There are condition check for characterFontData followed by same condition check for characterFontData in the FontGlyphs::glyphDataAndPageForCharacter(). I think this is not necessary.
Created attachment 235507 [details] Patch
Comment on attachment 235507 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=235507&action=review > Source/WebCore/platform/graphics/FontGlyphs.cpp:389 > characterFontData = characterFontData->variantFontData(description, variant); This function might return nullptr.
I'm not a reviewer but I would r- this.
Yes, I agree that variantFontData() could rarely return nullptr. So I close this issue.
Thank you for reviewing.
Comment on attachment 235507 [details] Patch As Myles says, this patch is only correct if variantFontData never returns null. Did you prove that variantFontData will never return null?