[widget audit] toga.Button #1761

freakboy3742 · 2023-01-31T06:13:45Z

A testing and documentation audit of the toga.Button widget.

Includes the content from #1794.

Audit checklist

freakboy3742 · 2023-01-31T06:23:45Z

As of cddb9e4 platform coverage is at:

core 100%
cocoa 100%
winforms 33.3% (?)
iOS 86.7%
Android 98.0%
Linux 85.7%

Most of the coverage omissions appear to be related to the button press. I suspect coverage on Windows is a misreport; it's not reporting coverage on any of the method bodies, even though the tests should be running. This might be a path issue in my virtualised Windows setup.

mhsmith · 2023-01-31T16:18:28Z

The Windows issue was caused by the same thing as nedbat/coveragepy#686: the main loop runs on a thread that's created through the Windows API, not the Python threading API, so it wasn't detected by coverage.

Fixed by manually installing the trace hook at the start of the thread.

freakboy3742 · 2023-02-07T06:59:00Z

testbed/tests/data.py

+        (50, 128, 200, 0.0),
+        (50, 128, 200, 0.5),
+        (50, 128, 200, 0.9),
+        (50, 128, 200, 1.0),


I've reduced the multiplicity of colours here as it didn't seem to be adding anything to test rigor, but it was difficult to see the "meaningful" colors appear in the test series.

freakboy3742 · 2023-02-07T07:00:48Z

As of 065bca3, the GTK colour tests pass, but only if you have UI animations turned of. At the command line, run:

gsettings set org.gnome.desktop.interface enable-animations false

You can reset the setting with:

gsettings reset org.gnome.desktop.interface enable-animations false

This is an OS-wide setting.

mhsmith · 2023-02-19T20:28:03Z

On Debian Buster, after disabling animations, I was able to reproduce some similar failures to the CI, plus a new one involving fonts and UTF-8:

Log

[testbed] Running test suite in dev environment...
===========================================================================
WARNING: Can't find icon resources/testbed; falling back to default icon
/home/smith/git/beeware/toga/gtk/src/toga_gtk/app.py:43: DeprecationWarning: Gtk.Window.set_wmclass is deprecated
  self.native.set_wmclass(app.interface.name, app.interface.name)

(testbed.py:15361): Pango-CRITICAL **: 20:21:45.025: pango_font_description_set_size: assertion 'size >= 0' failed
============================= test session starts ==============================
collecting ... collected 25 items

tests/widgets/test_button.py::test_background_color <- tests/widgets/properties.py FAILED [  4%]
tests/widgets/test_button.py::test_background_color_reset <- tests/widgets/properties.py FAILED [  8%]
tests/widgets/test_button.py::test_color <- tests/widgets/properties.py FAILED [ 12%]
tests/widgets/test_button.py::test_color_reset <- tests/widgets/properties.py FAILED [ 16%]
tests/widgets/test_button.py::test_font <- tests/widgets/properties.py FAILED [ 20%]
tests/widgets/test_button.py::test_text <- tests/widgets/properties.py PASSED [ 24%]
tests/widgets/test_button.py::test_text_width_change <- tests/widgets/properties.py PASSED [ 28%]
tests/widgets/test_button.py::test_press PASSED                          [ 32%]
tests/widgets/test_button.py::test_background_color_transparent PASSED   [ 36%]
tests/widgets/test_button.py::test_button_size PASSED                    [ 40%]
tests/widgets/test_label.py::test_background_color <- tests/widgets/properties.py FAILED [ 44%]
tests/widgets/test_label.py::test_background_color_reset <- tests/widgets/properties.py FAILED [ 48%]
tests/widgets/test_label.py::test_background_color_transparent <- tests/widgets/properties.py PASSED [ 52%]
tests/widgets/test_label.py::test_color <- tests/widgets/properties.py FAILED [ 56%]
tests/widgets/test_label.py::test_color_reset <- tests/widgets/properties.py FAILED [ 60%]
tests/widgets/test_label.py::test_font <- tests/widgets/properties.py FAILED [ 64%]
tests/widgets/test_label.py::test_text <- tests/widgets/properties.py PASSED [ 68%]
tests/widgets/test_label.py::test_text_width_change <- tests/widgets/properties.py SKIPPED (resizes not applying correctly) [ 72%]
tests/widgets/test_label.py::test_multiline SKIPPED (changing text does not trigger a refresh (#1289)) [ 76%]
tests/widgets/test_label.py::test_alignment SKIPPED (alignment probe not implemented) [ 80%]
tests/widgets/test_slider.py::test_init PASSED                           [ 84%]
tests/widgets/test_slider.py::test_value SKIPPED (on_change called 2 times) [ 88%]
tests/widgets/test_slider.py::test_change SKIPPED (min/max value not correct) [ 92%]
tests/widgets/test_slider.py::test_min PASSED                            [ 96%]
tests/widgets/test_slider.py::test_max PASSED                            [100%]

