Some combination of toggling fullscreen & changing the resolution in Windows 10 (VM) caused a crash.

Possibly selection outlines & highlights need to be torched/recached whenever the module cache is?

Possibly for reasons related to mysterious Jump-Point Search voodoo, moving modules that are far away from wires can still cause those wires to shift between two similar-length paths.

Not sure whether it's the lock screen or sleeping, or whether it also happens on Windows 7.

When picking up groups of modules (possibly also single modules), they sometimes flicker violently upward for a single frame.

After switching from fullscreen to windowed (or vice versa), the next mouse click event seems to go missing.

Numbers are fine. Others modules keep their color when overlapping.

Whether it is by clicking the button in the Right-click UI, the one in the toolbar at the bottom or hitting "c".

No crash report, only "Process has stopped working"

Nothing new in error.txt.

Trying to wire from one input to another, the destination input was colored red (invalid), but was still connected when the mouse released. (Both inputs were on latches; the source was a Number.)

Inconsistently, modules will fail to "push down" modules below them when edited to become larger, and will instead overlap lower modules. This may have something to do with the order in which modules were placed on the level.

I'm not sure I was clear : it actually did copy the game canvas in the clipboard, from when you added the screenshot button, until the last version.
Now, it copies the windows behind the game instead of the game.

I guess windows get the PrintScreen KeyDown event too, and handled it by copying the game. Now, for a reason I don't know, it copies another windows, as if it had the focus instead of the game.

https://msdn.microsoft.com/en-us/library/system.windows.forms.clipboard.setimage(v=vs.110).aspx
http://www.alecjacobson.com/weblog/?p=3816

Many times after testing a modification, I'd like to cancel it, so as soon as the test is finished, I hit Z, but that does not work. I'm forced to pick my mouse and click on the button.

Although there is no instructions to decode, OpSel is addable/pastable in "Shift one".

I'm a moron. Sorry.

Haven't seen this in a while, but never actually fixed it.

Sometimes this creates a cool 'rewinding' effect - it would be nice to make that happen consistently.

When the player has zoomed out, the minimap doesn't correctly reflect what's on-screen, either in terms of the visible rectangle or in the sizes & proportions of modules.

Wires between my RegFile and the I-Sel should be all grey, right ? (I verified each wire, they are each between an output slot of the RegFile and an Input slot of a I-Sel)

puzz-Shift One.zip

When clicking on wires that have control points (especially diagonal segments), an assertion sometimes fails in WiringState::trimWire().

Not really an issue, but when I click Test, the values in the modules briefly change (during the fade to black). When the test begins (after the fade), they are back to "normal".

Seem to be the last values of the last test set that stick in after the actual test by the algorithm.

I have noticed this in Spares, but I guess it's in every levels.

• ConstMarshalledData ignores control points when comparing equality of states.
• Copied & palette groups do not save control points.

Not sure if this is an OS X/Windows issue or something else, but it completely breaks time pars regardless.

The mouse cursor changes, modules highlight, etc, when the mouse is not actually inside the game window. This seems to be based on the last mouse location before it left the window.

Was never able to reproduce this.

Stashed as direct_connect_on_edit . Related to #21 .

Failure to create color cursor sometimes occurs on launch; seems inconsistent even on the same system. (Haven't reproduced it myself.)

In "Dragging", I made the second connection, and then immediatly grabbed a module and dragged it, and the screen didn't appear.

When ‘writing’ an already-existing value, writers’ light turns on, but mem sidebar’s doesn’t. Should be consistent, one way or another!

(Instead of passing currentmem to memsidebar, pass sim?)

I was hitting Y to redo some undone actions (wiring).

Can be moused even when over the bottom-right 'hollow' section, where the test # UI lives.

Which breaks the selection outlines (and other things).

Probably the same bug as #19 and #3 .

A wire was somehow stuck inside an adder, rather than pathing around.

Dragging from Writer input to writer control, when input was already wired to a Latch output, set the control to ‘-’.

Only ever saw this one once.

Only ever saw this once.

No crash report, nothing new in error.txt.

It does either if I add this by the new Right-click UI, or by the palette menu.
It does not crash if I don't include the I-Dec in the selection.

No crash report, only "Processus stopped working".

The selection included a counter (Latch, Adder and Number 1) and a I-Dec, all connected.

Error.txt :

1376044: CreateTexture(D3DPOOL_DEFAULT): INVALIDCALL
58551401: CreateTexture(D3DPOOL_DEFAULT): INVALIDCALL
525: CreateIconIndirect(): Paramètre incorrect.


178: CreateIconIndirect(): Paramètre incorrect.


175: CreateIconIndirect(): Paramètre incorrect.


186: CreateIconIndirect(): Paramètre incorrect.


169: CreateIconIndirect(): Paramètre incorrect.

but I suspect this to be from previous crashes (Cursor crash at start), so it seems there's nothing here for this issue.

Edit : error.txt confirmed to not change after other same crashes.

On Windows, clicking the fullscreen -> windowed toggle doesn’t immediately change the game window size.