Page 1 of 1

IDE repeatedly crashing! PB 4.2 beta 6.

Posted: Tue May 20, 2008 10:37 am
by srod
@Timo : please see my post below for details of the crash I am getting.

============================================


Hi,

I am finding with several tabs open (about 12 at the moment) in the IDE, that it is horrendously slow in closing down - can take two/three minutes sometimes. This has been happening for a couple of 4.2 beta versions now.

No problem with PB 4.1.

Posted: Tue May 20, 2008 11:41 am
by Tranquil
I encountered this problem even on some projects of mine. It takes more longer to shout them down as some versions earlier.

Maybe it has something ToDo with this issue:
http://www.purebasic.fr/english/viewtopic.php?t=32407

... and the system frees the resources itself? Just a thought.

Posted: Tue May 20, 2008 12:49 pm
by srod
Could be. I'll see if PB 4.2 final fixes the problem. :)

Posted: Sat May 31, 2008 9:44 am
by srod
Been using 4.2 final now since it was released and the IDE problems persist. With about 10 tabs open the thing can take an age to shut down and that's if it doesn't crash first!

XP pro.

Posted: Sat May 31, 2008 11:38 am
by Mistrel
Can you be more specific about how long it takes? I usually have a lot of tabs open as well and I don't remember having a problem like this.

Are you sure it doesn't happen in the 4.10 IDE?

Posted: Sat May 31, 2008 3:54 pm
by DoubleDutch
SRod: Check your drives - it may a sign that something else is wrong.

Posted: Sat May 31, 2008 5:06 pm
by dhouston
I see the same ultra-slow shutdown with 4.20 final and a single tab. It happens on two Windows machines so I don't think it's hardware related. It shuts down nearly instantly under Ubuntu 8.04 but takes an inordinate amount of time to load whether under Windows or Linux. It's so bad that had it been this slow when I first evaluated PB, I would never have considered it.

Also, while the compile speed appears to be much faster, there's now a very long delay between the time the little compiling window goes away and the executable starts.

I've been extremely happy with PB until 4.20 final which appears quite buggy.

There are no significant changes to my code between 4.10 & 4.20 that would account for the speed change. Under Windows, I switched from MVCOM to the native serial port library but under Linux, the serial port library is DOA so I'm still using API functions for serial comms.

Posted: Sat May 31, 2008 5:45 pm
by Fred
Well, we never experienced this, so that's why the bug is still here. Once we can reproduce it, we will indeed fix it... No need to call a release 'quite buggy' because you encountered a few bugs over the tousand of commands available.

Posted: Wed Jun 04, 2008 11:00 pm
by akee
buggy - a recreational vehicle with large wheels, and wide tires, designed for use on sand dunes or beaches.
Quite buggy...? Are you kidding? PB is like a my virtual super duper buggy on my PC. Fast and loads of fun... :p

Posted: Sun Jun 08, 2008 10:56 am
by srod
Just to say that there seems to be no problem with the slow closedown of the IDE with a new computer running Vista 32. Could be the fact that this machine is almost 3* faster than my old one though I guess! :)

Posted: Thu Jun 12, 2008 1:27 pm
by srod
@Freak; on having about 16 tabs open in the IDE, I was experiencing some very strange behaviour! I know that 16 tabs is a little extreme, but I was testing something out. Basically I had identical source files open, but from two different folders in order to contrast my application's behaviour with an older version.

I suspect that the ide was not saving my alterations to certain source files.

I was checking to see if any constants were redundant by commenting them out etc. I would comment out a constant, save the source include file in question, run the program and it would not report an error. Run it again and the compiler would then corretly report a 'undeclared constant' error. This continued on and on, allowing me to double check this strange occurrence. For example, I would reinstate the missing constant, save the file, run the app and an error would ensure until I ran it again.

I can repeat this problem, although only when there are a lot of tabs open.

As I say, the constant file was open in one tab, and an identical one was opened (but from another folder) in a separate tab.

Not a massive problem, but one which perhaps you should be aware of.

(Vista 32, PB 4.2, 2 GB ram.)

Posted: Sat Jun 21, 2008 3:02 pm
by srod
Right, I am now getting a reproduceable crash with the IDE (Vista premium).

Open about 6 tabs in the IDE, including 2 separate copies of the following program :

Code: Select all

If OpenWindow(0, 0, 0, 355, 180, "TreeGadget", #PB_Window_SystemMenu | #PB_Window_ScreenCentered) And CreateGadgetList(WindowID(0))
  TreeGadget(0, 10, 10, 160, 160)                                         ; TreeGadget standard
  ListIconGadget(1, 180, 10, 160, 160, "",100,#PB_ListIcon_GridLines|#PB_ListIcon_HeaderDragDrop)  ; TreeGadget with Checkboxes + NoLines
  AddGadgetColumn(1,1,"jj",100)

  For ID = 0 To 1
    For a = 0 To 0
      AddGadgetItem (ID, -1, "Normal Item "+Str(a), 0, 0) ; if you want to add an image, use 
      AddGadgetItem (ID, -1, "Node "+Str(a), 0, 0)        ; ImageID(x) as 4th parameter
      AddGadgetItem(ID, -1, "Sub-Item 1", 0, 1)    ; These are on the 1st sublevel  
      AddGadgetItem(ID, -1, "Sub-Item 2", 0, 2)
      AddGadgetItem(ID, -1, "Sub-Item 3", 0, 3)
    Next
  Next
  SetActiveGadget(0)
  SetGadgetItemState(0,1, #PB_Tree_Expanded)

  Repeat : Until WaitWindowEvent() = #PB_Event_CloseWindow
EndIf
Run these two copies, select an item in the respective tree gadgets and then close the two running programs. On my system, the IDE then thinks one of the two programs is still running and no amount of 'Kill program's alter that. Close the IDE at this point and the IDE crashes out. Actually I keep getting a repeated 'Application is not responding' message and have to use the task manager to close the IDE.

I don't know if this is just Vista or just my system etc?

I sometimes have to repeat this 3 or 4 times before I get a crash (just in case someone tries to reproduce this), but I always succeed in getting a crash! :)