pulseview crashes on the second run after first run success. Thereafer each and every attemp at running fails first time. Error is same "Pulseview.exe has encountered a problem and needs to close. We are sorry for the inconvienience" Running pulseview on windows XP SP(3) all updates installed. Error Signature ____ AppName pulseview.exe AppVer: 0.0.0.0 ModName pulseview.exe ModVer: 0.0.0.0 Offset: 000356ca Exception Information_____________ Code 0xc0000005 Flags: 0x00000000 Record 0x0000000000000000 Adress: 0x00000000004356ca ALL Puseview version etc #s that foolowed are zero's __________________________________________________________ Install version 03-11 and 03-16 were both "pulseview-NIGHTLY-32bit-static-release-installer.exe" No version number found. ____________________________________________________________ ISSUE DETAILS After many days of tests/contortions - even with a clean system with 'fresh' pulseview install and simplifying further by using pulseview with demo deivce -pulseview fails on the 2nd run. thereafter fails on the first attempted run even after restart. )) With or without device hardware - even in deom mode after first run and allowing to stop on second run pulseview crashes and closes. 2) On restart of pulseview after crash, then pressing the run button (even with demo device) pulseview again crashes. This is true each and every time after restart. 3) Rebooting computer 'resets' that fail on run state and Pulseview again allows a first run with hardware or demo but then crashes on the next run. 4) Uninstalling and reinstalling Pulseview after a crash (not rebooting windows after the prior install crsashed) Then puleeview agian operates as in 1 thru 3 5) Removing all PID of the zadig installed device driver and 'cleaning' windows of any presence of Sigrok and/or Zadig driver and installing Pulseview wihtout any hardware attached AND in Demo Device only and can get TWO runs of demo (only one time - never duplicated) but IF Saving 1st run and Saving 2nd run then crash. Thereafter crashes each time upon restart until reboot then crashes first time. (probably five such install and demo device only reinstalls attempts and most crash on 2nd run the once allwoing two runs was the exception) 6) I did search for ANY SUCH STATE change in a file etc.. that allows Pulseview to 'remember' previous crash but i found nothiner. Whatever that state memory is is reset on reboot or reinstall (without reboot). __________________ NOTE: this failre is very frustrating becuase the device I am trying to use, a LHT00SU1 will work (repeatedly) but only one of those First RUNS (Zadig WINUSB driver installed) WHEN I saw that run with pulseview (UART n I2C datalines on a uC ) IT WAS AMAXING that I had that power in my shop! Not the expensive (and cumbersome 1990's logic anayzers). And I have been doing so much code insertion to debug my code and seeing that decode and timing and addressing etc were all there with sigrok pulseview made my day.. well untill the crashes. And THEN that list of other busses/protocols that Pulseview/Sigrok can decode 'blew my mind' Unfortunately due to health reasons I only have a windows XP machine currently running. I have been punching at this for days now and reading bugzilla and google searches I find simialr issue but none pared down to just demo device fails like this. I am now at a loss Thanks for any help or simple clarity you might provide Bob
Hi Bob, I fixed a bug that caused a similar problem (#1125) a few days ago. Please let me know the PV version, you can see it by clicking on the "settings" icon at the top left and switching to the "about" tab. I'd need the characters after the "0.5" version as they represent the last change that was included in the build. That said, it's going to be a bug in PV itself, so uninstalling/reinstalling PV and/or Zadig won't have any impact. I would suggest you try running PV without any USB devices attached (except for keyboard/mouse). It's possible that there's some device that causes trouble.
Missed the #1125 will look and will try unplugging all the USB devices next. For now wanted to thank you for response and to get you the version info below. Pulseview 0.5.0-git9a0e130 GNU GPL, version 3 or later sigrok.org _______________________________ Libraries and features: Qt 5.7.1 glibmm 2.42.0 Boost 1_60 libsigrok 0.6.0-git-e843992/4:0:0 (rt: 0.6.0-git-e843992/4:0:0) - glib 2.50.2 (rt: 2.50.2/5002:2) - libzip 1.1.3 - libserialport 0.1.1/1:0:1 (rt: 0.1.1/1:0:1) - libusb-1.0 1.0.20.11003-rc3 - libftdi 1.2 - Host i686-w64-mingw32.static.posix, little-endian - SCPI backends TCP, serial, USBTMC libsigrokdecode 0.6.0-git-8111675/4:0:0 (rt: 0.6.0-git-8111675/4:0:0) - glib 2.50.2 (rt: 2.50.2/5002:2) - Python 3.4.4 / 0x30404f0 (API 1013, ABI 3) - Host i686-w64-mingw32.static.posix, little-endia
Per your response. 2018-03-17-1244. 1) Removed USB deivces other than keyboard/mouse (logitech dongle w_Microsoft drivers installed). Note The only USB devices that were present on the WinXP_SP(3) system during last few days of testing were flash memory devices. There are no built in USB devices on MB except hubs. And the LHT00SU1 and the Nano used for quick test were both removed and the driver instance of the LHT00SU1 was deleted in registry. 2) Tried restarting a 'prefailed' PV with USB devices just removed (before reboot ) and it started good but failed on 1st run attempt w_ demo device (SAME) 3) Rebooted computer started PV and 1st run good, 2nd run failed w_demo device (SAME). Note I have a 'portable' install of Python 34 and just in case it might be some confict w_sigrok, I removed the Windows nvironmental pointers to that Python34 for both User and System and rebooted. NOJOY - SAME failure modes. Bob
Interesting, thanks! I'm not able to reproduce this on a Windows 10 box, but I *can* reproduce it on a Windows XP VM indeed (installer as of a few minutes ago). We'll have to look into the details. I don't have any Python installed on the VM, so that's probably unrelated. No hardware attached either, just demo triggers it indeed.
