Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[gtoso]

Hi,
on Mac OSX Leopard/Intel I see some visualization defect in SVN r9623 and r9641 (r9602 ok):
kad graph appears for some seconds in Transfers, Networks and Searches.

See here.

EDIT: Tested with my packages.
EDIT2: The same on aMuleGUI.

[Stu Redman]

Cool recording!

Please try 9642, I've already changed something regarding startup of the OScopeCtrls there. Maybe this takes care of your problem too.

[gtoso]

Cool recording!

Please try 9642, I've already changed something regarding startup of the OScopeCtrls there. Maybe this takes care of your problem too.

Tried: it fix it only in part:
at the beginning and until I show a graph (kad or statistics) it's ok,
then as before or even worse: statistics tab, then switch to searches tab and statistics graphs appears for few seconds leaving Name field dirty.

EDIT: After show a graph it's exactely as before, I had not tried this before.

[Stu Redman]

EDIT: After show a graph it's exactely as before, I had not tried this before.

What do you mean with "before" - like in 9623(bad) or like in 9602(ok) ?

[gtoso]

In summary, no more problems at startup or in Transfers tab.
Problems only in Networks and Searches tabs.
In Networks tab begin only after show Kad graph.

For the detail see this video.

[gtoso]

EDIT: After show a graph it's exactely as before, I had not tried this before.

What do you mean with "before" - like in 9623(bad) or like in 9602(ok) ?

Like in 9623(bad) 

[Stu Redman]

Just remove this line (OScopeCtrl.cpp : 292)

Code: [Select]

void COScopeCtrl::OnPaint(wxPaintEvent& WXUNUSED(evt))
{
// m_onPaint = true;
and tell me if it helps. (It should redisable the whole timer logic.)

[gtoso]

Good, it works fine!!!

What does involve this modify?

[Stu Redman]

There is a logic in the OScopeCtrl using a timer to prevent permanent rescale and redraw when the window is resized. This logic was broken, and I fixed it. Looks like it has undesirable effects on Mac.

Let's try something else. Does this work too (with the line from above present again) ?

Code: [Select]

Index: OScopeCtrl.cpp
===================================================================
--- OScopeCtrl.cpp (revision 9642)
+++ OScopeCtrl.cpp (working copy)
@@ -527,7 +527,7 @@
  bRecreateGrid |= bInvalidateGrid;
  // To prevent startup problems don't start timer logic until
  // a native OnPaint event has been generated.
- if (m_onPaint) {
+ if (m_onPaint && IsShown()) {
  bRecreateAll  |= bInvalidateGraph && bInvalidateGrid;
  timerRedraw.Start(100, true); // One-shot timer
  }


[gtoso]

There is a logic in the OScopeCtrl using a timer to prevent permanent rescale and redraw when the window is resized. This logic was broken, and I fixed it. Looks like it has undesirable effects on Mac.

Let's try something else. Does this work too (with the line from above present again) ?

I'm going to try.

[gtoso]

Let's try something else. Does this work too (with the line from above present again) ?

No, doesn't work 
Same behavior  of 9642 without patch.

[Stu Redman]

OK, fixed in 9645 (disabling timer logic for Mac).

[gtoso]

OK, fixed in 9645 (disabling timer logic for Mac).

Thanks, I hope that soon or later we find a better solution.

I just noticed other problems, that I quite sure there aren't in 9602.
Quickly (I have no time now):
"amuled -f" seems to hang after fork, amulecmd blocks trying to connect to it
"amuled" works
amulecmd doesn't show connection status:

Code: [Select]

aMulecmd$ status
 > Download: 0 bytes/sec
 > Upload: 13 bytes/sec
 > Client in coda: 0
 > Fonti totali: 0
aMulecmd$