -D_INCLUDES_CLASS_DEVICE_MAP=1 -D_SUPPORTS_RESOURCES=1 -D_SUPPORTS_FEATURE_SCRIPTING=1  -Dstd=""              -o obj.ppc/PoseView.o
### mwcc Compiler Warning :
#       SetHasPosesInClipboard(hasPosesInClipboard | fHasPosesInClipboard);
#                                                                        ^
#   implicit arithmetic conversion from 'int' to 'bool'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/tracker/PoseView.cpp"; Line 2863
#----------------------------------------------------------


mwcc -c QueryPoseView.cpp  -I./  -I../shared -i-   -O7 -w all -w nonotinlined  -D_BUILDING_tracker=1 -DOPEN_TRACKER=1 -D_INCLUDES_CLASS_DEVICE_MAP=1 -D_SUPPORTS_RESOURCES=1 -D_SUPPORTS_FEATURE_SCRIPTING=1  -Dstd=""                 -o obj.ppc/QueryPoseView.o
### mwcc Compiler Warning :
#       if ((buffer = (char *)malloc(info.size)) != NULL
#                                             ^
#   implicit arithmetic conversion from 'long long' to 'unsigned long'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/tracker/QueryPoseView.cpp"; Line 457
#----------------------------------------------------------


mwcc -c StatusWindow.cpp  -I./  -I../shared -i-   -O7 -w all -w nonotinlined  -D_BUILDING_tracker=1 -DOPEN_TRACKER=1 -D_INCLUDES_CLASS_DEVICE_MAP=1 -D_SUPPORTS_RESOURCES=1 -D_SUPPORTS_FEATURE_SCRIPTING=1  -Dstd=""                  -o obj.ppc/StatusWindow.o
### mwcc Compiler Warning :
#       AddLine(BPoint(bounds.left, bounds.bottom - 1.0),
#                                                       ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/tracker/StatusWindow.cpp"; Line 655
#----------------------------------------------------------
### mwcc Compiler Warning :
#       AddLine(BPoint(bounds.left + 1.0, bounds.top),
#                                                    ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/tracker/StatusWindow.cpp"; Line 657
#----------------------------------------------------------
### mwcc Compiler Warning :
#       AddLine(BPoint(bounds.right, bounds.top + 1.0),
#                                                     ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/tracker/StatusWindow.cpp"; Line 659
#----------------------------------------------------------
### mwcc Compiler Warning :
#       AddLine(BPoint(bounds.right - 1.0, bounds.bottom),
#                                                        ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/tracker/StatusWindow.cpp"; Line 661
#----------------------------------------------------------


mwcc -c CalendarMenuItem.cpp  -I./  -I../shared  -I../tracker -i-   -O7 -w all -w nonotinlined  -DDB_ADDONS -DOPEN_TRACKER=1 -D_INCLUDES_CLASS_DEVICE_MAP=1 -D_SUPPORTS_RESOURCES=1 -D_SUPPORTS_FEATURE_SCRIPTING=1 -D_SHOW_CALENDAR_MENU_ITEM=1 -Dstd=""                  -o obj.ppc/CalendarMenuItem.o
### mwcc Compiler Warning :
#       TitleHeight = ceil(fontHeight.ascent + fontHeight.descent + fontHeight.leading);
#                                                                                      ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/deskbar/CalendarMenuItem.cpp"; Line 159
#----------------------------------------------------------
### mwcc Compiler Warning :
#       t = ceil(fontHeight.ascent + fontHeight.descent + fontHeight.leading + kRowGap);
#                                                                                      ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/deskbar/CalendarMenuItem.cpp"; Line 163
#----------------------------------------------------------
### mwcc Compiler Warning :
#       fFontHeight = ceil(fontHeight.ascent);
#                                            ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/deskbar/CalendarMenuItem.cpp"; Line 164
#----------------------------------------------------------


mwcc -c WindowMenuItem.cpp  -I./  -I../shared  -I../tracker -i-   -O7 -w all -w nonotinlined  -DDB_ADDONS -DOPEN_TRACKER=1 -D_INCLUDES_CLASS_DEVICE_MAP=1 -D_SUPPORTS_RESOURCES=1 -D_SUPPORTS_FEATURE_SCRIPTING=1 -D_SHOW_CALENDAR_MENU_ITEM=1 -Dstd=""   -o obj.ppc/WindowMenuItem.o
### mwcc Compiler Warning :
#       Frame().Width() - contLoc.x - 3.0);
#                                        ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/deskbar/WindowMenuItem.cpp"; Line 126
#----------------------------------------------------------
### mwcc Compiler Warning :
#       rgb_color shadow = tint_color(menuColor, (B_NO_TINT + B_DARKEN_1_TINT) / 2.0);
#                                                                                   ^
#   implicit arithmetic conversion from 'double' to 'float'
#----------------------------------------------------------
    File "/CommonPPC/agmsmith/OpenTracker/opentracker/deskbar/WindowMenuItem.cpp"; Line 180
#----------------------------------------------------------

