CJButler
Adobe Employee
CJButler
Adobe Employee
Activity
‎Feb 14, 2025
04:18 PM
Hi A.L.
Be advised that this patch is experimental and may have side effects. What the flag does is terminate the drag by reading the mouse-hardware directly, looking for button-up. Events which have not been processed yet would get processed later, after the hardware-driven button-up has been seen and handled, and might cause unexpected side effects, especially when working quickly. I'm just not sure how such side-effects might manifest in the user experience, and I'm much less confident of Pen-driven events with this flag enabled. It might be just fine for your purposes, but please be on the lookout for odd behavior. Thanks.
... View more
‎Feb 11, 2025
12:42 PM
1 Upvote
Hi A.L.,
Unfortunately I did not see the prior report dated October 11, 2024.
I've logged bug PS-151750 to track this issue specifically with the Transform Handle Mouse-Up not being detected. We'll accumulate information in that bug and see if we can reproduce.
Edit add: And thanks for the video!
... View more
‎Aug 26, 2024
10:34 AM
@DiodorS So we have this issue Open in our bug backlog. This weekend, I took a look at what it would take to sort it out. Unfortunately a blanket change to GetKeyState is not a safe thing to do. There are about 50 instances of direct calls to GetKeyState and GetAsyncKeyState, about evenly mixed, and from a casual survey of those calls, it's clear the original authurs had reasons for using deliberately one function or the other. Furthermore, there are ~ 6 additional code paths that abstract GetKeyState or GetAsyncKeyState into generic functions which are then abstracted into various utility function that are used all over the code base. So the implications of making such a change are unknown and risky. Things like using the ecsape key to cancel gestures may stop working, for example. I then looked into seeing if I could spot the specific code that trapped the ctrl key while the laggy brush waas painting, and while I can see the WM_KEYDOWN event, I cannot see which functions are calling GetAsyncKeyState while the paint stroke is being generated without adding a bunch of debugging code.
I have to get back to my assigned tasks, so I'll have to put this issue on hold for a while. The bug will go back into out backlog with some notes. I had hoped it would be easy to do, but it is not looking that way. I'll update this thread when I find some more time.
... View more
‎Jul 25, 2024
01:42 PM
Thanks - I'll add your comments to the bug and they will be seen and discussed during the review.
I think if we took your approach concept, we'd do it at the source level. I understand it well enough. The unintended ripple effects are hard to foresee when we consider all the workflows of our larger user base, so whatever we do, we will proceed with great caution.
... View more
‎Jul 25, 2024
09:57 AM
Good job - and interesting. I've added your comments and findings to the bug report. The bug will be will be reviewed next Wednesday.
As you might guess, the Photoshop Event Handling implementation is ... complex. Probably more than you can imagine. It has layers and layers of implementations applied and patched over the years. It requires support for multiple platforms (Mac, Windows, etc.). It has an unfortunate mix of an Event Queue pump, Paint Brush Stroke smoothing, real time hardware polling (we've worked hard to minimize this, but it is still required in places), Event Queue Peeking, and more. We've talked about rewriting the whole system, but our assessment is that it will be an extraordinary effort, and it will likely have a whole new set of bugs that will take years to find and fix.
One of the more useful bits of input you can provide between now and next Wednesday is your honest assessment of how this negatively impacts your day-to-day workflow, and whether your workarounds are tolerable or not.
I'll add your comments to the bug. Thanks again!
... View more
‎Jul 24, 2024
03:49 PM
Good info and Steps to Reproduce. Thanks. Logged a bug, PS-137040.
But I have to be honest, it's probably not going to get ranked very high by the Product Managers unless a lot of people complain, despite the impact on your workflows. It's probably been this way for several major releases. A Smaller Brush will reduce the lag, as might changing the Memory Tile sizes (not sure). It also depends on the Brush Tip Style; some complex brushes are lag machines when they are made big. It will take a major overhaul to change that behavior.
One thing that might be done is to somehow reject KB modifiers while a brush stroke is still painting. That code a very fussy, so we'd have to be very careful how we did that. We'd almost certainly upset someone else by changing that behavior without a great deal of care. So: High cost, which is also considered when evaluating what to fix.
... View more
‎Jul 07, 2024
01:56 PM
@Coreylehmanenterprises I am unaware of any such tools and this is not my area of expertise. This thread needs a CAI expert to chime in. Maybe some other User knows the answer to your question.
