Re: [Kicad-developers] GAL canvas deadlock

Hmm... That makes sense. In the calls, we only unlock for std::runtime_error. Since you can trigger this, could you add a default catch with an unlock call? (opengl_gal.cpp:357 and 2009) -Seth Am Do., 11. Okt. 2018 um 16:51 Uhr schrieb Jeff Young : > Hi Seth, > > I agree that it would be

Re: [Kicad-developers] GAL canvas deadlock

Hi Seth, I agree that it would be better to get to the bottom of this. I just wasn’t having any luck. The context locking issue has to do with trying to go to fallback canvas while in the middle of rendering. But we do that (in both Tom’s and my case) as a result of catching an error from

Re: [Kicad-developers] Windows build symbol files?

I was not aware of that change, but I have made a test build if anyone likes to testrun it. This is master built with debug as the build type and the option enabled. http://downloads.kicad-pcb.org/windows/testing/patched/kicad-patched-144-ddea4e11d-x86_64.exe On Thu, 11 Oct 2018 at 22:54, Seth

Re: [Kicad-developers] 5.0.1 status

FWIW everything should be tagged now. So packagers should be able to bump the pkgver. On Wed, 10 Oct 2018 at 12:18, Marco Ciampa wrote: > > On Mon, Oct 08, 2018 at 09:24:02AM -0400, Wayne Stambaugh wrote: > > How much time will our translators and librarians need to tag 5.0.1? > > Hopefully not

Re: [Kicad-developers] Windows build symbol files?

Hi Simon- Am Do., 11. Okt. 2018 um 11:33 Uhr schrieb Simon Richter < simon.rich...@hogyros.de>: > These should be Release builds, so no debug information. > OK. Could we not strip the symbols in the linking to give a bit more information in the dump files? Or does that incur a large penalty

Re: [Kicad-developers] GAL canvas deadlock

Am Do., 11. Okt. 2018 um 12:20 Uhr schrieb Jeff Young : > I pushed a change that keeps the error from happening for me. It now > makes sure it updates the GAL settings in the active window first and then > any others. > > @Seth, if you want to look into it on your Mac, the change is in >

Re: [Kicad-developers] GAL canvas deadlock

I pushed a change that keeps the error from happening for me. It now makes sure it updates the GAL settings in the active window first and then any others. @Seth, if you want to look into it on your Mac, the change is in KIWAY::CommonSettingsChanged(). Cheers, Jeff. > On 11 Oct 2018, at

Re: [Kicad-developers] Windows build symbol files?

Hi Seth, On 10/11/2018 07:39 PM, Seth Hillbrand wrote: > Are you generating symbol files for KiCad on the nightly builds?  If so, > where can I find them in the install? These should be Release builds, so no debug information. The toplevel CMakeLists.txt complains if someone tries to build

Re: [Kicad-developers] Net selector

Cool. Thanks for running down the right number. Cheers, Jeff. > On 11 Oct 2018, at 18:39, John Beard wrote: > > Hi Jeff, > > I'm not a committer, so you might as well go ahead, rather than > messing with a one-line patch from me! > > Cheers, > > John > On Thu, Oct 11, 2018 at 6:34 PM Jeff

[Kicad-developers] Windows build symbol files?

Hi Simon- Are you generating symbol files for KiCad on the nightly builds? If so, where can I find them in the install? Thanks- Seth ___ Mailing list: https://launchpad.net/~kicad-developers Post to : kicad-developers@lists.launchpad.net

Re: [Kicad-developers] Net selector

Hi Jeff, I'm not a committer, so you might as well go ahead, rather than messing with a one-line patch from me! Cheers, John On Thu, Oct 11, 2018 at 6:34 PM Jeff Young wrote: > > Hi John, > > That’s the right place to adjust it. There’s also LIST_PADDING which gets > put in once (rather than

Re: [Kicad-developers] Net selector

Hi John, That’s the right place to adjust it. There’s also LIST_PADDING which gets put in once (rather than per-item), but it looks like LIST_ITEM_PADDING of 6 is doing the trick. Are you going to commit that, or do you want me to? Cheers, Jeff. > On 11 Oct 2018, at 18:33, John Beard