Re: [Opentracker-develop]Re: Nightly build tarball not up-to-date

[Axel <ax...@pi...>]

"Simon Taylor" <sim...@ga...> wrote:
> >Sure, I could also introduce a modifier key (which is 
> > probably a good idea), but I think keeping the button pressed for a 
> > little longer to get that calendar is an almost natural extension - 
> > 
> > even if you'll probably need some time to find it.
> Yes that seems sensible.
> 
> Would this also work the other way round? If deskbar is
> currently displaying date, you click it and it switches to time,
> hold button for a bit longer and calendar pops up?

That would be possible, but I've disabled this as I think it would be a 
bit strange - although on second thought, it might be a good idea to 
allow this in case the user has released the button already when 
looking at the date and now wants to get a calendar overview...

Bye,
   Axel.

[Opentracker-develop]Re: Nightly build tarball not up-to-date

[Simon T. <sim...@ga...>]

Hey Axel,

>> I've never seen any other app that uses this method - click and
>> hold to do something. I don't know how it's implemented (a
> 
> Oh, okay, go to your desktop, and hold the left mouse button for a 
> while and see what will happen. Also, this is the standard way to get a 
> context menu under MacOS X (besides pressing Ctrl while clicking the 
> one and lonely mouse button).

Yes, OK. However that is just a secondary way of accessing a
context menu. If holding left triggers the calendary, so should
simple click right I think.
 
>> deskbar replicant?), but I think the volume control in the
>> deskbar is a good example - it appears as soon as you click on
>> it, and disappears as soon as you release the mouse button. It's
>> not at all intrusive IMHO - I can't see any reason you would
>> click on it unless you wanted to change the volume.
>> 
>> If the main reason for this addition is to see for example, what
>> date it is next friday, I dont see why any more is needed -
>> click and hold, look at the date, let go.
> 
> Ahem, are you contradicting yourself or can I just not follow you? Are 
> you arguing against or for the way this feature is implemented?
> When you click on the time, the Deskbar will make the date visible for 
> a few seconds.

That's why I said I didn't know how it was implemented - now i
see it is accessed from the Time/Date view on the deskbar. I
thought you might have a seperate replicant (with a calendar
icon on it), in which case I thoght that a single click should
instantly cause the calendar to appear for the duration the
mouse button is held down.

>Sure, I could also introduce a modifier key (which is 
> probably a good idea), but I think keeping the button pressed for a 
> little longer to get that calendar is an almost natural extension - 
> even if you'll probably need some time to find it.

Yes that seems sensible.

Would this also work the other way round? If deskbar is
currently displaying date, you click it and it switches to time,
hold button for a bit longer and calendar pops up?

>> One of
>> the most annoying things on R5 for me is the fact that fonts <
>> 11 pts don't render bold any differently to roman.
> 
> Luckily, that's not the case in Dano/Zeta, nor will it be in Haiku. BTW 
> a good work-around to that problem is to buy a monitor with greater 
> resolution ;-)

Running 1024x768, but still with 10pt default, not sure why :)

> Bye,
>    Axel.

Simon
---------------------------------------------
Free POP3 Email from www.Gawab.com 
Sign up NOW and get your account @gawab.com!!

Re: [Opentracker-develop]Re: Nightly build tarball not up-to-date

[Axel <ax...@pi...>]

"Simon Taylor" <sim...@ga...> wrote:
> >> - It took me some time to figure how to display it (In fact, I had 
> > > to 
> >> look at TimeView.cpp to understand LMB must be kept pressed a few 
> >> seconds). On Fedora, the calendar appears at first LMB click. It's 
> >> more "intrusive" but it's easy to figure how it works.
> > At first, I had it in the right click menu; but that didn't look 
> > good, 
> > and, as you say, was kinda intrusive. And misplaced in my opinion.
> > BTW you only need to press the button (left or middle) less than 
> > one 
> > second - although that depends on the time you've chosen as a 
> > double-
> > click frame.
> I've never seen any other app that uses this method - click and
> hold to do something. I don't know how it's implemented (a

Oh, okay, go to your desktop, and hold the left mouse button for a 
while and see what will happen. Also, this is the standard way to get a 
context menu under MacOS X (besides pressing Ctrl while clicking the 
one and lonely mouse button).

