• Content Count

  • Joined

  • Last visited

Community Reputation

3 Neutral

About Alan_S

  • Rank
    Active Member

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Sorry, a Sorry, the above should read "... try to launch a daughter window that doesn't have a scrollbar. "
  2. I've sent you some logs by PM. Perhaps the reason you couldn't reproduce it was Windows. I doubt you are using Win 7 32 bit... Having said that, I've found that sometimes, very rarely, It does work OK for me so perhaps you were in luck! Oh yes, as I've mentioned in 'notes' in the material I sent, I've since discovered that the issue can also occur on log list entries that will try to launch a daughter window that doesn't have a sidebar.
  3. Thanks for the tip! But unless you mean other logs than those I was talking about above, no - it's the same. Still, now I've discovered that launching from the sidebar doesn't give the problem I'll use that until a fix surfaces.
  4. Some good news and some sad news. The good news is that the UI response failure I reported right at the beginning has been fixed by version 2019.2.0.9269 The sad news is that it hasn't fixed the 'overlay' problem - see above post of December 9, 2018. I've looked into this further and narrowed it down: The issue seems to concern the 'Logs' section, when selecting 'View Details' for an entry that will cause the daughter pane to have a scrollbar. In practice this only appears to be the case for an 'Updates' entry. The following will demonstrate the problem: Launch the UI by clicking the EAM icon in the notification area, a.k.a. system tray. Click the blue 'Logs' box (not the sidebar icon). Mark an 'Update' entry in the list presented and click 'View Details'. The daughter pane is not launched and the UI is unresponsive, presumably waiting until the (invisible) daughter pane is closed. There's probably a window behind the UI. Click this and then click the UI window. The part of the daughter pane that overlays the window behind the UI now appears - see post of December 9, 2018. I mention this simply as something that might aid fault finding. In practice it's superfluous. To display the whole daughter pane, click any window on the desktop (start something if the UI was all that was on the desktop) and swipe it over the UI. Magic! . And a couple of observations: The issue only occurs one time per UI launch. Having once got it working, other entries launch the daughter pane correctly. It only seems to occur when launching 'Logs' by clicking the blue box and appears to work properly if launched from the sidebar. Now, I realise that this problem is a bit esoteric and that indeed, the vast majority never look at the logs. And, you may ask, “What's he whinging about? He's found two ways to bypass the problem”. True, but others might not be so tenacious in delving into it and most would simply give up. Also, although I can't find any case other than the above, who knows? And what might lie in wait in the future? Look at it as a suggestion to improve the product. Surely it ought to be relatively simple to fix?
  5. Reporting that the situation hasn't occurred for me since the shower of incidents mid December. That's nearly 10 weeks. I haven't noticed any announcement that sounds like a fix, neither in the beta forum nor the Change Blog (though perhaps it's hidden under “Improved stability”) but something must have happened. It took over 6 months but, as the Bard of Avon said, “All's well that ends well”! Thank you all who have got this to happen!
  6. Thank you! That explains it. I suppose I must have “invented” the EAM_yyyy-mm-dd folder name myself and been a bit more vigilant in previous runs.
  7. I wanted to export the settings. I didn't examine the folder suggesed by the tool very carefully, but saw it where I had exported previously so clicked the OK-button. Expecting the name to be today's date, I looked for the backup but couldn't find it. So, tried again and saw that the tool wasn't suggesting a unique folder name but an existing folder – not the latest even, but one from 2016: EAM_2016-11-28, which apparently got overwritten (see 'Date modified'). What I would expect is that the suggestion would be a new folder and I believe that this was what it did when I used it last. And if one already existed, that either a suffix would be added (e.g. EAM_yyyy-mm-dd-01) or that the user would be warned that he/she was about to overwrite it unless changed. Now I know what happens I will be a bit more careful, but existing data should't get overwritten. If I've done something wrong then please tell me! And yes, it's high time for a cleanup...
  8. Apologies for the belated response - been a bit busy. In Windows 7, the basic DirectX support is version 11.0. Apparently, the full version of 11.1 is supported by Win 8.1 but says that “Portions of the 'DirectX 11.1 Runtime' are being made available on Windows 7 Service Pack 1 via the Platform Update for Windows 7 Service Pack 1 and Windows Server 2008 R2 Service Pack 1 (KB 2670838)” describes what KB2670838 does. KB2670838 was installed January 25th 2016. And that seems to be as good as it gets unless I upgrade to Win 8.1 or Win 10. But the UI worked perfectly until 2018.11.0.9073.