I have continued to try and find where a 'state' is saved in Windows that allows PV to 'rememer' that a first run with subsequent 2nd run failures have occured Note I have verified that if PV is closed after a succssful 1st run and then if PV is restarted and then run is clicked - PV fails. SO it must be a state associated with the first run that causes subsequent runs to fail - whether PV is left open, and a subsequent 2nd run is intiated and fails, or PV is closed and then restarted and a subsequent "2nd run" fails (whether same day or many days later - unless windows has been rebooted). There must be state saved somewhere in memory or file system that PV 'reads' to know that it has finished a 1st run. First because an unitstall / reinstall of PV on same 'boot' of Windows 'resets' that remembered 'state' and PV again does a run (but only once). OR 2nd because a reboot of Windows erases or resets that saved 'state' and allows PV to run (again only once). So not able to search memory for a saved area - and having used my small windows xp machine to maxiumum for days and the PV saved state still is persistant as restarting PV (after having failed a 2nd run) initial failure a failure on run - I chose ot focus on a file state change as the location for PV saved state. For time saving (as a reboot of Windows takes a long time) I 'reset' PV saved state with an uninstall/reistall of PV ((BTW - Kudos on clean uninstall!)) and then focused on searching windows for any file changes after a PV first run / 2nd run fail. MANY MANY MANY searches for modified file dates/times (and completely thru two SSD Drives) and found VERY FEW files change.. NOTE NO CHANGES OCCUR IN THE sigrok pv program files folders - not in library nor fw nor code etc etc... Two files that are changed by PV occur after the run fails and probably by windows Dr Watson etal : C:\Documents and Settings\Administrator\Cookies\index.dat C:\Documents and Settings\NetworkService\Cookies\index.dat These change occur on PV restart (after a 1st run - close or after 2nd run fail) not during Startup- initialzation nor device searching but only AFTER clicking the run button and only AFTER the close of PV error popup windows. These index.dat files are normally blank (I do not use Internet Explorer). The data is the same for both and is created (Windows Dr Watson etal most likely) after the PV fail popup window is closed. I have deleted the contents and the same data is entered. However there is NO CHANGE in the file (whether file is deleted or the file left with contents emptied) IF PV is started (after a 1st run fail) but then is NOT RUN but closed - no error windows. And if I force delete the file or blank it out and then run and PV still fails on 1st run. So these files must NOT BE the "saved state" The "2nd run fails" produce files from Dr Watson reporting, but only on the fail of PV and only after the closing of the PV fail popup window so are not the "saved state". Note that curiously the PV fail produces change in two log files for ScanSnap scanner service. I have stopped that service and there is no change in PV responses and I doubt a commonality with your virtual WinXP install that duplicates the PV failure. Again only after the PV fail window is closed, so I have not followed this rabbit hole further - just noted for completeness here. Note I installed PV Debug version. On PV "2nd run fail" there is data shown on debug windows but PV closes os fast that only an instantaeous flash of characters appears in the window before closing - but not readable at all and i have not found a way to capture it. Note I installed PV onto different drive than Windows drive and same failure with same file changes. I simply have not been able to find any file saved state - at least with windows explorer searches for modified date/time. I have searched thru the registry but no joy and just a wild shot at it. Again not able to search / compare memory areas. Robert
The only state that PulseView saves is that it remembers the device that was used last time, the files that were open the last time etc. The location of this saved state is OS-dependent, on Windows it's a registry key. You can use "-l 5" when running the debug PulseView version (run it from a DOS box so that the window doesn't close immediately) to see where exactly it's saved. Search for "Settings:" in the "-l 5" output (should be the third line or so). You can force PulseView to forget/delete all its state by running it with the "-c" option.
I'm pretty sure now that this is the same issue as in #1125, closing here, discussions will continue there. *** This bug has been marked as a duplicate of bug 1125 ***
I'd just like to add that we performed a bug fix yesterday which likely fixes the issue you're encountering. Please download a new Windows nightly build and check if the issue is now resolved for you.
Now PulseView 0.5.0-git-e4cc0ee works great on Windows XP, tested static-relase on demodevice and Usbee (f2xlafw). Big thanks to contributors for fixing bug.
FIXED - new verions PV 0.5.0git-e4cc03 dwnld 04-06-0200 Runs (demo and with LHT00SU1 'as clone cwav device) FIRST, 2nd 3rd... (really tried every comvination of restart - reboot start restart etc and all good) THANK YOU SO MUCH ! Please excuse my delay in responding to your debug w_ switch '-l 5' request. I was not good to go until early this AM. Irrelavent but for record with switch failed first run (install & reboot) but as before debug w_o switch good 1st run (reboot or install) then fail 2nd ... same ole) Robert