> deskbar replicant?), but I think the volume control in the
> deskbar is a good example - it appears as soon as you click on
> it, and disappears as soon as you release the mouse button. It's
> not at all intrusive IMHO - I can't see any reason you would
> click on it unless you wanted to change the volume.
> 
> If the main reason for this addition is to see for example, what
> date it is next friday, I dont see why any more is needed -
> click and hold, look at the date, let go.

Ahem, are you contradicting yourself or can I just not follow you? Are 
you arguing against or for the way this feature is implemented?
When you click on the time, the Deskbar will make the date visible for 
a few seconds. Sure, I could also introduce a modifier key (which is 
probably a good idea), but I think keeping the button pressed for a 
little longer to get that calendar is an almost natural extension - 
even if you'll probably need some time to find it.

> > be_bold_font may have a different size, so this could look a bit 
> > odd. 
> > Anyway, I'll see how I could make it more prominent. Maybe an 
> > underline 
> > would do.
> Yes, or a light box around the current day or something. One of
> the most annoying things on R5 for me is the fact that fonts <
> 11 pts don't render bold any differently to roman.

Luckily, that's not the case in Dano/Zeta, nor will it be in Haiku. BTW 
a good work-around to that problem is to buy a monitor with greater 
resolution ;-)

Bye,
   Axel.

Re: [Opentracker-develop]Deskbar's Calendar

[Axel <ax...@pi...>]

Benjamin Thery <ben...@bu...> wrote:
> > be_bold_font may have a different size, so this could look a bit 
> > odd. 
> > Anyway, I'll see how I could make it more prominent. Maybe an 
> > underline 
> > would do.
> What about a lighter background for 'Today'?
> See:
[...]

Thanks! I've changed it a bit, and have commited it to the repository.

Bye,
   Axel.

[Opentracker-develop]Deskbar's Calendar

[Benjamin T. <ben...@bu...>]

Axel D=F6rfler wrote:

> Although I don't know that one, all I wanted was something very simple=20
> :-)
> I was thinking about adding a way to see the calendar for other months=20
> than the current one, but I didn't find a simple way, so I left that=20
> out.

Here is a screenshot of the calendar in Fedora:

http://oxben.free.fr/tmp/fedora-calendar.png

It can display the calendar for other months and (other years).

>>- Finally something specific to my settings: I cannot distinguish the=20
>>current day from the other days. Indeed, my plain font is quite small=20
>>(swis721 BT Roman 10) and therefore its bold face is not different=20
>>from=20
>>the Roman face. May be you could also alter the background color of=20
>>the=20
>>current day (kind of inverse video... not sure if this is feasible=20
>>easily) or you could simply use be_bold_font to draw it?
>=20
>=20
> be_bold_font may have a different size, so this could look a bit odd.=20
> Anyway, I'll see how I could make it more prominent. Maybe an underline=
=20
> would do.

What about a lighter background for 'Today'?
See:

http://oxben.free.fr/tmp/deskbar-calendar-01.png

void
CalendarMenuItem::DrawContent()
...
rgb_color todayColor =3D tint_color(Menu()->LowColor(), B_LIGHTEN_2_TINT)=
;
...
if (today) {
	// draw a rectangle around today's day number
	Menu()->SetHighColor(todayColor);
	Menu()->FillRect(BRect(point.x - 2, point.y - 1 - fFontHeight,
				point.x + width + 1, point.y + 2));
	Menu()->SetHighColor(dayColor);
}

Menu()->DrawString(text, point);

if (today)
	Menu()->SetFont(be_plain_font);
}
...