  9. Regarding my post above "Just taken a look at it and it's a bit different." Of course it's different! I was looking at it on my other computer (the Dell). The HP is still as in my previous experiment. It's a little bit faster than the Dell. Sorry, I wasn't paying attention but it's getting a bit late this side of the pond! As to the motherboards. Neither are Asus. Speccy says the manufacturers are Dell and HP respecively. Both machines are ancient. The Dell is from November 2007 and there is no newer BIOS. The HP is from 2010. Perhaps the installed BIOS is not the best, but the problem surfaced with build 9073. The UI worked fine previously. There appears to be a couple of slightly newer versions but I don't understand (on the HP site) which would be correct. To be honest I'm loathe to attempt an update it as once it's changed, it seems there's no going back.
  10. Just taken a look at it and it's a bit different. With Settings selected, but not doing anything to change etc. a2start is taking a bit more - changing continually between less than 1 up to 7%. But there's still lots of CPU left!
  11. Aha! Yes Marko, I did misunderstand. I haven't experienced any problems when altering the debug logging state. I'll leave it on though as it might give a bit more info for anyone who is working on the problem. Thanks for the clarification.
  12. So do I stapp. With a proper a mechanical switch too... To business: It's very early days but the good news is that I haven't experienced any freezes at start-up with build 9073, installed a week ago. The bad news is that I've experienced 3 freezes during normal activity: last Wednesday, Friday and Sunday. I've sent a dump for each of these to [email protected] asking that it be forwarded to you, GT500. Also, a doc file for each, describing the scenarios. Were I a developer I'd want as much material as possible, but I leave it to you to decide what your people need. I really hope they can find out what's going on and fix it, as a freeze every other day is just not viable. Or perhaps the problem is that my two computers and EAM are simply not compatible. I'm going to try Marko's advice and have debug logging enabled continually. It does agree with my own experience when all this started (see my original post). But even if it works, I can't accept it as a solution. Just a very temporary workaround until the developers fix it properly. Fingers crossed...
  13. I did a very simple and brief test, clicking Settings and then Advanced. a2guard is in action very little: now and then a brief flash of 1 - 4% but that's all. a2start uses CPU continually when in Settings: about 1 – 3%, most of the time exactly 1.92%. This is what I did: Start Win Launch UI but do nothing a2start 1.5 – 3% continually, mainly 1.92% a2guard zero 13:43 Click Settings a2start 1.5 – 3% continually, mainly 1.92% a2guard briefly 4% 13:45 Click Advanced – no response a2start 1.5 – 3% continually, mainly 1.92% a2guard zero 13:48 Click PX then EAM still no response a2start 1.5 – 3% continually, mainly 1.92% a2guard zero 13:49 Get reply a2start 1.5 – 3% continually, mainly 1.92% a2guard zero Close UI Then, a look at logs. Start Win Launch UI and do a Quick scan 14:15 Click Logs Hightlight the scan and click View Details No response a2start zero a2guard zero 14:18 Click the Process Explorer window then the UI window and the reply pop-up appears! The 'other' window doesn't have to be PX: Notepad, Windows Explorer - anything will do the trick. But look at the attached pop-up screenshot. Note that the reply pop-up is not complete: only what overlays the window behind it (Process Explorer here). The rest can be coaxed out of hiding by moving the window behind, to increase the 'shadow', in this case moving Process Explorer right. Well, it's a workaround but it can be tricky with a cluttered desktop. If the whole pop-up can't be retrieved there's another problem: it has to be closed before anything else can be done and how to do that when neither the Close button nor the 'X' in the top right hand corner are accessable? Two other observations: When waiting for a reply from 'View details' the UI window is dead to the user. Can't move it on the desktop and can't close it. It appears that the problems only occur when a new log entry is in the list, even if the 'View details' concerns an older entry. Once past the hurdle it seems to work normally, at least until next time Logs is invoked, even if there are no new entries. Sorry, I see that this is all a bit incoherent. But it's not easy to invoke and observe on one computer and at the same time document on another computer! But perhaps it's of use. The bottom line is that a2guard uses very little CPU and only briefly. a2start uses CPU continually when in Settings - almost all the time 1.92% but almost nothing when trying to look at logs, even when waiting for a reply that should be instantaneous.
  14. Marko, your comments regarding debug logs rang a bell but couldn't put my finger on it. Now I remember: Look at the very first post I made when I started this thread. I recall now that debug logging appeared to help back in June but I seem to remember that it wasn't a 100% cure - a bit hazy... But even so, I don't really feel that it's an acceptable solution.
  15. Thanks Marko. It might just be something to do with logging. I'll keep an eye on it. But in my case the problem only occurs at around 2-3 week intervals (at least it did...) but then a shower of several occurrences.