=================================== FAILURES ===================================
____________________________ test_background_color _____________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 102, in test_background_color
    assert_color(probe.background_color, color)
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AttributeError: 'NoneType' object has no attribute 'r'
_________________________ test_background_color_reset __________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 113, in test_background_color_reset
    assert_color(probe.background_color, named_color(RED))
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AttributeError: 'NoneType' object has no attribute 'r'
__________________________________ test_color __________________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 74, in test_color
    assert_color(probe.color, color)
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AssertionError: assert 54 == 0
 +  where 54 = getattr(rgba(0, 0, 54, 1.0), 'b')
 +  and   0 = getattr(rgba(0, 0, 0, 1.0), 'b')
_______________________________ test_color_reset _______________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 89, in test_color_reset
    assert_color(probe.color, named_color(RED))
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AssertionError: assert 0 == 255
 +  where 0 = getattr(rgba(0, 0, 54, 1.0), 'r')
 +  and   255 = getattr(rgb(255, 0, 0), 'r')
__________________________________ test_font ___________________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 47, in test_font
    new_size_font = probe.font
  File "/home/smith/git/beeware/toga/gtk/tests_backend/widgets/base.py", line 78, in font
    return toga_font(sc.get_property("font", sc.get_state()))
  File "/home/smith/git/beeware/toga/gtk/tests_backend/widgets/properties.py", line 28, in toga_font
    family=font.get_family(),
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xa0 in position 0: invalid start byte
____________________________ test_background_color _____________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 102, in test_background_color
    assert_color(probe.background_color, color)
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AssertionError: assert 0 == 128
 +  where 0 = getattr(rgba(0, 0, 128, 1.0), 'r')
 +  and   128 = getattr(rgba(128, 128, 128, 1.0), 'r')
_________________________ test_background_color_reset __________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 113, in test_background_color_reset
    assert_color(probe.background_color, named_color(RED))
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AssertionError: assert 0 == 255
 +  where 0 = getattr(rgba(0, 0, 0, 1.0), 'r')
 +  and   255 = getattr(rgb(255, 0, 0), 'r')
__________________________________ test_color __________________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 74, in test_color
    assert_color(probe.color, color)
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AssertionError: assert 0 == 128
 +  where 0 = getattr(rgba(0, 0, 128, 1.0), 'r')
 +  and   128 = getattr(rgba(128, 128, 128, 1.0), 'r')
_______________________________ test_color_reset _______________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 89, in test_color_reset
    assert_color(probe.color, named_color(RED))
  File "/home/smith/git/beeware/toga/testbed/tests/assertions.py", line 26, in assert_color
    assert getattr(actual, component) == getattr(expected, component)
AssertionError: assert 0 == 255
 +  where 0 = getattr(rgba(0, 0, 0, 1.0), 'r')
 +  and   255 = getattr(rgb(255, 0, 0), 'r')
__________________________________ test_font ___________________________________
Traceback (most recent call last):
  File "/home/smith/git/beeware/toga/testbed/tests/widgets/properties.py", line 40, in test_font
    orig_font = probe.font
  File "/home/smith/git/beeware/toga/gtk/tests_backend/widgets/base.py", line 78, in font
    return toga_font(sc.get_property("font", sc.get_state()))
  File "/home/smith/git/beeware/toga/gtk/tests_backend/widgets/properties.py", line 28, in toga_font
    family=font.get_family(),
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x90 in position 0: invalid start byte
==================================== PASSES ====================================
__________________________________ test_init ___________________________________
---------------------------- Captured stdout setup -----------------------------
[GTK+] Not implemented: Slider.tick_count()
[GTK+] Not implemented: Slider.set_on_press()
[GTK+] Not implemented: Slider.set_on_release()
___________________________________ test_min ___________________________________
---------------------------- Captured stdout setup -----------------------------
[GTK+] Not implemented: Slider.tick_count()
[GTK+] Not implemented: Slider.set_on_press()
[GTK+] Not implemented: Slider.set_on_release()
___________________________________ test_max ___________________________________
---------------------------- Captured stdout setup -----------------------------
[GTK+] Not implemented: Slider.tick_count()
[GTK+] Not implemented: Slider.set_on_press()
[GTK+] Not implemented: Slider.set_on_release()
=================== 10 failed, 10 passed, 5 skipped in 3.38s ===================