... View more
‎Jul 05, 2024
10:26 AM
At this point, we cannot reproduce without more information. Reproducing is critical for us to resolve the issue. I've logged an internal ticket to track all infomation, There was a fix for a freeze when using Flick Panning with the Hand Tool that went into 25.9. The fix was designed to prevent unsafe numerical values from being used. It's possible we missed a code path when making this fix; we'll certainly review the change. If you roll back to 25.8 AND disable Flick Panning, does the problem go away.? We'd like to collect an Event Log per these instructions: https://helpx.adobe.com/photoshop/kb/photoshop-event-logger.html
An Adobe Support Representative will reach out to you directly to collect the resulting log files.
... View more
‎May 13, 2024
10:18 AM
Thanks. Forwarded to trhe CAI team to see if that helps them reproduce. Hopefully one of them will chime in on this thread.
... View more
‎May 13, 2024
08:42 AM
I'm trying to get someone on the team engaged. They cannont reproduce. I presume you are on Windows (10? 11?) and Photoshop 25.7. Can you share a screen shot of your Preferences > History and Content Credentials Screen? I will make sure they see that.
... View more
‎May 12, 2024
10:28 AM
Is this unexpected behavior blocking any workflows? That will elevate the priority.
I did find these threads. Did you see them? (Never mind - I see you did)
(deleted links)
Cory is the one to carry this forward. (Cory: this looks like CAI? ref CAI-4608.) Caity is now in another role in the organization. A bug will need to be logged, the issue will need to be reproduced, and it will need to be assigned to the correct team. (I'm not a cumminity manager, I'm just a developer in one corner of Photoshop. And I'm not the right guy to hunt this particular issue down.)
But - from a comment in the ticket above, it suggests that there is "an old style sidecar has the PS preferences ignored". That might be a clue to preventing future creation. (But I don't know since I don't have t his problem.). Also, notes indicates it is safe to delete these side-car files.
... View more
‎May 06, 2024
10:33 AM
Just to be clear, Rick is The Guy for ACR support issues. But he (or someone on the team) needs to be able to reproduce the problem before progress can be made.
... View more
‎May 05, 2024
04:43 PM
2 Upvotes
@Kuttyjoe Might you have installed 24.7.3? That splash screen looks like it belongs to major version 24. I think.
... View more
‎May 05, 2024
04:40 PM
2 Upvotes
I'm not sure what version that is. I've seen the graphic before, but I don't associate it with a version (I've see a LOT of splash screens over the years, some release, some beta, some prerelease, some internal). As far as I know, 25.7.0 is the latest offical release on Windows. The Main branch is currently tagged as 25.9, but it has not yet been moved to Release Staging. I don't ever recall seeing a 25.8 in the pipeline, but I am not sure. I think there was a special release for demonstration at London MAX a few weeks ago, but I'm not sure what version that was; Maybe another Adobe person watching this thread knows.
I also checked the internal communications on offical build release versions. I did not see any mention of a version 25.7.1.
So, I don't know what you managed to install. In any case, 25.7.0 is what we should looking at. And I think 25.9 will be the next official release.
... View more
‎May 05, 2024
02:55 PM
2 Upvotes
So far, I can't reproduce with 25.7. I've been trying for a bit, given what you described, so I'm probably missing some crucial ingredient.
... View more
‎May 05, 2024
02:35 PM
1 Upvote
@Kuttyjoe Let me clarify.
>> Are you saying, if I were to select the hand tool, use it, then switch to another tool, then see if it's still on the hand tool?
I'm looking to see if you get into a stuck mouse-down state without keyboard chording being involved. i.e. Just use the Hand Tool and move things around. Mouse-up. Are you stuck? Or just use the Zoom Tool and mouse-down, sweep left or right, then Mouse-Up. Are you stuck?
Keyboard chording can affect which tool is being used when a mouse-down happens. Event handling properties (for that Tool) are established at mouse-down. If the Tool is switched to an Alternate Tool due to keyboard chording or some other behavior that closely follows the mouse-down, like the mouse being over a special control surface in the document (e.g. a rotate grab handle), then problems may occur, especially if the Tool cursor has not changed yet. I just found a bug with Pen and the Marquee Tool, where if you were working very quickly, and you tried to create a new selection over an existing one, we try to switch to the Move Tool. In that case, we might miss POINTERUP, because the Move Tool only deals with Mouse-based events and is not Pen aware. That's why the precise details of what Tool you start with and comes first (mouse-down or key-down) matters, as if affects how control flows through the code. And it is also why I'm trying to be as precise as I can, because without those steps, we are very unlikely to reproduce. Our muscle gestures may be just slighly different from yours, and you see the bug, and we don't.
