Bug Reports against Recent Builds -- Report bugs here - page 23 - CHDK Releases - CHDK Forum supplierdeeply

Bug Reports against Recent Builds -- Report bugs here

  • 281 Replies
  • 100842 Views
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #220 on: 31 / January / 2013, 22:04:52 »
Advertisements
I think the file corruption was caused by the bug I fixed in 2527 - so it shouldn't happen again (fingers crossed).
Still seems kind of strange it used different params at startup than script reload from SD card.  One more thing "that goes bump in the night" I guess.

Another solution would  be to add a new option to the menu item that selects the parameter set to use. Currently the allowed values are 0 - 9; but it would be possible to add 'Default' as an option which would bypass loading any param file.
Probably makes sense to change the Save parms [ ] menu item to be a MENUITEM_PROC rather than MENUITEM_BOOL if you do that?
Ported :   A1200    SD940   G10    Powershot N    G16

Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #221 on: 31 / January / 2013, 22:58:53 »
Second question of the night.

Why do the OSD icons display in <ALT> mode even when you have the OSD disabled in the OSD menu ? 


Update :   I must be getting lazy - I can figure this one out for myself I'm sure.
« Last Edit: 01 / February / 2013, 00:06:46 by waterwingz »
Ported :   A1200    SD940   G10    Powershot N    G16

*

Offline philmoz

  • *****
  • 3332
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #222 on: 09 / February / 2013, 05:12:20 »
Another solution would  be to add a new option to the menu item that selects the parameter set to use. Currently the allowed values are 0 - 9; but it would be possible to add 'Default' as an option which would bypass loading any param file.

I've added this to the trunk version in revision 2559.

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)

Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #223 on: 09 / February / 2013, 09:54:06 »
Another solution would  be to add a new option to the menu item that selects the parameter set to use. Currently the allowed values are 0 - 9; but it would be possible to add 'Default' as an option which would bypass loading any param file.
I've added this to the trunk version in revision 2559.
Very nice - thanks.   I'll see if I can update some of the docs. 

For reference, when the Save Params value is set to On,  when are the parameters actually saved?  When you exit the Script menu?  When the script is started? When the Script stops?  When you exit <ALT> mode?
Ported :   A1200    SD940   G10    Powershot N    G16


*

Offline philmoz

  • *****
  • 3332
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #224 on: 09 / February / 2013, 15:56:29 »
Another solution would  be to add a new option to the menu item that selects the parameter set to use. Currently the allowed values are 0 - 9; but it would be possible to add 'Default' as an option which would bypass loading any param file.
I've added this to the trunk version in revision 2559.
Very nice - thanks.   I'll see if I can update some of the docs. 

For reference, when the Save Params value is set to On,  when are the parameters actually saved?  When you exit the Script menu?  When the script is started? When the Script stops?  When you exit <ALT> mode?


Parameters are saved when:
- Parameter set is changed (previously selected set is saved)
- 'Save Params' is set to On (current set is saved)
- Script is run (before script starts)
- Load a new script (old script paramaters saved before new one loaded)
- 'Load default param values' menu option (defaults saved to current set)

Script parameters will also be saved if either the user menu was edited or another 'conf' value was changed when:
- Exit <ALT> mode
- Exit CHDK menu back to <ALT> mode (including the script menu)

Note sure if the last two are really needed; but it doesn't cause problems so I'm reluctant to change it.

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 srsa_4c

  • ******
  • 4426
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #225 on: 14 / February / 2013, 21:10:30 »
Benchmark causes cam to switch off.
Tried current trunk build on ixus65_sd630, a410, a470 (and s1is). Log is not written.
It happens before the "Flash-Card" part. Screen benchmark shows lower values than it used to (probably unrelated).

The uart output on the s1is shows "Watch Dog Timer Timeout !!" after the cam silently restarts (which means there are no asserts or exceptions).  The a410 doesn't show anything on its uart output (haven't tried to make it show more output).

Doesn't happen on current 1.1 builds.

CHDK 1.2, ixus110_sd960:
Test executes seemingly without a problem.
However, PTP communication (including ptp live view) halts right after starting the test, resumes only when the card benchmarks start.


I have a feeling that the test somehow manages to halt some tasks, resulting in watchdog timeout on slower (<=DIGIC III) cameras.

The problem is also unrelated to blocking the PhySw task, since the s1is port has a separate task to handle CHDK keyboard.
I've compared the benchmark code to an earlier version (r1984), and the differences seem pretty minor (some differences in register variables and the added optimization-related gcc attribute). I don't see how it could block other tasks completely...


late edit: the issue was reported before me: http://chdk.setepontos.com/index.php?topic=5560.msg95495#msg95495 (the A3100 is DIGIC III).
« Last Edit: 21 / April / 2013, 21:09:41 by srsa_4c »

*

Offline srsa_4c

  • ******
  • 4426
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #226 on: 16 / February / 2013, 06:42:05 »
the added optimization-related gcc attribute
This is the cause. When removed, gcc optimizes out some tests (the ones where "Calculating..." doesn't disappear). I split out the first 4 memory related tests to separate functions with __attribute__((optimize("O0"))), and left the rest of gui_bench_run() without that attribute. Now it executes on all slower cams without shutdown.

*

Offline lapser

  • *****
  • 1093
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #227 on: 16 / February / 2013, 14:01:17 »
I've noticed that the size of the screen console isn't initialized when a script starts. The screen console size depends on what the last program set it to.

It's easily fixed by changing line 225 in script.c from:

    console_clear();
to
    console_close();
EOS-M3_120f / SX50_100b / SX260_101a / G1X_100g / D20_100b
https://www.youtube.com/user/DrLapser/videos


*

Offline philmoz

  • *****
  • 3332
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #228 on: 16 / February / 2013, 18:04:07 »
the added optimization-related gcc attribute
This is the cause. When removed, gcc optimizes out some tests (the ones where "Calculating..." doesn't disappear). I split out the first 4 memory related tests to separate functions with __attribute__((optimize("O0"))), and left the rest of gui_bench_run() without that attribute. Now it executes on all slower cams without shutdown.

Strange behaviour, not sure I understand why this would fix it; but stranger things have happened.

Added in revision 2577. I changed the function names to match the operation being done.

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 philmoz

  • *****
  • 3332
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #229 on: 16 / February / 2013, 18:04:54 »
I've noticed that the size of the screen console isn't initialized when a script starts. The screen console size depends on what the last program set it to.

It's easily fixed by changing line 225 in script.c from:

    console_clear();
to
    console_close();


Fixed in revision 2577.

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)

 

Related Topics