Specify the Culture while searching the font

[Bykiev]

Specify the Culture while searching the font

[lahma]

What's the benefit? Isn't current culture the default and thus causes difference between in CI runs and end use experience?

[Bykiev]

@lahma, hello, nice to see you here!
The default culture is InvariantCulture, that's why if the system/thread culture is not English you'll be unable to find culture-specific fonts. See here for more information

[Bykiev]

Btw, there are a lot of tests in NPOI are failing in non-English environment because of not specified culture (e.g. comparing text 0.9 while in some cultures where comma is decimal separator it's 0,9)

[lahma]

Thanks for the clarification. I was just wondering about how the culture changes the matching, is it just string-matching the name and family? Getting the CI run properly on Linux seems to be a battle but it doesn't seem to be a problem on Windows so probably just Font presence issue - not about the name.

This might be a problem in a long run when people test them in "current culture" and then run on cloud environment like Azure where it's en-US IIRC.

[Bykiev]

Thanks for the clarification. I was just wondering about how the culture changes the matching, is it just string-matching the name and family?

Yes, you're right,you can find code here. I believe it's ok, many places in NPOI where CurrentCulture is used

[lahma]

Thanks for the clarification. I was just wondering about how the culture changes the matching, is it just string-matching the name and family?

Yes, you're right,you can find code here. I believe it's ok, many places in NPOI where CurrentCulture is used

Sorry for repeating myself here.. 🙂 As I can see it should be case-insensitive so there are some cases where it fails, I'm trying to understand what's the input and wrong output.

[tonyqus]

LGTM