As far as your frustrations go, I certainly understand. All I can do is try to find and fix the problem. To find it, I need to reproduce it.
... View more
‎May 05, 2024
01:51 PM
3 Upvotes
@Kuttyjoe Please confirm the following:
Windows (version?) Thanks I see you added that.
No Pen attached?
New default document?
After you open the document, are you in the Move Tool by default? If not, which Tool?
What happen when you don't use a keyboard chording gesture? i.e specifically choose the Hand Tool or Zoom Tool and no space bar or control key on another Tool. (I'm not suggesting this is a fix, it's just data. It is important that keyboard chording works.)
Are you using the keyboard chording at the same instant that your mouse down? i.e. Is it key down and mouse-down at very close to the same instant, or is key down first, see the cursor change, then mouse-down?
It sounds like a lost-mouse-up situation similar to what we've been hunting all this time. After clariifying steps, we'll try to reproduce. If that fails, we'll probably ask you to generate an Event Log file. Thanks.
... View more
‎May 05, 2024
01:10 PM
3 Upvotes
@Kuttyjoe Can you give us a more detailed description and more precise steps on exactly what you are doing in 25.7 that exhibits "Tools still sticking within 3 seconds of opening a new document". Step-by-step, what OS platform, exactly which tools you are using, exactly where do you click , and what does "sticking" mean? Is this the lost mouse-up problem? or something else? We'd like to try and reproduce. Thanks.
... View more
‎Apr 24, 2024
07:09 PM
@NikVladi You're welcome.
... View more
‎Apr 24, 2024
05:35 PM
@NikVladi Please try version 25.7. You can leave older versions installed. (I have 8 versions on my machine).
... View more
‎Apr 24, 2024
04:58 PM
@NikVladi In my earlier reply to you, I asked you several questions and I provided information on how to invoke the Event Logging mechanism. You did not reply. Did you miss that?
Event logging has performance overhead, which is why it is not enabled by default. There is also the Photoshop Help > System Info... report, but it does not have information which is helpful for understanding Event Issues. Most of the information in that report are focused on plugins and GPU information, which have been historic problem areas. It does confirm things like platform, OS version, Photoshop version, and basic system configuration, which can be helpful at times.
Ultimately, though, we have to reproduce the issue before we can make meaningful progress. Photoshop version 25.7 has been released. Can you confirm this is the version you are using?
... View more
‎Apr 24, 2024
09:31 AM
@Benjamin32760768enyn I'll just add to what Chad said. I have no idea from your description how to even try to reproduce the problem you are experiencing. Can we get some basic infomration?
OS Platform (Win, Mac)? Photoshop Version (e.g. 25.6)?
A more precise series of steps that exhibits the problem (e.g. New default document, etc.). The fewer steps the better.
Do you have any third-party plugins installed?
When Photoshop becomes unresponsive can you recall what you were doing just before that point in time? In as precise detail as you can.
When we have this information, we'll try to reproduce on our side.
... View more
‎Apr 22, 2024
05:11 PM
@Jqqerry I don't think that drawing problem has anything to do with the original problem as reported in this thread (stuck mouse-up).
This is out of my areas of expertise, but it sounds possibly like a drawing framerate issue. It may be related to the character of the document content or the GPU settings. Or both. I recommend you start a new thread on this issue, especially if it is blocking your use of the latest version of the product. Of course, you'll need to be clear as to why it is impacting your workflows. A new bug will need to be logged; it will be probably assigned to the Photoshop GPU team for investigation.
In general, when using the Hand Tool, I would expect screen caches to be in effect, and for most modern machines, this should be very fast and seamless. If the caches are being invalidated for some reason, that could the cause a redraw to occur, and could be the reason for the laggy drawing experience. Adobe QE will want a sample document and precise steps to reproduce.
(Maybe some other Adobe person watching this thread has some ideas for you.)
... View more
‎Apr 22, 2024
07:42 AM
1 Upvote
> I haven't worked in Photoshop for a long time, today there was something I needed to do... the latest version...