Re: [Kicad-developers] Net selector

Hi Jeff, It's better now for <= 3 items. For more, it still cuts a little bit off, but less than before. I adjusted LIST_ITEM_PADDING to 6, and it seems to hove done the trick. However, I haven't checked very carefully if I have adjusted the right thing. Attached are two screenshots, so if you

Re: [Kicad-developers] Net selector

Am Do., 11. Okt. 2018 um 08:58 Uhr schrieb Jeff Young : > I’ve been finding that saving the filter is unexpected, even when useful. > And unexpected is always bad, because it erodes one’s confidence in their > user model. So I think I’ll just remove it. > > Anyone else (with current code) able

Re: [Kicad-developers] GAL canvas deadlock

Sorry, Jeff but my Mac is currently burning a hole in the desk trying to get the 5.0.1 build out. I can look this weekend probably On Thu, Oct 11, 2018, 3:01 AM Jeff Young wrote: > There’s a whole pile of OSX frames in the stack before the paint event, so > it’s possible it has to do with the

Re: [Kicad-developers] GAL canvas deadlock

Yeah, that’s the same end result as I’m getting. The pending error in my case is "Error: switching framebuffer: invalid framebuffer operation”. Does that suggest anything to you that I might look for? > On 11 Oct 2018, at 14:22, Tomasz Wlostowski wrote: > > On 11/10/18 14:55, Seth Hillbrand

Re: [Kicad-developers] GAL canvas deadlock

No worries. > On 11 Oct 2018, at 13:55, Seth Hillbrand wrote: > > Sorry, Jeff but my Mac is currently burning a hole in the desk trying to get > the 5.0.1 build out. I can look this weekend probably > > On Thu, Oct 11, 2018, 3:01 AM Jeff Young > wrote: > There’s a

Re: [Kicad-developers] GAL canvas deadlock

On 11/10/18 14:55, Seth Hillbrand wrote: > Sorry, Jeff but my Mac is currently burning a hole in the desk trying to > get the 5.0.1 build out. I can look this weekend probably Hi Jeff, I don't know if this is very relevant for your case, but I noticed a deadlock in: #2 0x7fffe3b2a53c in

Re: [Kicad-developers] GAL canvas deadlock

I tried a bunch of different scenarios to see how reliant it was on a particular set of steps. It reproduces (for me) whether you change the properties *from* PcbNew or Eeschema. It reproduces whether you change the properties *on* PcbNew or Eeschema. It does *not* reproduce when changing the

Re: [Kicad-developers] [PATCH] Fuzzable PCB parsing test harness

Hi, Here's a quick patch to add some documentation about print-debugging and trace. It's a follow-up to the fuzz tool, as there would be a conflict in the MD file, as one section follows the other, so it needs to go in afterwards, or otherwise be manually resolved. Also includes a handy list of

Re: [Kicad-developers] Net selector

Hi John, I removed the filter retention. I also adjusted the spacing on GTK. I’ve changed it from what Windows uses to what OSX uses. It may need yet a 3rd value, so let me know if it’s still too short with only a few list items (or too tall with a bunch of list items). Cheers, Jeff. > On

Re: [Kicad-developers] Net selector

Hi Jeff, It doesn't seem to accept spaces for me on GTK (bea75753d). I don't think retaining the last filter is extremely useful, though perhaps if you were going around changing a lot of things to the same net? But that will be a better job for the Object Inspector in future anyway. Cheers,

Re: [Kicad-developers] Net selector

I’ve been finding that saving the filter is unexpected, even when useful. And unexpected is always bad, because it erodes one’s confidence in their user model. So I think I’ll just remove it. Anyone else (with current code) able to type a space into the filter? The commits Eeli mentions use

Re: [Kicad-developers] GAL canvas deadlock

There’s a whole pile of OSX frames in the stack before the paint event, so it’s possible it has to do with the order of event processing. Seth, can you try on Mac? ___ Mailing list: https://launchpad.net/~kicad-developers Post to :