Notes:
PEP compliance discussion: https://phabricator.kde.org/T6588

Forgot to reference this in the commit :(
347e631531eb

Update to diff - add some documentation, update cmakelists.txt

BTW If the ui file is in the same directory as the script it will run from the command line. So you can run it stand alone to test.

"The user's resource directory is fixed on each different OS. Settings/Manage Resources/Open Resource folder will always take you there."

So, what's the story with PyQt5 - if you get the Windows build or the linux app image it's bundled into the package?
But if you build from source (and presumably for package management users) you have to have it installed as a dependency to run the scripts?

I think I can harvest QIcons from existing widgets. I was more wondering whether there was some sort of icon resource where I could find them more easily.
Accoding to the console if I try Krita.icon("name") it can't find it but falls back on QIcon::FromTheme("name")

https://phabricator.kde.org/R37:28547a852ee98330ef2a985b7b74594854623e0b

https://phabricator.kde.org/R37:6dad1dd2ca67a442ff397e8135e0c26cfc9e8cb7

A user can test for whether or not they can import krita to determine whether they are executing in krita's context I guess.

N802 is lower case function names - will need to be ignored
The sample scripts will not pass PEP8 requirements.

I had thought about doing it as a setting that stored the previous size (perhaps geometry) of the scripter window as at close. Anything else would require a line entry in the settings dialog => new strings in any event.
I need to work out how the Krita settings work first.

Agreed. This was faster.

https://phabricator.kde.org/R37:bec0891da4edc684a0b2aa659a946cf813a04ec7

https://phabricator.kde.org/R37:bec0891da4edc684a0b2aa659a946cf813a04ec7

Addressed in:
https://phabricator.kde.org/D9317

Removed magic string "users_script"
Consolidated https://phabricator.kde.org/D9527 - ie remove unused imports, exec compiled object code

Is there a way to access the button bitmaps from Python? For example, if I want to put an undo button in a docker can I get a copy of the undo bitmap to apply it to the button?

So, I don't have to stick to exactly what I've put in the revision?
For example, the commenting code had a couple of bugs in it, some of which have been approved and some of which haven't. Can I just push the code as debugged?

I have organised the developer access but am not sure what now? Are changes pushed separately through git, quoting this review?

I am not sure how to implement tests for Python plugins. The plugins rely on the Krita environment. Does Krita have a headless mode?

Hmmmm...
I was losing track of what local commit matched what diff.

I did some edits to the intro para of one of those pages.

For things (like resources) where you need to already know the resource type in order to get the resources include a list of what strings will work.

On both pages
python -> Python
Me editing a comment
woohoo!

There seems to be another bug when commenting the first line of a script. I need to fix.

on https://docs.krita.org/How_to_make_a_Krita_Python_plugin

on https://docs.krita.org/Introduction_to_Python_Scripting

Earlier diff was too complicated.
More succinct code attached. Should work under Python 3.3 and Python 3.4
Patched as against git b37324dc77

In T6588#121885, @victorw wrote:

In T6588#121427, @brendanbd wrote:

There should be guidelines about where scripts ought to persist their data.

Good point. Do you have any actual use-cases in mind?

As it turns out, the coding style will be inconsistent either way.
If you choose Qt style, then references to the Python standard library will be inconsistent and vice versa.

You should also think about how scripts are distributed and installed. Will any scripts be hosted on phabricator for example? How do people install scripts? How should scripts be packaged?

I have assumed that you want to conform to C++ variableName style, ugly as it is.
There should be guidelines about where scripts ought to persist their data.
I also think there should be thought given to how scripts are listed in menus the flat tools-scripts structure could quickly run out of space.

I have worked out how to do a consolidated diff against master if anyone wants these revisions as one consolidated revision

Oh, thanks. Done
Not sure that's working though.
I have ended up with diffs that are cross dependent over two revisions. :(

Previous patch introduced bug - pressing tab in the middle of text when there is no selection inserted whitespace in the middle of the code.
This checks that the cursor is at the end of the text block before inserting spaces

Hmmm.... This doesn't really allow me to lodge diffs...

Additional changes to allow case:
press return, tab to indent, start code

Actually, this has a UI issue for when there is no selection. Will update.

I have added tab/shift tab functionality for the python-editor in scripter.
Not sure whether I am supposed to also open an issue in the bug tracker.
I am thinking of other ui improvements to the editor that may affect this file. Am not sure what I'm supposed to do to handle this conflict if it arises. What do I diff against?