Hi Nik, can I get some more information from you?
Can you confirm the actual version number where you observed the problem? And the platform (presumed to be Windows)?
Are you using a Mouse only, or are you using a Pen/Stylus? Per the last Sticky Note for this thread, there have been changes in 25.6 for Mouse users, and there are changes in 25.7 for Pen/Stylus users.
Can you describe exactly the gesture that is failing for you? Is it the lost mouse-up when using the Hand Tool as described in the original thread, or is it something else? If something else, can you describe the steps in greater detail as precisely as you can: What Tool are you using? Are you using a keyboard chording method (e.g. Space Key in Move Tool to get to the Hand Tool)? What are you clicking on (some control point?) How are you moving your Mouse (or Pen)? When are you releasing the Mouse Button (or lifting the Pen), and then what happens? Please give as precise steps as possible so we can try to reproduce.
I crafted an Event Logging Tool that can help us look at things in greater detail. If you can reproduce stuck mouse-up on 25.6 or later, we would like to examine a log file from that event when it happens.
If you are curious as to why this bug took so long, there are a number of earlier posts that have more information. The major blocker for us was our inability to reproduce the bug at Adobe.
... View more
‎Apr 13, 2024
03:18 PM
And, this is a bit of a long shot, but again presuming Windows: Do you have "ClickLock" enabled? It is a Windows Accessibility thing for Mice. If enabled, it has some settings for when ClickLock engages.
... View more
‎Apr 13, 2024
02:59 PM
> This mainly happens for me with the transform tool. I am just clicking the transform handles in order to free transform a layer, and it happens immediately when I click the transform handle.
OK, in the meantime, let me clarify the presumed gesture as precisely as I can so I can try to reproduce if possible:
Presumed: Windows
Presumed: Multi-layer document with a single layer selected. The Document Window is docked (not floating).
Presumed: You are using the Move Tool, "Auto-Select" is enabled and "Show Transform Controls" is enabled.
Presumed: Contextual Task Bars are enabled (small button bar over the content)
Presumed: You do NOT use ctrl-T to initiate the Transform command
Presumed: You hover over a control point (resize cursor), or near a corner control point (rotate cursor)
Presumed: You click (mouse-down, mouse-up) on a control point, and the mouse sticks to the control point and initiates the transform as the mouse moves. You have to hit the Escape or mouse-down again followed by Undo to get back to no-transform state.
Is this what you are doing? (I'm looking for very precise instructions for the fail case, as there are a lot of ways to get into Transform Mode.) Feel free to correct or clarify.; be as precise as you can, breaking clicks down into mouse-down, mouse-move and mouse-up.
Does this happen every time for you, or just occasionally? Does the speed of the click (mouse-down, mouse-up) appear to affect the behavior?
(Note that we also have two very experienced quality engineers watching this thread - they will also try to reproduce on their machines.)
... View more
‎Apr 13, 2024
10:34 AM
> ust to update on this: Unfortunately this bug is NOT fixed in this latest update, it is still happening for me after updating to 25.6
Since you are first person on this thread who says they can still reproduce with 25.6, a few questions if you don't mind:
Are you using just a Mouse, or are you using a Pen/Stylus? There are fixes in 25.7 intended to help with lost Pen Events.
What kind of gestures are you doing? Is it the Hand Tool as originally reported, or is it something else?
Would you mind generating an Event Log file using the instructions here? And making that log file available to us on a site like DropBox?
Thanks.
... View more
‎Mar 22, 2024
06:31 PM
4 Upvotes
> I'd love to see a technical write-up of exactly what was causing this
I posted something earlier with details (March 8).
Our inability to reliably reproduce was what prevented us from making progress on the bug.
