2.6.0 Beta Build RC1 (3058)

[Patrick Connor]

 Hi All,

We have just made available an Release Candidate RC1 build of 2.6.0 beta - version 2.6.0.3058 - This was created to address bugs and make improvements to the new features announced in this post

The fix list for this build is here.

Improvements & Changes in 3058

• Colour Tag options added to top Layer menu (for reference in macros)
  • This allows colour tag commands to be recorded as part of macros. Previously, they were only available on the Layers panel right click context menu, which is blocked whilst macro recording is active.
• Colour Picker Tool: Alt modifier to "apply to selection" added to the tool, to act like the Colour picker on the panel
• Machine Learning features are limited to Apple Silicon iPads.
• CMD+Return on macOS (CTRL+Return on Windows) to commit text input: Enter removed (just use Return)
  • When in text input mode (e.g. actively typing with a text tool), you can now use CMD+Return on macOS, Ctrl+Return on Windows to instantly commit the input and switch to the Move Tool. Accommodations have been made to allow other mappings around this for users who want to bind similar shortcuts to text operations: additional modifiers can be used, so for example you could map page break to Shift+CMD+Return. On macOS, Return and Enter are regarded as separate inputs, so you can also map a command to CMD+Enter (on the numpad).

 

[loukash]

1 hour ago, Patrick Connor said:

On macOS, Return and Enter are regarded as separate inputs, so you can also map a command to CMD+Enter (on the numpad).

For the record, fn+Return has also always acted as Enter on Macs with the fn key.

[Latens]

2 hours ago, Patrick Connor said:

Colour Picker Tool: Alt modifier to "apply to selection" added to the tool, to act like the Colour picker on the panel

Sorry, but this is just another speedbump on the road.
Alt+dragging should commit the color in all pipettes/pickers.
The raster brush tool can only work with Alt key pressed to collect a color from anywhere but you still need to apply it yourself.
I think for consistency sake it should apply to the others as well.
e.g. Alt+drag is collecting and Applying
Dragging alone is adding to the well.
In a way one could collect the color and add the Alt key to the action later, while still holding the mouse button down, to commit the color chosen.

Perhaps you could explain the adding of a (child)pixel layer when painting over a vector object with the raster brush tool because there's no setting in the assistant for it?
There's an option to Painting with no layer selected>add pixel layer and paint
There's an option for Erasing from vector layers>place mask and erase.
There's an option for Other brushes on vector layers where there's either take no action or rasterize and apply brush.
But there's no setting for painting over selected vector objects getting a child pixel layer.
*Edit: There should be an option like adjustments/masks>add based upon selection(aka child nesting) or as new layer.

[kaffeeundsalz]

2 hours ago, Patrick Connor said:

Machine Learning features are limited to Apple Silicon iPads.

This is sad. I can see why legacy Macs were cut off. But iPads never had any Intel processors inside.

[Frozen Death Knight]

Good to see the Smudge Brush Tool getting the bug fixed that caused brushing artefacts. Reported it many years ago and it was a big turnoff when I first started using Affinity due to how much I use this tool. Thanks for fixing it.

[fde101]

On 1/21/2025 at 7:00 AM, kaffeeundsalz said:

iPads never had any Intel processors inside.

Apple doesn't support Apple Intelligence on the A series processors either, only on the M series processors, even on the iPad.  I don't think this is strictly an Intel vs. ARM thing; there is probably some other resource that the A series chips are short on or missing which is causing a problem for this.

[mdriftmeyer]

On 1/21/2025 at 4:00 AM, kaffeeundsalz said:

This is sad. I can see why legacy Macs were cut off. But iPads never had any Intel processors inside.

I think it has to do with the performance cores. The A17 bionic only has 2 performance cores. The M1 has 4 performance cores alongside 4 efficiency cores. The M2 offered 4, 6, 8, 16 performance core configurations. 

With the advent of the M4 line only the entry level iPad and iPad mini lacked an M processor. This way they can focus on M series only moving forward. Sooner or later even the iPad entry will have an M series processor.

[kaffeeundsalz]

This is plausible. Yet, Affinity's ML features worked just fine on my non Apple Silicon iPad Pro. A bit slowly I admit, but not to the extent that it would have been unuseable.

[fde101]

5 hours ago, mdriftmeyer said:

I think it has to do with the performance cores

There is also a difference in how much memory is available; some of the A series chips used in the iPad line only have 3 GB, with the larger ones having at most 6 GB, compared to at least 8 GB for the M series chips.  With all of the resources of the chip (CPU, GPU, etc.) sharing the same pool of memory that could be a meaningful factor as well, depending on the size of the model in use, combined with  memory used by the OS, application, document content, etc.

I suspect that is a bigger factor than performance cores, but it is just a guess on my part.

 

4 minutes ago, kaffeeundsalz said:

Yet, Affinity's ML features worked just fine on my non Apple Silicon iPad Pro.

Even if the current model reasonably fits in available memory, Serif could have concerns that they will not be able to update the model going forward if they constrain themselves by allowing current chips with tighter memory constraints to make use of the feature.  Requirements could increase fast, and if they don't take steps to allow for future expansion, then they could wind up cutting off currently supported models too quickly in order to advance the capabilities of the feature.

Also, while I recognize that you are referring to the M series chips, note that ALL iPad models ever released have used Apple Silicon, as the A series chips are also technically Apple Silicon.

[Schindi]

No ML Futures on my IPad Pro M1 anymore! Why?

[Patrick Connor]

43 minutes ago, Schindi said:

No ML Futures on my IPad Pro M1 anymore! Why?

It's a mistake and will be addressed before release