freakboy3742 · 2023-02-19T22:35:08Z

Most of those errors look like the Pygobject issue - The font bug, and the failures where there's a B and A channel value, but no R and G value (e.g., test_color). Are you sure you're running with the bugfix fork of pygobject?

However, some of the other failures do give me an idea - I wonder if the issue is color resets, and the tests not correctly identifying the background color (which is essentially platform dependent, but will be reliably reproducible on any given machine, and appears to be rgba(0,0,0,0 on my machine).

freakboy3742 · 2023-02-20T06:41:38Z

Nope. It was Wayland. Of course it was.

freakboy3742 · 2023-02-20T06:53:00Z

The good news is that the global gsettings change is no longer needed; I've found a programmatic GTK setting that both X and Wayland honor.

mhsmith · 2023-02-20T12:34:48Z

OK, after pulling the current version of this PR, and running briefcase dev -r to change the pygobject version, everything now passes on my machine.

mhsmith · 2023-03-14T16:33:00Z

test_implementation doesn't allow widgets to implement methods by inheriting them from a base class. This is often the most readable and maintainable way of doing it, but will inevitably be different from platform to platform.

Also, it's fairly arbitrary which methods it's checking for: e.g. TextInput has a set_font method, but Label doesn't.

Now that we have the testbed, I think test_implementation is more trouble than it's worth, and can just be removed.

mhsmith · 2023-03-14T19:17:41Z

I've added a font_size example app for manual testing. Not all the widgets implement font size yet (that can wait until future PRs), but those that do are mostly consistent across the 5 platforms, and proportionate to the default font size.

The exception is Windows, where the default font size is 8 points. However, I think it's actually bigger than 8/72 inches on most screens, because it was fixed at a time when Windows assumed 96 dpi but the typical monitor was less dense than that (see this article from 1999). And since Windows is so serious about backward compatibility, we're still stuck with it. [EDIT 2023-09: Disregard that, I've now measured 8 point text on Windows 10 at 100% scale, and it's 11 pixels tall from top of ascender to bottom of descender, which is exactly what you would expect from 8 * 96 / 72 == 10.66.]

So Toga widgets with fixed font sizes currently appear proportionately larger on Windows than on the other platforms. I don't think there's anything we can do about that now, but in the future we could deal with it by implementing the CSS font size keywords small, medium, large, etc. Created #1814 to track this idea.

freakboy3742

I've done a full review pass; this mostly looks good, but I've flagged a couple of minor issues. Most of them are likely fixed with a docstring explaining reasoning; the only "big" item is the dependency inversion on font family resolution in the test probe (and even that is that big).

freakboy3742 · 2023-03-20T01:10:48Z

android/src/toga_android/fonts.py

+            # will give relative sizes that are consistent with desktop platforms. It also
+            # means font sizes will all change proportionately if the user adjusts the
+            # text size in the system settings.
+            tv.setTextSize(TypedValue.COMPLEX_UNIT_SP, self.interface.size)


This is definitely a good approximation; I only wonder if it might be worth introducing a fudge factor so that "12pt == 14sp" (i.e., "size in toga points" * 1.1666 == size in dp).

(for the record - this is on the "strong opinion, weakly held" list; I won't give much pushback on a different fudge factor, or no fudge factor at all).

Windows defaults to 8 points (see above), and according to the fonts.py files in the respective backends, Cocoa defaults to 12 points and iOS 17. So I don't see any strong reason to fix 12 as the default. Android is already within the range of the other platforms, and making 1 Toga pt = 1sp brings it closer to iOS, which is probably the one it's most useful to be consistent with.

In the long run I think the best solution to this issue is the CSS font size keywords, as mentioned above.

