help: how to debug script interrupting? - page 17 - General Discussion and Assistance - CHDK Forum supplierdeeply

help: how to debug script interrupting?

  • 224 Replies
  • 43457 Views
*

Offline reyalp

  • ******
  • 13619
Re: help: how to debug script interrupting?
« Reply #160 on: 02 / May / 2013, 23:22:23 »
Advertisements
From  some discussion in IRC with waterwingz:

We don't really know how camera settings affect the likelihood of hitting the bug.  All else being equal, if you use settings like AE lock and AF lock to speed up the half shoot cycle, it will spend less time sleeping in half shoot and so have less time to hit the bug.

On D10, I've used the following camera settings
P mode
both MF and "focus at infinity" mode for different runs (maybe regular AF too, not sure...)
AE was normal, with AF assist turned off.
ISO 400
whitebalance daylight
IS = "shoot only" (because I kept forgetting to turn it off)
MF point zoom ON
AF point zoom ON
I contrast off
review off
Light is generally low (indoor, indirect window and/or incandescent light)

I don't know if any of these settings are relevant.

On the A540, I used P mode, MF, ISO 400. No IS on that camera.

script settings (from last D10 run)

NumofShots 10000
delay 1
deley step 500
mode accurat
bv correction 0
disable shoot on
log on
get_shooting sleep 10
buffer log on
verbose log off
debug info off
Don't forget what the H stands for.

Re: help: how to debug script interrupting?
« Reply #161 on: 02 / May / 2013, 23:27:16 »
rerunning the tests ... it takes a lot longer with AF and AE enabled ... but here's my "team" hard at work ..



gonna have to dig out the external supplies .. battery mode won't get me to 10,000 cycles with the new test mode
« Last Edit: 02 / May / 2013, 23:59:46 by waterwingz »
Ported :   A1200    SD940   G10    Powershot N    G16

*

Offline reyalp

  • ******
  • 13619
Re: help: how to debug script interrupting?
« Reply #162 on: 03 / May / 2013, 00:05:49 »
Some info from the last noshoot run on the d10

8 errors in 4942 shoot cycles, when I ended the run early.

err   elapsed timedtshotds
start00:00:00   00:00:000   0
1   00:34:01   00:34:0114041404
2   00:37:20   00:03:191574170
3   00:47:29   00:10:091994420
4   00:51:39   00:04:102188194
5   00:55:03   00:03:242333145
6   01:16:50   00:21:473242909
7   01:26:35   00:09:453647405
8   01:55:47   00:29:1248581211
end   01:57:48   00:02:01494284

ccd temp was ~50c throughout most of the run.
Don't forget what the H stands for.

Re: help: how to debug script interrupting?
« Reply #163 on: 03 / May / 2013, 00:17:23 »
Some info from the last noshoot run on the d10
8 errors in 4942 shoot cycles, when I ended the run early.
Are you running a modified CHDK version that continues after the error?   Maybe I missed something but I've been assuming the autobuild versions will halt/crash at the first occurance of the error?
Ported :   A1200    SD940   G10    Powershot N    G16


*

Offline reyalp

  • ******
  • 13619
Re: help: how to debug script interrupting?
« Reply #164 on: 03 / May / 2013, 00:23:43 »
Are you running a modified CHDK version that continues after the error?   Maybe I missed something but I've been assuming the autobuild versions will halt/crash at the first occurance of the error?
Yes, I'm running my debug patch that takes dumps and ignores the error. However, running stock should be fine, it will just die with a null or otherwise bogus error message.
Don't forget what the H stands for.

*

Offline philmoz

  • *****
  • 3376
    • Photos
Re: help: how to debug script interrupting?
« Reply #165 on: 03 / May / 2013, 07:00:51 »
From  some discussion in IRC with waterwingz:

We don't really know how camera settings affect the likelihood of hitting the bug.  All else being equal, if you use settings like AE lock and AF lock to speed up the half shoot cycle, it will spend less time sleeping in half shoot and so have less time to hit the bug.

On D10, I've used the following camera settings
P mode
both MF and "focus at infinity" mode for different runs (maybe regular AF too, not sure...)
AE was normal, with AF assist turned off.
ISO 400
whitebalance daylight
IS = "shoot only" (because I kept forgetting to turn it off)
MF point zoom ON
AF point zoom ON
I contrast off
review off
Light is generally low (indoor, indirect window and/or incandescent light)

I don't know if any of these settings are relevant.

