Sandboxie Plus 0.7

@Survivor
I never said it couldn't be complicated or that additional info about different scenarios which do or do not show a certain problem wouldn't help. And the "veteran"-comment just meant that if users report a problem in this forum they usually should be taken seriously and that the problem reported by them does exist for real and in all probability hasn't been caused by some recent blunder they might have introduced themselves into their systems by culpable negligence.

So if someone reports in here that the problem started with Chromium-version-upgrade from v.89 to v.90 and I am able to confirm, that current Opera_stable is not affected (still using Chromium v.89) while Opera_Developer refuses to launch (using Chromium v.91) - now that should tell you something! This does not look, because of reproducibility, like some strange Windows-upgrade-glitch any longer.

As for my suggestions - if you just want to open the GUI from the tray-icon, well, a left double-click will do, without any danger of mixing up options in the right-click-menu. Personally I've never had any problems with choosing the proper menu-entry in a right-click-menu, not even in explorer where there may be dozens of entries.

And yes, I'm well aware of the fact that I could edit the ini-file with a text-editor at any time and temporarily comment-out the entry in question. But then I would have to search and find the appropriate entry first and know about its syntax before being able to successfully edit it and there would still linger the question if that change would take effect immediately or only after a reload or maybe even only after a re-boot of the PC.

A GUI is meant to make things easier in such cases, allowing changes without the knowledge of the exact syntax or location of the appropriate command/entry. And I really can't see why it shouldn't offer an easy way of temporarily changing a forced folder/program entry to off and on again for testing-purposes as well.

 

Disabling hardware acceleration does not allow Vivaldi to successfully run inside sandboxie with a chromium core < 90 . That might work for Edge / Chrome (or whatever) but it does not work with vivaldi.

 

The issue is caused by anything over Chrome/90. Try Vivaldi 3.8.2259.27, which uses Chrome/90.0.4430.86. Does that work?

 

Maybe this is not a useful idea, but since some of us have no issue with sandboxing Chrome while others do, I wonder if it would help David -who seems to be the one working on Sandboxie Plus if I understand it well- had access to a machine that has no problems with SB + and Chrome. Using the "Help at a distance" option.
I just installed the latest Brave again unsandboxed, but it still didn't work when used sandboxed, so I removed it again. But Chrome works fine.
I would be willing to hand over control of my laptop to David when I'm not using it, in case that would help.

 

Damn. I was really hoping I had messed up in some way and there really was no problem!

 

Point was, some have an issue, some others not, both infos are valuable and might lead to solutions. Windows and or Program glitches, do not start me with that. It was not said, that "you" did cause the issue by doing some, I said, it can be anywhere. Sometimes it is even difficult to pinpoint to a specific action, maybe another one you did a day before, caused it, but only the update from Chrome is showing the error now.

For the context menu, sure you can have hundred options, you can make all items available, but where does it get to much. It was just my opinion. For example due to the behaviour that the sandboxes are all in the send to folder, this menu is not usable, can I change it, yeah, could, but in the end I don't care that much about it anymore. In older years I customized it to death.


The ini file takes immediate effect, after restart of the program in the box. Tested with the closed path of drives and fileexplorer. Change, start explorer again and drive is open or closed depending on the session. So no restart of the PC or interface, the settings seem to be read fresh anytime you start a program in the box. In this case, the first instance had the drive locked, after change second instance had it available. So it is on the fly.

And yes I agree to the GUI, I prefer this over tweaking inis, regs, .... I like it if both works, flexibility.

 

There is no disputing the fact that reasons for sudden malfunction of a group of interdependent programs can be very complex and to find out about the true cause can be very tricky - but very often after some research the culprit becomes quite obvious.

Once it has been established that 3 or more Chromium-based browsers reproducibly will start in a version using Chromium <v.90 and will NOT start in any version using Chromium >=v.90 on different computers showing the problem - now that should tell you something!

As for the context menu the option to delete the contents of the active sandbox is not just any other garden-variety-option. Every now and then (with the option set to auto-delete the box-contents after certain leader-programs, often browsers, will have shut down) Sandboxie will "stumble" upon shut-down and give an error-message that the contents could not be deleted successfully. This is presumably due to the effect that the (temporarily hanging?) OS might keep some files open for a delayed period of time and therefore will deny access to deleting those files. So Sandboxie advises the user to delete those manually after shutdown of all processes. And since such behaviour will occur every now and then it would be convenient if the user could remedy this issue right from the tray-icon without each time having to open the main UI first, then looking for the sandbox affected (most recently closed), right-clicking it and then choose delete-contents.

Btw. this delete-contents tray-icon-option I'm asking for has been - and still is - part of the few tray-icon-options for the Sandboxie-"classic"-tray-icon, for good reasons one might assume!