Sure - that makes sense.

android/src/toga_android/fonts.py

freakboy3742 · 2023-03-20T01:12:57Z

android/src/toga_android/fonts.py


-    def get_typeface(self):
+        cache_key = (self.interface, default_typeface)


Why is the default typeface part of the cache key?

Because it may affect the resulting font when font_family is SYSTEM. The default font on most widgets is sans-serif, but on buttons it's sans-serif-medium, which is intermediate between normal and bold.

freakboy3742 · 2023-03-20T01:27:50Z

cocoa/src/toga_cocoa/fonts.py

                )

-                if font is None:


Are you certain this error handling is no longer required? Core and system fonts will have all the weights/variants; but if a user-supplied font doesn't have a particular variant/weight, will this still succeed?

(And if it will - a comment explaining how/why would be worthwhile)

It's no longer possible for font to be None here, because:

When a custom font can't be found, I've added a line above to fall back on the system font. Previously in this case, the size, style and weight settings would all be ignored.

According to the convertFont documentation if it fails to find a font with the given traits, it returns the original font unchanged. I'll add a comment explaining this.

Also, note I've changed the indentation here so that the SYSTEM font uses the same bold/italic logic as everything else. Previously the italic setting would be ignored with a SYSTEM font.

if a user-supplied font doesn't have a particular variant/weight

It would be good to have a test for this, but that would require actually supplying such a font, and it looks like loading fonts from a file is currently broken on Cocoa. But this won't be a very common thing for an app to do, so I'm happy to defer it until we come to test the Font API directly.

freakboy3742 · 2023-03-20T01:34:15Z

testbed/tests/assertions.py

+        ".AppleSystemUIFont",
+        "sans-serif-medium",  # Android buttons
+    ],
+}


What's the reasoning behind having this table of "generic" fonts in the testbed assertions, rather than being something that is normalised out by the test backend? Previously, toga_font on each backend would normalise ".AppleSystemUIFont" into SYSTEM, but do so in the Cocoa backend.

Having this table here seems like an odd inversion - we've got platform-specific details in the generic test definitions.

The problem I was trying to solve was the many-to-one mapping between Toga font names and actual fonts. For example, SANS_SERIF and SYSTEM will often give you the same visible result, as will explicitly specifying the font name. But whether these 3 options can be distinguished by the probe varies from platform to platform.

But you're right, it would be better for platform-specific differences to be handled by the test backend, so I'll do that.

freakboy3742 · 2023-03-20T01:35:02Z

cocoa/tests_backend/widgets/properties.py

-            ".AppleSystemUIFont": SYSTEM,
-            "Papyrus": FANTASY,
-        }.get(str(font.familyName), str(font.familyName)),
+        family=str(font.familyName),


See Q below about this inversion, removing the ".AppleSystemUIFont" normalisation from the platform backend.

freakboy3742 · 2023-03-20T01:47:04Z

iOS/src/toga_iOS/fonts.py

-                if font is None:
-                    print(f"Unable to load font: {self.interface.size}pt {full_name}")
-                else:
-                    _FONT_CACHE[self.interface] = font


As with Cocoa; we either need to report the missing font issue, or docstring the reason why it isn't an issue.

freakboy3742 · 2023-03-20T01:51:59Z

testbed/tests/widgets/properties.py

+                await probe.redraw()
+                assert_font_family(probe.font.family, family)
+                assert probe.font.weight == weight
+                assert probe.font.style == style


This makes me wonder whether what is needed here is a set of comprehensive "font management" tests operating on a single widget (say, Label), and a basic "can a font be set" test on each widget. If basic font setting works on a widget, we don't have any particular reason to believe that weight/variants will be problematic.

That's true if we assume that the widgets are all using a common font implementation, and using it correctly, but that's something that could easily be broken by accident in the future. So I think it's worth being comprehensive as long as it only adds a few seconds to the tests.

mhsmith · 2023-03-20T13:15:42Z

testbed/tests/widgets/test_button.py

+    assert 10 <= probe.width <= 150, f"Width ({probe.width}) not in range (10, 150)"
+    assert 10 <= probe.height <= 50, f"Height ({probe.height}) not in range (10, 50)"


As discussed, these messages should be generated automatically by pytest's "assertion rewriting", but that isn't currently working on Android. But since this PR is blocking all the other widget testing, let's fix that separately.

