Skip to main content
Topic: OLS client 0.9.4 RC2 (Read 17856 times) previous topic - next topic

OLS client 0.9.4 RC2

It took longer than originally anticipated, but it is finally available: the [s:]first[/s:] second Release Candidate (RC) of the 0.9.4 version!

Implemented features and Issues solved in this release (including already closed ones):

  • #4, VCD exports ignore selected channel groups;
  • #5, Verify frequency measurement with RLE captures;
  • #6, support OLS-device profiles for BP, regular OLS and SUMP;
  • [s:]#30, Add icon to window/application;[/s:]
  • #31, 200MHz capture fails halfway;
  • #32, Add warning message about highest channel icw RLE;
  • [s:]#33, when RLE is enabled the first value rendered is always low;[/s:]
  • [s:]#34, Is it possible to disable DTR when opening the serial;[/s:]
  • [s:]#37, No menu entry for 'capture' functions;[/s:]
  • #43, Preferences in Edit/Preferences dialog are not saved;
  • #47, Stop button now always enabled;
  • #50, RLE channel group issue with 0.9.4-b1;
  • [s:]#51, I2C decoder produces invalid address;[/s:]
  • [s:]#54, Add "Jump to Marker" buttons for remaining cursors;[/s:]
  • #56, Stop button not working correctly;
  • #57, 200MHz 8 bit capture incorrect if upper channel groups selected;
  • #58, 24 bit capture set incorrect memory size with auto;
  • #60, (UART analyser) settings are not retained when dialog closed;
  • #62, Open existing project "current project has changed" dialog inverted.

A call to all beta testers: grab an archive and test away using the testing charter:

Please do not use this topic for feature/improvement requests! Create either a new Client topic, or file an issue request on GitHub...

All issues found should be reported as issue on GitHub and can be discussed in this topic.

If tested, could you post a message in this topic stating the platforms (operating system + version, 32/64-bit, and used hardware) on which you've tested it, and possibly the issue numbers verified?
when good software is not an alternative...

Re: OLS client 0.9.4 RC1

Reply #1
Hi Jawi, have done some testing with RC1 on 32bit Fedora 14

Platform tests;

Not a new issue (or a problem with the client), but the test charter did remind me that Fedora 14 broke non-root RXTX lock file usage some time ago - see http://https://bugzilla.redhat.com/show_bug.cgi?id=581884 for background. The Fedora RXTX package has been patched to create lock files in another directory, but for bundled RXTX I need to change permissions on the /var/lock subdirectory.

Acquisition tests;

Can confirm issue 50 is solved.
I do see that if a long capture is stopped before completion, then the data is not decoded. Will raise a new issue for this on GitHub. [edit: I see issue 56 has already been raised for this]

Tooling tests;

UART analyser settings  (channel selection) are not retained when the dialog is closed. If this is unexpected I'll raise an issue for it?

Cheers,
Jon

Re: OLS client 0.9.4 RC1

Reply #2
[quote author="jone"]
Platform tests;

Not a new issue (or a problem with the client), but the test charter did remind me that Fedora 14 broke non-root RXTX lock file usage some time ago - see http://https://bugzilla.redhat.com/show_bug.cgi?id=581884 for background. The Fedora RXTX package has been patched to create lock files in another directory, but for bundled RXTX I need to change permissions on the /var/lock subdirectory.
[/quote]

Hmm, I'm thinking to switch over to NrJavaSerial, which appears to be more actively developed. I see they've already fiddled around with lockfiles, so perhaps it is already fixed, or they can fix it pretty quickly...

[quote author="jone"]
Acquisition tests;

Can confirm issue 50 is solved.
I do see that if a long capture is stopped before completion, then the data is not decoded. Will raise a new issue for this on GitHub. [edit: I see issue 56 has already been raised for this]
[/quote]

Yeah, thanks for confirming this. This issue will be solved in the final 0.9.4 release.

[quote author="jone"]
Tooling tests;

UART analyser settings  (channel selection) are not retained when the dialog is closed. If this is unexpected I'll raise an issue for it?
[/quote]

You can raise an issue for this, as it is not how it should work...
when good software is not an alternative...

Re: OLS client 0.9.4 RC1

Reply #3
Jawi,

A new release is great, give me the word when this Release Candidate is ready to be a final release and I will put together the 3.08 release package. I have a couple more tricks that I'm ready to release with a 3.08 package.

On a side note, what do you think it would take to add support for a real time mode? Kinsa released that test bitstream that has continuous mode as a proof of concept, its not very fast at 10K samples but if the client supported a real time mode we might be able to start working on faster solutions.

Jack.

