-
-
Notifications
You must be signed in to change notification settings - Fork 780
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
core: Extend the embedded fake device font with more characters #6224
core: Extend the embedded fake device font with more characters #6224
Conversation
By also embedding the "Basic Latin", "Latin I", and "Latin Extended A" sections.
Hmm, I suppose the failing |
Sounds like a very good tradeoff, compared to hundreds of kB we add for some dependencies that we use 1% of ;)
Maybe run the test SWFs and compare the screenshots? |
TBH, to me it looks better in some places and worse in others... not sure what to think about it :D |
I don't have a good eye for these things, and I find both versions to be readable (legible?) just fine. Isn't the purpose of approx-tests to handle the uncertainties that are outside of our control? Say, differences across platforms, backends, or slight changes in external dependencies. |
Well, as we can see, a font in a way is outside of our control :) If some other person updates the embedded font and gets slightly different results again, they'll need to update the test again - which kinda defeats the point :) The previous font also wasn't 100% guaranteed to match FP, I think. (Further, even some of existing tests are are approx because "I implemented this to the best of my ability, but there are still some minor differences from FP - I probably missed something, but let's consider that out of scope for now". You can see some in the comments there, like (though I wish every single one had an explanation comment... if you do decide to make it an approx test, maybe let's add a top-level comment to always document the reason for being approx) |
But then again, if you don't want to do it, just updating the test now and waiting for the next time this happens is a valid approach too :D |
Truth be told, I'm leaning towards this at the moment. :D |
By also embedding the "Basic Latin", "Latin I", and "Latin Extended A" sections.
This is all I did to the
![image](https://user-images.githubusercontent.com/288816/152695490-252d3dc1-92a3-4253-a8a7-9945a0e546f6.png)
![image](https://user-images.githubusercontent.com/288816/152695499-5fa39182-0396-49de-9004-414f409dbce8.png)
.fla
file before exporting the.swf
and extracting theDefineFont3
data:While this is still just putting out a larger fire temporarily (cc: #1862), and also increases the embedded (uncompressed) data size by about 60 kB, I still think it's worth it.
After a quick glance at the "Latin Extended B" and "Latin Extended Add'l" sections, they seemed less crucial to include to me personally, relative to their size.
If I didn't miss anything, this fixes #1099, #5357, and probably many more related issues that will have to be checked one by one.