[Webkit-unassigned] [Bug 193036] [FreeType] Restore conditional compilation logic for recent HarfBuzz refactoring

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=193036

Michael Catanzaro <mcatanzaro at igalia.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #358077|review?                     |review+
              Flags|                            |

--- Comment #2 from Michael Catanzaro <mcatanzaro at igalia.com> ---
Comment on attachment 358077
  --> https://bugs.webkit.org/attachment.cgi?id=358077
Patch

r=me to get back to the previous status quo, but it looks problematic. Jim, what port are you building, and with what build flags?

Hey Don, here's another one. ENABLE(OPENTYPE_MATH) is not a valid feature flag, but it's being used throughout several files under platform/graphics/opentype. It's defined only in OptionsGTK.cmake using cpp flags:

add_definitions(-DENABLE_OPENTYPE_MATH=1)

Then looking at USE(HARFBUZZ), my initial reaction is confusion. Surely USE(FREETYPE) must imply USE(HARFBUZZ), but it looks like USE(HARFBUZZ) is only set in OptionsPlayStation.cmake, which is nuts because GTK and WPE both use harfbuzz too, so it must be set somewhere else, but I can't find it. It's used in FontCascade.cpp, HbUniquePtr.h, and OpenTypeMathData.[cpp,h] so that's surely not enough places to actually build if harfbuzz were actually not present.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20181226/614c4d88/attachment.html>