My assessment is that the dominant element that was preventing us from reproducing the bug and making progress was that Photoshop was not calling the Windows OS function EnableMouseInPointer(false) on startup. Photoshop is an old application, and this function was added when Microsft rolled out Pen/Stylus support. Because of how Microsoft implemented the behavior of this function, if it is not set at application launch, the application property bit can be set to 'true' ANY time by ANY subprocess. For most of us at Adobe, we theorize it was not being set and so stayed 'false' for the session. For users consistently experiencing the problem, we theorize that something, some code somewhere, was consistently setting EnableMouseInPointer(true) for them when they ran Photoshop. We still do not know what that was, but it literally could be anything, including Chrome libraries, or any third-party technology we embed in Photoshop. It's also possible that Microsoft was setting this property to 'false' earlier for us as a legacy application, but that we fell off their "special cases list" (yes - they do that sort of thing. So does Apple.) and things started breaking as soon as the Photoshop version advanced enough to not longer be considered a special case. Or perhaps a Windows update somehow lost track of treating Photoshop as a special case. We don't know. EnableMouseInPointer(true) tells Windows that Photoshop is "pointer-aware", and that it is ok to send us just Pointer messages, and not synthesize mouse messages. But Photoshop is a legacy application and it is only partially pointer aware: for example, Painting is largely pointer aware with things like pen pressure, but much of the other Tracker code (like the Hand Tool or Marquee Tool) is not. The older code assumes just mouse behavior and we expect to get mouse-button events. No WM_LBUTTONUP event for those trackers means that we never see the mouse-up. And that's the observed bug.
Once we figured out that we needed to explicitly call this function and declare ourselves a legacy application, this allowed us to make progress. Windows was not consistent when it would decide to not synthesize WM_LBUTTONUP, leading to further unpredictability. In our explorations and debugging, we also found a few other problems with our usage of the Windows OS function PeekMessage when it was called with PM_REMOVE, especially when the Windows Message queue was very busy with a lot of queued messages (e.g. fast mouse or pen work). We're still reviewing and fixing those bits of code up to make them more reliable.
Personally, I think this method of requiring an application to declare itself as a legacy application via EnableMouseInPointer(false) was a terrible approach by Microsoft when it implemented Pen/Stylus event handling years ago. It is a very obscure OS function call with poor default behaviors.
... View more
‎Mar 22, 2024
02:20 PM
1 Upvote
> Now GPU acceleration turns off randomly as I'm trying to work.
Please start a new thread with this problem, along with your system information and anything you can think of that might help Adobe reproduce this problem. GPU issues are handled by a different team, and we'll need to let them investigate. This thread is too large; any information posted here runs a high risk of being lost. Thanks.
(This is something that would normally be done by a Community Manager. Alas, I am not such a person.)
... View more
‎Mar 08, 2024
06:24 PM
5 Upvotes
@chrisell99 >> can you summarise (at a high level) what the code change(s) is that might have cured this?
Two things changed:
(1) We now call the Windows OS function "EnableMouseInPointer(false)" on start-up. This is the correct behavior for legacy apps like Photoshop, where the majority of event handling assumes a mouse (like.the Hand Tool tracker), even when working with pointer devices. We were not calling it at all before this. This function was created by Microsft when they started supporting pointing devices: it has the unfortunate property that anyone can set it at any time - and fix the behavior for the session. We were not calling it, but that did not prevent any of the multitude of embedded 3rd party technology within Photoshop from calling it. E.g. Chrome libraries. Through experimentation we know that when set to "true", bad behavior in Photoshop follows. My theory is that there is some startup sequeunce (or later gesture) that traverses some of our 3rd party libraries, causing this to be set to "true" consistently for some users. I don't know where that might be. We've looked for the culprit, but have not found it since we have had such a difficult time reproducing the behavior and much of the source code for such libraries is not visible to us. In any case, the setting is now established correctly at Photoshop launch.
(2) We have changed how we interact with the Windows OS function PeekEvent in certain locations, especially when dealing with Mouse events and Trackers. When the Windows event queue gets busy, this function is capable of behaviors contrary to the expectations of the original developer who called PeekEvent. Photoshop uses PeekEvent throughout the code base, and has done so literally for decades. But it does have non-intutive behaviors under some circumstances.
As mentioned earflier, we're not done with scrutinizing our PeekEvent calls. We still have a known problem with Pointer (Pen) events, especially when when the Windows event queue gets busy. We're working on that.
@Jqqerry >> Can you do a research based on this user's experience?
I'm sorry, I don't understand what you are asking.
If it is because he is seeing problems in 24.2, then that is not suprising given what we now know. We have had reports of lost mouse-up behavior going back to 23.2, and possibly earlier, but very rarely. As @Goyko suspects, there is are other bits of shared technology that get updated through triggers other than a Photoshop update. For example, a Lightroom update will also update Camera Raw, which Photoshop uses. I am very sure there are other such things. But even if you asked me, I could not enumerate them all.
... View more