Ben

[Opentracker-develop]Re: Nightly build tarball not up-to-date

[Simon T. <sim...@ga...>]

>> - It took me some time to figure how to display it (In fact, I had to 
>> look at TimeView.cpp to understand LMB must be kept pressed a few 
>> seconds). On Fedora, the calendar appears at first LMB click. It's 
>> more 
>> "intrusive" but it's easy to figure how it works.
> 
> At first, I had it in the right click menu; but that didn't look good, 
> and, as you say, was kinda intrusive. And misplaced in my opinion.
> BTW you only need to press the button (left or middle) less than one 
> second - although that depends on the time you've chosen as a double-
> click frame.

I've never seen any other app that uses this method - click and
hold to do something. I don't know how it's implemented (a
deskbar replicant?), but I think the volume control in the
deskbar is a good example - it appears as soon as you click on
it, and disappears as soon as you release the mouse button. It's
not at all intrusive IMHO - I can't see any reason you would
click on it unless you wanted to change the volume.

If the main reason for this addition is to see for example, what
date it is next friday, I dont see why any more is needed -
click and hold, look at the date, let go.

> be_bold_font may have a different size, so this could look a bit odd. 
> Anyway, I'll see how I could make it more prominent. Maybe an underline 
> would do.

Yes, or a light box around the current day or something. One of
the most annoying things on R5 for me is the fact that fonts <
11 pts don't render bold any differently to roman.

>> Great work.
>> It's nice to see some new features added to OpenTracker.
> 
> And yet there is still plenty to be done :)
> 
> Bye,
>    Axel.

Simon
---------------------------------------------
Free POP3 Email from www.Gawab.com 
Sign up NOW and get your account @gawab.com!!

Re: [Opentracker-develop]Nightly build tarball not up-to-date

[Axel <ax...@pi...>]

Benjamin Thery <ben...@bu...> wrote:
> Thanks for the updated tarball!
> I've just compiled the new OpenTracker and tested the calendar view.
> The calendar applets in my Fedora Core installation is something I 
> really like and I wanted to see what you implemented for OpenTracker.

Although I don't know that one, all I wanted was something very simple 
:-)
I was thinking about adding a way to see the calendar for other months 
than the current one, but I didn't find a simple way, so I left that 
out.

> Here is my feedback:
> 
> - First, it looks good to me and I think its a useful addition to the 
> Deskbar!

You're welcome :)

> - It took me some time to figure how to display it (In fact, I had to 
> look at TimeView.cpp to understand LMB must be kept pressed a few 
> seconds). On Fedora, the calendar appears at first LMB click. It's 
> more 
> "intrusive" but it's easy to figure how it works.

At first, I had it in the right click menu; but that didn't look good, 
and, as you say, was kinda intrusive. And misplaced in my opinion.
BTW you only need to press the button (left or middle) less than one 
second - although that depends on the time you've chosen as a double-
click frame.

> - Months, year and days of the week appear a bit to light on my 
> display 
> compared to the background.

That's actually intended, even if I might have done too much - since 
it's always the current month, and you should know the weekdays, that 
information is not so important compared to the rest.
I'll try to make it a bit darker without coming to close to the day 
numbers.

> - Here, in France, weeks start on monday. Do you think it's worth to 
> add 
> a new deskbar setting (after "European date") or will this be handled 
> automagically by the locale kit one day? ;-)

That is supposed to be fixed with the Locale Kit, yes; then all those 
settings will be removed from the Deskbar anyway. Since I intend to 
ship the Locale Kit with the next Tracker release (just don't ask when 
;-)), there is no need to change that now.

> - Finally something specific to my settings: I cannot distinguish the 
> current day from the other days. Indeed, my plain font is quite small 
> (swis721 BT Roman 10) and therefore its bold face is not different 
> from 
> the Roman face. May be you could also alter the background color of 
> the 
> current day (kind of inverse video... not sure if this is feasible 
> easily) or you could simply use be_bold_font to draw it?

be_bold_font may have a different size, so this could look a bit odd. 
Anyway, I'll see how I could make it more prominent. Maybe an underline 
would do.

> Great work.
> It's nice to see some new features added to OpenTracker.

And yet there is still plenty to be done :)

Bye,
   Axel.

Re: [Opentracker-develop]Nightly build tarball not up-to-date

[Benjamin T. <ben...@bu...>]

Thanks for the updated tarball!
I've just compiled the new OpenTracker and tested the calendar view.
The calendar applets in my Fedora Core installation is something I=20
really like and I wanted to see what you implemented for OpenTracker.

Here is my feedback:

- First, it looks good to me and I think its a useful addition to the=20
Deskbar!