…, iOS)

core/tests/widgets/test_activityindicator.py

docs/reference/api/widgets/label.rst

core/tests/widgets/test_progressbar.py

…play a single space to prevent the widget from collapsing

mhsmith

From the user's point of view, banning empty text on buttons seems like an arbitrary restriction. I see the problem was that it caused the button height to collapse on Winforms, so I've worked around that by making it display a single space instead, which is visually identical, and only requires a small special case in the tests.

I saw a similar issue with empty labels on iOS last week, so I'll probably end up doing something similar when I review the Label PR tomorrow. However, in the case of a label with a background color, the difference between a space and an empty string may be visually significant, so maybe there's a Unicode zero-width character that would be better.

mhsmith · 2023-03-20T20:31:27Z

docs/reference/api/widgets/button.rst

-Notes
-----
-* Any "truthy" object (i.e, object that evaluates as ``True``) can be provided
-  as the text for the button; if the object is not a string, it will be
-  automatically converted to a string using ``str()``.
-
-* The button text cannot contain newlines. Any content after the first newline
-  will be ignored.


I've moved the property documentation into the docstring so it'll appear next to the property itself, rather than in a separate section.

mhsmith · 2023-03-20T20:32:50Z

testbed/tests/widgets/properties.py

-async def test_text_empty(widget, probe):
-    "The text displayed on a widget can be empty"
-    # Set the text to the empty string
-    widget.text = ""


Empty strings are now covered by test_text above, and conversion of non-string objects is covered by test_button.py in the core.

freakboy3742

A couple more notes and tweaks, hopefully not too controversial.

freakboy3742 · 2023-03-20T23:52:08Z

android/tests_backend/widgets/base.py

+        if expected == SYSTEM:
+            assert actual == (
+                "sans-serif-medium"
+                if isinstance(self.widget, toga.Button)


We can use subclassing here, rather than an isinstance check.

freakboy3742 · 2023-03-21T00:38:57Z

core/src/toga/widgets/button.py


-        self._impl.set_text(value)
+        self._impl.set_text(self._text)


I'm generally against the introduction of interface state/shadow variables; in this case, the shadow variable is only required if we have problems round-tripping the original value. In this implementation we're not round tripping anyway (because we're losing None and multiline input); to support the winforms edge case, I'm entirely comfortable with a normalizing a user-supplied "\u0200" (zero width space) to ""). It's highly unlikely to ever be user supplied in that form, and it's functionally indistinguishable from "".

I'm entirely comfortable with a normalizing a user-supplied "\u0200" (zero width space) to "").

Makes sense, but \u0200 is LATIN CAPITAL LETTER A WITH DOUBLE GRAVE. I've replaced that with \u200B.

freakboy3742 · 2023-03-21T00:52:25Z

docs/reference/api/widgets/button.rst

@@ -29,14 +29,8 @@ A button has a text label. A handler can be associated with button press events.

    button = toga.Button('Click me', on_press=my_callback)

-Notes


I think we should keep this as a "Notes" section. While the notes at present are all style related, we're very likely going to come across platform specific eccentricities that don't fit the description being "style", and don't map to a specific attribute like the text handling of None et al.

…ound in testbed

freakboy3742 marked this pull request as draft January 31, 2023 06:25

mhsmith mentioned this pull request Jan 31, 2023

Distinguish beta from stable support in docs #1762

Merged

4 tasks

freakboy3742 commented Feb 7, 2023

View reviewed changes

freakboy3742 force-pushed the audit-button branch 4 times, most recently from 227e187 to 4a1ce8a Compare February 16, 2023 05:24

freakboy3742 force-pushed the audit-button branch from df8737e to f7a1006 Compare February 21, 2023 05:25

freakboy3742 and others added 14 commits February 25, 2023 11:27

Whitespace change to use as the starting point for a PR.

92fd082

Add changenote.

d206715

Fix tracing on Windows

9fdf46e

Migrate dummy widget tools to pytest.

38bd4d6

Remove use of shadow attribute for button label.

b541c09

Convert Button tests to pytest.

7077562

Fix Android tests, get to 100% coverage.

ec04ddc

Fix get_text implementation on gtk and winforms.

859ad48

100% coverage on iOS.

f93158e

Remove deprecated code from Button.