Re: OLS client 0.9.4 RC1

Reply #4
@Jack: I've put the "real-time" acquisition mode on the roadmap, currently for the 1.0 release, though I've to see if that would be feasible. There's also an issue (#16) for this requirement as well...
when good software is not an alternative...

Re: OLS client 0.9.4 RC1

Reply #5
I tried 0.9.4 rc1 on Arch Linux and found that the I2C protocol analyzer doesn't remember the settings (bit order, pin assignments, mode) between one analysis and the next.  This worked fine in 0.9.3.  No errors are reported on the console.

 

Re: OLS client 0.9.4 RC1

Reply #6
[quote author="DeKay"]I tried 0.9.4 rc1 on Arch Linux and found that the I2C protocol analyzer doesn't remember the settings (bit order, pin assignments, mode) between one analysis and the next.  This worked fine in 0.9.3.  No errors are reported on the console.[/quote]

Thanks for reporting, DeKay. In fact, the issue you found is related to issue #60: none of the user settings were correctly retained. It was easy to fix, and should be solved in the latest GitHub repository. In any case, consider this fixed for 0.9.4 final.
when good software is not an alternative...

Re: OLS client 0.9.4 RC2

Reply #7
Released the RC2 of the 0.9.4 just now!

Five additional bugs were fixed, and some additional regression tests were added. I expect this RC will become the final release...

Enjoy!
when good software is not an alternative...

Re: OLS client 0.9.4 RC2

Reply #8
One little note, that might be insignificant once your new display is implemented, but I thought I'd mention it.

Mousescroll doesn't work if the cursor is over the menus/icons at the top, which is not a big deal really.

If you mousescroll anywhere in the window it works fine, but if you try to control-mousescroll to zoom, it only works if the cursor is actually over the traces. If the cursor is over the grey area below the traces it does a normal time scroll instead of a zoom even if control is held.

Even if the client is in the foreground, if you mouse scroll with the cursor outside the window (if it's not maximized) the scrolling does nothing. There might be reasons this is desirable, but it's also kind of annoying if you don't keep the client maximized.

Re: OLS client 0.9.4 RC2

Reply #9
[quote author="Gridstop"]Mousescroll doesn't work if the cursor is over the menus/icons at the top, which is not a big deal really.

If you mousescroll anywhere in the window it works fine, but if you try to control-mousescroll to zoom, it only works if the cursor is actually over the traces. If the cursor is over the grey area below the traces it does a normal time scroll instead of a zoom even if control is held.

Even if the client is in the foreground, if you mouse scroll with the cursor outside the window (if it's not maximized) the scrolling does nothing. There might be reasons this is desirable, but it's also kind of annoying if you don't keep the client maximized.[/quote]

I'm not sure I entirely follow your request, you want to be able to scroll/zoom the signal display when the cursor is anywhere on screen, but as long as the client window has the focus?
when good software is not an alternative...

Re: OLS client 0.9.4 RC2

Reply #10
Well I guess the biggest thing is that the control key (for zooming) is ignored even if the cursor is over the client, if it's over the grey area below the traces. The cursor has to be hovering over the actual traces for zooming to work.

But yeah, in addition to that, if it's not a big deal it would be great if panning/zooming worked whenever the client had focus, even if the mouse cursor were outside the window.

Re: OLS client 0.9.4 RC2

Reply #11
Hi. I just downloaded the RC2 and I get the following error when trying to launch the application from a different location than its install directory. For example:

This produces an error:
Code: [Select]
armandas@ouroboros:~$ ./Apps/OLS/run.sh

This starts the application:
Code: [Select]
armandas@ouroboros:~/Apps/OLS$ ./run.sh

Note to mods: I had to put the error messages in the file as the filter kept asking to remove off-site links.

Re: OLS client 0.9.4 RC2

Reply #12
@armandas: on what platform are you seeing this? I've tried to reproduce this on my Linux (work) laptop, and the client runs just fine...
when good software is not an alternative...

Re: OLS client 0.9.4 RC2

Reply #13
[quote author="jawi"]@armandas: on what platform are you seeing this? I've tried to reproduce this on my Linux (work) laptop, and the client runs just fine...[/quote]

I run Ubuntu 10.04. I did not have this problem with last version (not sure whether it was RC1 or beta).

Re: OLS client 0.9.4 RC2

Reply #14
[quote author="armandas"][quote author="jawi"]@armandas: on what platform are you seeing this? I've tried to reproduce this on my Linux (work) laptop, and the client runs just fine...[/quote]

I run Ubuntu 10.04. I did not have this problem with last version (not sure whether it was RC1 or beta).[/quote]

Ah, is there a "felix-cache" folder in your home folder? Does the problem resolve if you remove that folder?
when good software is not an alternative...