- It took me some time to figure how to display it (In fact, I had to=20
look at TimeView.cpp to understand LMB must be kept pressed a few=20
seconds). On Fedora, the calendar appears at first LMB click. It's more=20
"intrusive" but it's easy to figure how it works.

- Months, year and days of the week appear a bit to light on my display=20
compared to the background.

- Here, in France, weeks start on monday. Do you think it's worth to add=20
a new deskbar setting (after "European date") or will this be handled=20
automagically by the locale kit one day? ;-)

- Finally something specific to my settings: I cannot distinguish the=20
current day from the other days. Indeed, my plain font is quite small=20
(swis721 BT Roman 10) and therefore its bold face is not different from=20
the Roman face. May be you could also alter the background color of the=20
current day (kind of inverse video... not sure if this is feasible=20
easily) or you could simply use be_bold_font to draw it?

Great work.
It's nice to see some new features added to OpenTracker.

Ben

Axel D=F6rfler wrote:
> Benjamin Thery <ben...@bu...> wrote:
>=20
>>It seems to me that the supposed "nightly built" sources tarball=20
>>available on the OpenTracker site is not up-to-date=20
>>(http://opentracker.sourceforge.net/downloads.html)
>>. The youngest file in this tarball dates from October 2004.
>>
>>Can anyone confirm this? Am I forced to extract the whole source tree=20
>>from CVS to build the latest OpenTracker with Axel's new calendar=20
>>view?
>=20
>=20
> Yes, I can confirm this. It's just another SourceForge related problem;=
=20
> project cronjobs are not executed anymore (a periodically ran script=20
> that updates the sources package).
> Anyway, I've rebuild the archive manually just now, so you should be=20
> able to retrieve the latest archive again, at least for the next couple=
=20
> of hours/days :-)
> I will look into providing a better solution to this. Thanks for the=20
> note!
>=20
> Bye,
>    Axel.
>=20
>=20
>=20
> -------------------------------------------------------
> The SF.Net email is sponsored by: Beat the post-holiday blues
> Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
> It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
> _______________________________________________
> OpenTracker-develop mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opentracker-develop
>=20


--=20
B e n j a m i n   T h e r y  - BULL/DT/OSw R&D/AIX development

  mailto:Ben...@bu...
     tel: (047629)7133
    http://www.bull.com

Re: [Opentracker-develop]Nightly build tarball not up-to-date

[Axel <ax...@pi...>]

Benjamin Thery <ben...@bu...> wrote:
> It seems to me that the supposed "nightly built" sources tarball 
> available on the OpenTracker site is not up-to-date 
> (http://opentracker.sourceforge.net/downloads.html)
> . The youngest file in this tarball dates from October 2004.
> 
> Can anyone confirm this? Am I forced to extract the whole source tree 
> from CVS to build the latest OpenTracker with Axel's new calendar 
> view?

Yes, I can confirm this. It's just another SourceForge related problem; 
project cronjobs are not executed anymore (a periodically ran script 
that updates the sources package).
Anyway, I've rebuild the archive manually just now, so you should be 
able to retrieve the latest archive again, at least for the next couple 
of hours/days :-)
I will look into providing a better solution to this. Thanks for the 
note!

Bye,
   Axel.

[Opentracker-develop]Nightly build tarball not up-to-date

[Benjamin T. <ben...@bu...>]

Hi all,

It seems to me that the supposed "nightly built" sources tarball 
available on the OpenTracker site is not up-to-date 
(http://opentracker.sourceforge.net/downloads.html)
. The youngest file in this tarball dates from October 2004.

Can anyone confirm this? Am I forced to extract the whole source tree 
from CVS to build the latest OpenTracker with Axel's new calendar view?

Thanks.

Ben

Re: [Opentracker-develop]Deskbar broken?

[Ingo W. <bon...@cs...>]

On 2004-10-26 at 15:35:12 [+0200], Stefano Ceccherini wrote:
[...]
> >PS: I'm using Oliver's brand new gcc, though it would surprise me, if that
> >would be the reason.
> 
> You can never know. I'm using the geekgadgets compiler, and as I said it 
> works here.

Whoa, good thinking. With the GG compiler it works here, too. Same goes for 
the gcc 2.95.3 with binutils 2.11.2 and the previous binutils 2.15 version 
(from 2004-10-03). So it really seems to be the new compiler. Poor Oliver...

CU, Ingo

Re: [Opentracker-develop]Deskbar broken?

