-
Notifications
You must be signed in to change notification settings - Fork 45
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix SCRIPT
font family in Kiva
#975
Conversation
This should fail CI.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Basically LGTM
self.assertIn(query_family, font_family_aliases) | ||
self.assertIn(query_family, preferred_fonts) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There are items in font_family_aliases
which are not keys in preferred_fonts
. Perhaps you should only check against preferred_fonts
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We want to check it is in both, because of this line here:
enable/kiva/fonttools/_score.py
Line 38 in 4d51db2
if family1 in font_family_aliases: |
If it is not on font_family_aliases
then preferred_fonts
will never be accessed.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm saying that query_family in font_family_aliases
can be True
at the same time that query_family in preferred_fonts
is False
.
The more complete snippet from score_family
...
enable/kiva/fonttools/_score.py
Lines 38 to 40 in 4d51db2
if family1 in font_family_aliases: | |
if family1 in {"sans", "sans serif", "modern"}: | |
family1 = "sans-serif" |
... is not really being tested here because converting from the integer Kiva family constants via Font.familymap
will never result in the values "sans"
, "sans serif"
, or "modern"
.
Is the point of this test to keep Font.familymap
values in sync with font_family_aliases
and preferred_fonts
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The entries in familymap
need to be in both, yes (and that's what I care about for this PR).
font_family_aliases
(and presumably preferred_fonts
) can have extra values which come from the face_name
(this is what derailed #977). Eg. if someone names a font "sans" (ie. Font("sans", 12)
) it will also go through the above codepath. That likely needs some re-engineering at some point, but it is outside the scope of this PR)
* Use the "sphinx-copybutton" extension in documentation (#948) * DOC: Use the sphinx-copybutton extension in documentation modified: ci/edmtool.py modified: docs/source/conf.py modified: enable/__init__.py * Update ci/edmtool.py * Get things working on wxPython again (#950) Addresses these warnigns/exceptions: - TypeError: BitmapFromImage() got an unexpected keyword argument 'depth' - AttributeError: 'PaintDC' object has no attribute 'BeginDrawing' - wxPyDeprecationWarning: Call to deprecated item EmptyImage. Use :class:`Image` instead. image = wx.EmptyImage(*sz) - wxPyDeprecationWarning: Call to deprecated item BitmapFromImage. Use :class:`wx.Bitmap` instead bmp = wx.BitmapFromImage(image, depth=-1) See https://forums.wxwidgets.org/viewtopic.php?t=39930 why the Begin/EndDrawing calls can simply be removed. Related issues: #798 and #531 * Add SWIG to pyproject.toml build dependencies (#954) * Experiment with installing SWIG from PyPI * Update the EDM-based workflow * Remove restriction on SWIG version * Add Cython back as a development dependency; remove verify_swig_install * Remove generated C++ file. (#958) * Fix Kiva Quartz backend string encoding (#966) * Add regression test for Enable #964. * Properly decode strings in Kiva Quartz font code. * Properly guard against running Quartz tests on non-Mac systems. * Ensure Wx Quartz test terminates. * Make test success externally visible. * Fix tests, use UTF-8. * Add a github workflow that publishes to sdists to PyPI on release (#967) * Add a github workflow that publishes to sdists to PyPI on release. * Use build to build the sdist. * Take into account all font features when drawing SVG text. (#980) * Fix Quartz font rendering (#978) * Use font family in quartz when no font name given. * Render quartz text with fill color rather than stroke color. * Fix `SCRIPT` font family in Kiva (#975) * Add a test for #971 This should fail CI. * Fix bug and clean-up style. * Correct first point and orientation of Kiva QPainter arcs (#970) * Correct first point and orientation of Kiva QPainter arcs. This fixes #962 but not #960. * Update kiva/qpainter.py * Render font families better in QPainter backend (#973) * Fall back to kiva.fonttools if Qt can't come up with a font name. * Add a comment about setFamilies() * Fix WEIGHT_EXTRAHEAVY on Qt6 (#990) Some changes to tests to make sure eveything is tested correctly. * Fix the rendering of polygons in basecore2d-based backends (#987) * Fix the rendering of polygons in basecore2d-based backends This: - consistently uses Nx2 numpy arrays to store path points - doesn't create a new subpath after a LINES function A driveby fix gives a default implementation for show_text_at_point(). * Clean-up comments. * Fix some unused imports. * Some of the unused imports are in fact used. * Improve the situation of line_state_equal a bit. * Ensure all subpath arguments are numpy arrays. * Ensure all backends implement `arc_to` and `quad_curve_to` (#988) * Ensure all backends implement arc_to and quad_curve_to. * Fix 'infomration' typo in comments in multiple files. * Fixes suggested from PR review. Co-authored-by: Poruri Sai Rahul <rporuri@enthought.com> Co-authored-by: Brecht Machiels <brecht@mos6581.org> Co-authored-by: Mark Dickinson <mdickinson@enthought.com>
Fix is a matter of having
Font.familymap
map to the right thing. This is a very old bug.This is the benchmark rendering in the agg backend after the fix (note the "script" column):
Fixes #971.