Beta 6 -JaPBe- Editor functions crawl with 96% CPU usage

[Randy Walker]

As I was editing my code to accomodate PB4beta6 I encountered extremely painfull delays in many editor functions ranging from simple cursor movement to text searches.

These delays "initially" occured in both the PB editor and the jaPB editor. Now that I have all the code change modifications done the PB editor seems almost normal (still many delays compared to 3.94 final).

The jaPB editor however continues to exibit excruciating and tourturous delays. Pretty much as it did from the start of my code conversion. I don't think jaPBe is any of Fred's concern directly, but I didn't know where else to mention it. Thought it wouldn't hurt to mention it here since the standard PB4 editor was exibiting the same behavior too initially.

Some symptoms that occured in both editors and remain in jaPBe:

Fold All/Unfold All - Just go to bed and come back in the morning!!! (no exageration - hours on end w/20000 lines of code) This is the worst offender. CPU goes to 96% and editor will not respond until it completes the fold/unfold function. Search behavior almost as bad before I worked out all code incompatibilities.

Scrolling cursor - If cursor lands at beginning line of any procedure, the editor pauses milliseconds to several full seconds before allowing further movement. Varies - maybe up to 10 seconds (on PIII/600Mhz) This applies to all 4 cursor movements, page up/down, Ctrl+End and Ctrl+Home.

Also using scrollbar to shift text to begiining or end of listing works fine but try clicking the text to position cursor into view may take several seconds to appear and become active.

Scrolling line at a time using up/down arrows exibits random delays about 50% of the time. Elsewise cursor respone seems normal.

So ... Is it just me?

Thanks

[gnozal]

Some symptoms that occured in both editors and remain in jaPBe:

Fold All/Unfold All - Just go to bed and come back in the morning!!! (no exageration - hours on end w/20000 lines of code) This is the worst offender. CPU goes to 96% and editor will not respond until it completes the fold/unfold function. Search behavior almost as bad before I worked out all code incompatibilities.

I will try to make it a little faster and less freezing

Scrolling cursor - If cursor lands at beginning line of any procedure, the editor pauses milliseconds to several full seconds before allowing further movement. Varies - maybe up to 10 seconds (on PIII/600Mhz) This applies to all 4 cursor movements, page up/down, Ctrl+End and Ctrl+Home.

No problems here with my PII-300MHz

[Fangbeast]

Strange things have ahappened to me in the Beta 6 editor as well.

1. When typing in a pb keyword, the case is not corrected till you move to a new line. (and sometiles the colour as well!!)

2. Sometimes scrolling through the code, you might encounter up to 100 apparently empty lines that suddenly appear when you stop scrolling. I thought I had lost the lines!

[Randy Walker]

Scrolling cursor - If cursor lands at beginning line of any procedure, the editor pauses milliseconds to several full seconds before allowing further movement. Varies - maybe up to 10 seconds (on PIII/600Mhz) This applies to all 4 cursor movements, page up/down, Ctrl+End and Ctrl+Home.

No problems here with my PII-300MHz

I don't know if this is a clue but I just removed this from the tail end of my listing:

Code: Select all

; IDE Options = PureBasic v4.00 - Beta 6 (Windows - x86)
; CursorPosition = 1087
; FirstLine = 649
; Folding = QIAAAAAAAAAAAAAAAAAAAAAAAAAAA5
; IDE Options = PureBasic v4.00 - Beta 6 (Windows - x86)
; CursorPosition = 20311
; Folding = AAAAAAAAAAAAAAAAAAAAAAAAAAAAA5

I'm guessing there are nearly 200 procedures in my listing. Same cursor delay occurs if I position the cursor on the first line of a procedure near the center of the listing, scrollbar down to the end of the listing and click to bring the cursor into view. Same thing going the other way. Top line of a procedure near the bottom of the listing, scrollbar to the top of the listing and click to bring mouse into view. Further away the procedure is , the loonngger the delay.

Strange thing is it seems to "learn" my testing strategy and adapts to an almost livable response. I go back to 3.94 to see if I'm crazy, reboot, come back and same results. PB4beta6 is Ok and jaPBeV3 has severe delays, whereas 3.94 via jaPBe still works great. Try the new jaPBe again and severe delays that again seem to adapt. Strange.

[gnozal]

Strange thing is it seems to "learn" my testing strategy and adapts to an almost livable response. I go back to 3.94 to see if I'm crazy, reboot, come back and same results. PB4beta6 is Ok and jaPBeV3 has severe delays, whereas 3.94 via jaPBe still works great. Try the new jaPBe again and severe delays that again seem to adapt. Strange.

Strange. I don't have this kind of problems with jaPBe V3, even with a lot of procedures [100] and lines [8600] (with a PII-300MHz).

[Trond]

I have noticed in Beta 6 that sometimes a search and replace all is almost instant, while sometimes it takes several seconds on the same source file.

[Randy Walker]

I have noticed in Beta 6 that sometimes a search and replace all is almost instant, while sometimes it takes several seconds on the same source file.

Exactly the same results here ... except again, if my cursor happens to be sitting on top of the fiorst line of "any" procedure, the search will often have an additional delay.

[Randy Walker]

New senario that might help you duplicate the cursor delay.

I left jaPBe open all night sitting with my code loaded up and today and it was acting a little better. Then I did something I had never tried since converting my code to 4.0 syntax. I ran my code through the degubber (F6). Then jaPBe immediately stopped with all the extreme delays. Seemingly totally back to the old behaviour I was use to seeing.

BUT, then I decided to test folding, so I unfolded my entire listing (Alt+Shift+D) which took some time as it normally does with 200 procs and 20000 lines. Then I folded the entrie listing (Alt+D) which may have taken longer than before (not sure). After the folding completed, the editor was back to it's torturous ways. CPU usage pegged again and wondering if my cursor would ever return. Again, the more I try to evaluate it, the less it misbehaves. Tryed the Debugger trick again and it seemd to have no affect this time. So strange.

[freak]

Please discuss the jaPBe problems somewhere else. This forum is for PureBasic bugs.

[Randy Walker]

Please discuss the jaPBe problems somewhere else. This forum is for PureBasic bugs.

Hi Freak ... Trying to be conciencious so hope you will pardon my thread here. Originally, I was having issues in both editors although they tended to linger more in jaPBe (evidently my own fault). Could you suggest an appropriate forum for jaPBe posts in the future?

NOTE TO ALL:
Sorry for this whole thread - problems all went away after installing the latest jaPBe release. My bad

[Randy Walker]

Delay problems still occuring in jaPBe though not severe as I had seen before. Only very annoying now. Most annoying is in folding and unfolding. I click, wonder if I did because nothing happens, click again and then watch my procedure unfold and then fold back up because the first click was good. Click again and wait 2 seconds before it unfolds again as I wanted.

Haven't noted the delay in the native editor, and Berinko didn't comment on a proper place for me to post so opted for the original post. (Trying to keep it together anyway.)

[Trond]

Posting about jaPBe here is like posting about Linux at a Windows forum.

[blueznl]

not sure about that, japbe is for many users an essential part of purebasic, and it certainly was the reason i picked up purebasic again

so i'd say give japbe the advantage of the doubt

on top of that, japbe is WRITTEN IN purebasic, that makes it defintely on topic, unless we're no longer allowed to discuss our own bugs...

[Berikco]

Sure, this General discussion is for all purebasic stuff thats dont fit in the other forums.
Bug reports is not for JaPBe, and it does not belong in off topic, announcements etc...So here is the perfect place

[Randy Walker]

Thanks for the direction Berikco