[Stefano C. <bur...@fr...>]

>is anyone running the Deskbar from the current HEAD? I ask, 

I am.



>because I now 

>do and it reproducibly crashes when I try to open the Be 



>menu. In 

>TBeMenu::AddStandardBeMenuItems(), line 283. There



  be_app->WindowAt(1)->Feel()



>is invoked, but `be_app->WindowAt(1)' returns NULL. I find it a bit scary 

>to access a window that way, but the line exists since the first version in 

>the repository, so it must have worked once (and maybe still does for 

>others?).



It works here, actually.



>Any ideas?



CU, Ingo



>PS: I'm using Oliver's brand new gcc, though it would >surprise me, if that 

>would be the reason.





You can never know. I'm using the geekgadgets compiler, and as I said it works here.



Stefano Ceccherini aka Jack Burton
---------------------------------------------------------------
Scegli il tuo dominio preferito e attiva la tua email! Da oggi
l'eMail di superEva e' ancora piu' veloce e ricca di funzioni!
http://webmail.supereva.it/new/
---------------------------------------------------------------

[Opentracker-develop]Deskbar broken?

[Ingo W. <bon...@cs...>]

Howdy,

is anyone running the Deskbar from the current HEAD? I ask, because I now 
do and it reproducibly crashes when I try to open the Be menu. In 
TBeMenu::AddStandardBeMenuItems(), line 283. There

  be_app->WindowAt(1)->Feel()

is invoked, but `be_app->WindowAt(1)' returns NULL. I find it a bit scary 
to access a window that way, but the line exists since the first version in 
the repository, so it must have worked once (and maybe still does for 
others?).

Any ideas?

CU, Ingo

PS: I'm using Oliver's brand new gcc, though it would surprise me, if that 
would be the reason.

Re: [Opentracker-develop]Feature request. (fwd)

[Simon T. <sim...@nt...>]

Sf don't like my gawab account...don't know if this will make it 
through (not subscribed with this addy) or if it will have to be let 
through by an admin (sorry Axel!)

------ Forwarded Message: ------
To: ope...@li...
From: "Simon Taylor" <sim...@ga...>
Subject: Re: [Opentracker-develop]Feature request.
Date: Thu, 30 Sep 2004 17:00:04 +0100 BST

How about a scheme where you can hold down shift and then define 
the full order=3F First column you click on becomes 1st sort column, 
click again to reverse sort order, click next column to select 2nd 
sort depth level, etc.

It would mean you could set however many sorting levels you wanted 
fairly intuitively just by clicking on the column titles, without 
having to delve into sub menus or anything.

Clicking without shift would clear all the current sort order and just 
sort by the one column clicked on.

Simon


> "Axel D=F6rfler" <ax...@pi...> wrote:
> > I would have the following suggestions:
> > 1) add a menu to the attributes title where you can select the 
> > ordering
> Attribute titles has a attributes menu right now (accessed by right 
> mouse button). I think its wouldn't be a loss to replace them with 
> column menu or somethig. You can add more options when needed. I like 
> it.
> 
> "Axel D=F6rfler" <ax...@pi...> wrote:
> > I'm not sure if there is enough need to introduce a third sort 
> > order, 
> > though.
> BFS is database like. Let's extend tracker to take more advantage 
> over 
> that database. It will be useful, I'm sure. Not only 3 deep order. I 
> suggest infinite deep order (maded once - last forever). So...
> 
> 
> Stefano Ceccherini <bur...@fr...> wrote:
> > PL-SQL developer (a tool for doing sql development for windows) 
> > uses 
> > this kind of interface to be able to order columns, and I find it 
> > nice. 
> If you click on a column, it get sorted. If you click on it again, 
> the 
> order gets reversed. If you click on it again, it gets unsorted. The 
> bad thing is that to sort with more than one order, you just click on 
> another column's title, while with tracker we need to keep SHIFT 
> pressed, IIRC. I think it's better how we did it.
> > 
> > 
> > 
> > And it uses a triangle to show the sort order. In case of multiple 
> > sort order, only the second column has a triangle, the first one 
> > has a 
> circle.
> 
> ...I like it. But wee will need more geometric symbols. I suggest 
> roman 
> style numbers (it could be customizable in menu mentioned above) 
> representing deep level.
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement 
> on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc=5Fcontest.php
> =5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F
> OpenTracker-develop mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opentracker-develop