On the A540, I used P mode, MF, ISO 400. No IS on that camera.

script settings (from last D10 run)

NumofShots 10000
delay 1
deley step 500
mode accurat
bv correction 0
disable shoot on
log on
get_shooting sleep 10
buffer log on
verbose log off
debug info off

I've run the G12, G1X and SX30 all for 3000+ cycles using these settings.
No errors encountered.

Even tried the G12 both with and without using EXMEM.

On the D10 how low does your free memory drop in the console display.
The lowest I get is ~300K on the SX30 (no EXMEM).

Phil.
CHDK ports:
  sx30is (1.00c, 1.00h, 1.00l, 1.00n & 1.00p)
  g12 (1.00c, 1.00e, 1.00f & 1.00g)
  sx130is (1.01d & 1.01f)
  ixus310hs (1.00a & 1.01a)
  sx40hs (1.00d, 1.00g & 1.00i)
  g1x (1.00e, 1.00f & 1.00g)
  g5x (1.00c, 1.01a, 1.01b)
  g7x2 (1.01a, 1.01b, 1.10b)

*

Offline reyalp

  • ******
  • 13619
Re: help: how to debug script interrupting?
« Reply #166 on: 03 / May / 2013, 13:02:14 »
On the D10 how low does your free memory drop in the console display.
The lowest I get is ~300K on the SX30 (no EXMEM).
From the run where I was shooting, the lowest was about 345K. For the no-shoot run it was 330k (the buffered logging uses a bit more memory)

I'm very confident this is not a low memory error.

Note that the D10 has run > 3K shots without hitting the error, although this is fairly unusual.

I *think* I trigger the error last night using the chdkptp msgtest script. I'll look into this more today.

I left it running a sleep loop in play mode overnight (~9 hours) and did not hit the error.

what toolchain are you using? Can you post a copy of the lua.elf from your build?

I don't think this is a compiler issue per se, but it's possible that differences could cause or prevent the underlying issue from being triggered.
Don't forget what the H stands for.

*

Offline reyalp

  • ******
  • 13619
Re: help: how to debug script interrupting?
« Reply #167 on: 03 / May / 2013, 15:50:37 »
OK, this is odd.

On  a hunch, I modified the chdkptp msgtest script to allow just repeatedly sending / receiving messages of a fixed size. If anyone else wants to try this, you need to use the .lua files from r339. The executable from the latest binaries should be fine.

With the camera in play mode I ran
Code: [Select]
!m=require'extras/msgtest'
!m.load()
set cli_time=true
!m.test(100,100000)
which (with the new code) sends a 100 byte message and waits for it to be echoed back, 100k times.

This took 26 minutes and triggered the error 3 times.

I rebooted the camera, switched to rec mode, and tried again. Run time 31 minutes, no errors. The increased time is consistent with other tests of running ptp + lua in rec mode, not a surprise.

Repeated above with 250k messages. Run time 78 minutes, still no errors.

 :-[

note
msgtest load() starts a script on the camera. To use normal chdkptp commands, you can end it with

putm quit

and use load() again if you want to use msgtest

edit:
I went to switch the camera back to play mode using putm switch_mode_usb(0) and it crashed with a data abort in MetaCtg task
« Last Edit: 03 / May / 2013, 15:54:30 by reyalp »
Don't forget what the H stands for.


Re: help: how to debug script interrupting?
« Reply #168 on: 03 / May / 2013, 16:54:58 »
Since I had bump this thread I feel compelled to say that I'm not able to perform tests on my sx130 since my wife takes it for three months and I'll stay without any camera. I hope to buy some cheap, maybe somehow broken camera, but I have no idea when.

However - I can confirm that this bug appears somehow randomly - I had series of 900+ photos without triggering bug as well as series of photos, when for each 10 1-2 were affected by bug. This get's annoying when doing timelapses:) I have no clue what determines the bug. I tried with and without photos on SD, with and withount garbage collection. I thought that my script crashes only in accurate mode, but now I see that the fast mode fails as well (or as bad rather).

I'm trying to follow your posts but sometimes their are too technical for me:) However, I feel that the CHDK developers as you are much more helpfull in debuging this really strange piece of program written for completely undocumented environment than the Cannon helpdesk for such common problem as E18...:)

If I could help in any way I'll do that with pleasure - but without a cam I'm a little bit out of the stream...

Regards!
if (2*b || !2*b) {
    cout<<question
}

Compile error: poor Yorick

*

Offline ahull

  • *****
  • 634
Re: help: how to debug script interrupting?
« Reply #169 on: 03 / May / 2013, 17:11:19 »
If you need another pair of eyes, tell me what tests you want to run, I have a couple of cameras to test on, and have been watching the hunt for the cause of this problem for a while. 

 

Related Topics