67f3374

Winforms to 100%.

ca8a760

GTK to 100%.

8a03783

Winforms to 100%, no skips.

37985ae

Add test selection from the command line, and a slow mode.

2b9714e

Merge branch 'system-ci' into audit-button

8bd9528

mhsmith added 2 commits March 14, 2023 16:37

Allow default font to be returned as SANS_SERIF as well as SYSTEM

e83f2fa

Add examples/font_size

81e2d29

mhsmith mentioned this pull request Mar 15, 2023

Support CSS font size keywords #1814

Open

mhsmith and others added 4 commits March 16, 2023 20:09

Add bold and italic font tests (working on Android, Cocoa and iOS)

02a345d

Add bold and italic font tests (working on GTK and Winforms)

9248973

Remove redundant set_font methods in dummy widgets

45e52e6

Merge branch 'main' into audit-button

8445dae

freakboy3742 commented Mar 20, 2023

View reviewed changes

mhsmith reviewed Mar 20, 2023

View reviewed changes

mhsmith added 4 commits March 20, 2023 13:47

Improve comments

c646ef0

Move generic font mapping assertions to tests_backend (Android, Cocoa…

e2da16b

…, iOS)

Move generic font mapping assertions to tests_backend (GTK)

926c3d6

Move generic font mapping assertions to tests_backend (Winforms)

7f96efa

mhsmith reviewed Mar 20, 2023

View reviewed changes

core/tests/widgets/test_activityindicator.py Outdated Show resolved Hide resolved

docs/reference/api/widgets/label.rst Outdated Show resolved Hide resolved

core/tests/widgets/test_progressbar.py Outdated Show resolved Hide resolved

mhsmith added 2 commits March 20, 2023 19:09

Fix typos

36ed554

Allow Button text to be set to an empty string, and make Winforms dis…

091e3ec

…play a single space to prevent the widget from collapsing

mhsmith approved these changes Mar 20, 2023

View reviewed changes

freakboy3742 added 4 commits March 21, 2023 07:58

Use subclassing rather than an isinstance check.

dd73a9b

Revert the use of an interface state variable for button text.

cfdc3e5

Mark some classes with Test in the name as not-tests.

ce54201

Tweak section heading.

5abe2f1

freakboy3742 commented Mar 21, 2023

View reviewed changes

freakboy3742 and others added 2 commits March 21, 2023 09:01

Modify the winforms button probe to do ZWS normalization.

f76c90a

Use correct zero width space character, and remove unnecessary workar…

8f55716

…ound in testbed

mhsmith merged commit fa8ae75 into beeware:main Mar 21, 2023

freakboy3742 deleted the audit-button branch March 21, 2023 22:35

freakboy3742 mentioned this pull request Apr 15, 2023

Button / select box - font size attribute #1037

Closed


		def get_typeface(self):
		cache_key = (self.interface, default_typeface)

		assert 10 <= probe.width <= 150, f"Width ({probe.width}) not in range (10, 150)"
		assert 10 <= probe.height <= 50, f"Height ({probe.height}) not in range (10, 50)"

		@@ -29,14 +29,8 @@ A button has a text label. A handler can be associated with button press events.

		button = toga.Button('Click me', on_press=my_callback)

		Notes

[widget audit] toga.Button #1761

[widget audit] toga.Button #1761

Conversation

freakboy3742 commented Jan 31, 2023 • edited Loading

Audit checklist

freakboy3742 commented Jan 31, 2023

mhsmith commented Jan 31, 2023

Choose a reason for hiding this comment

freakboy3742 commented Feb 7, 2023

mhsmith commented Feb 19, 2023 • edited Loading

freakboy3742 commented Feb 19, 2023

freakboy3742 commented Feb 20, 2023

freakboy3742 commented Feb 20, 2023

mhsmith commented Feb 20, 2023

mhsmith commented Mar 14, 2023

mhsmith commented Mar 14, 2023 • edited Loading

freakboy3742 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mhsmith Mar 20, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mhsmith left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

freakboy3742 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

freakboy3742 commented Jan 31, 2023 •

edited

Loading

mhsmith commented Feb 19, 2023 •

edited

Loading

mhsmith commented Mar 14, 2023 •

edited

Loading

mhsmith Mar 20, 2023 •

edited

Loading

mhsmith left a comment •

edited

Loading