 

Reporting the same Edge hang as others have reported now, with and without hardware acceleration. So, is this universal among users now? Edge version is 90.0.818.49. There were like four updates to Edge in about 10 days' time: .41, .42, .46 and the current .49. Hard to say what exactly changed over those four versions to trigger this latest issue.

Firefox 88 still working fine w/HA off and on. But I think 89 will be coming out in like five weeks, they say. Anyone running a Nightly build with Sandboxie Plus?

 

Um, have you tried Microsoft Edge > Modify > Repair. I'm not suggesting Repair will sort reported Edge hang. Just curious if you've tried Modify > Repair.
FWIW ~ I'm not seeing Edge 90.0.818.49 hang.

 

OK, well that helps, that it still works w/HA disabled.

I can't repair Edge--this is error. I can't uninstall it now either. Looks like it's stuck where it is. I'll look up possible fixes then. Thanks I don't typically run it sandboxed anyway but this needs some attention.

Spoiler

 

Edge 90 works perfectly for me with hardware acceleration enabled. I use Sandboxie Classic 5.49.7. It opens immediately in his sandbox.

 

@plat1098 Oh....jeez...sorry for mentioning Repair.
Note: remove Force Edge in Sandboxie Settings before running Edge Repair. Edge Repair calls Edge after installing Edge.

Spoiler: Edge Repair

20H2 (19042.928)

 

Thanks, I just put Edge in the Default box and allowed access to my profile, now it works as it always did: beautifully, but only with hardware acceleration off. Good, that's a load off.

by the way: Edge just updated to version 90.0.818.51

 

FWIW ~ I run discrete Edge box with restrictions.

Curious, how/did you sort error in #239?

 

I just switched to the Default box from my Edge discrete box. Sometimes these things seem to get corrupted somehow. All I know is that Edge works in the Sandbox now, provided hardware accel. is disabled. Like you and stapp reported already.

Next time this or another glitch happens, I'll try this or creating a new sandbox first before posting unnecessarily here.

 

No worries. I was confused by "Windows Installer prompt" in #239.
Not what I see with my Edge Modify Repair #241.
Happy to read your Edge is back in business.
Thanks, please excuse my confusion. Regards w Respect.

 

Anyone any idea how to force Google Chrome to run without its own sandboxing-protection?

The start-option "--no-sandbox" doesn't seem to work any longer for Chrome itself. At least it won't make any difference as for not starting up the browser within Sandboxie-plus on those systems affected, again with or without hardware acceleration

 

@DavidXanatos the old issue with games is back. 0.7.2 was fine and 0.7.3 brought the old issue with Sandboxies RPC waiting for thread and complete crash of the small game Noita is back.
Noita brings (WerFault.exe (6364): SBIE2101 Object name not found: Unnamed object, error OpenProcess (C0000022) access=001FFFFF initialized=1)

As reported in the other tread as fixed, with 0.7.1 and OK with 0.7.2 https://www.wilderssecurity.com/thr...ith-signed-driver.434924/page-36#post-2991535
Now back with 0.7.3 and 0.7.4
The good thing is, this time I can pinpoint to the version. Whatever was done in 0.7.3 got the error back. This time I can exclude Windows, as switching back to 0.7.2 all works flawless again.

On a side note, the installer is working great, overwriting all without restart, even if you used the SBIE+ already.

 

algol1, my previous post, just makes me think, did you try to go back some versions of SBIE and see if this makes a difference with Chrome. Worth a try.

 

Indeed I've done that and reported about it earlier in this thread.

Sbie v0.7.4 = v.0.7.3 = v.0.7.2 Opera v75 (using Chromium 89) works flawlessly. Still does.
Sbie v0.7.4 = v.0.7.3 = v.0.7.2 Opera Developer v77 (using Chromium 91) refuses to launch. And the current Chrome doesn't launch either. In Sbie v.0.7.2 though there was no error-message-popup yet, the launch-command just went "into the void".

Chrome itself before v.90 I couldn't try because of its sucking auto-update but after reading about it I've tried if the "--no-sandbox" option with current Chrome would make any difference. Sorry to say it didn't (not sure if that option has been recognized any longer at all by Chrome as has been the case in earlier versions) and neither does turning hardware-acceleration off in Chrome.

However your above cited Sbie-error-message in correlation with the game DOES SOUND QUITE FAMILIAR!

"WerFault.exe (11992): SBIE2101 Object name not found: Unnamed object, error OpenProcess (C0000022) access=001FFFFF initialized=1"

reads mine when Chrome refuses to launch.

 

@algol1, sorry if I missed that mate
So we have the same error, which we have in common with @Harmony_ , which might give a hint to David, even if it seems to be quite generic.
Difference is, that it happens for you also with 0.7.2 while for me with in this case Noita game (the little witch) is OK with .2 but .3 starts to throw this error.

Do you also have some weird anomaly with the sandboxies RPCS, or is there no delay?