Currently there is a compilation path for a generic implementation of MemoryPressureHandler and MemoryFootprint but they're in multiple places. They should be split into MemoryPressureHandlerGeneric and MemoryFootprintGeneric.
Created attachment 347858 [details] Patch
Created attachment 347861 [details] Patch
Created attachment 347875 [details] Patch
Comment on attachment 347875 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=347875&action=review > Source/WTF/wtf/PlatformJSCOnly.cmake:59 > +elseif (CMAKE_SYSTEM_NAME MATCHES ".*Linux") You don't need '.*'. https://github.com/WebKit/webkit/blob/master/Source/cmake/OptionsGTK.cmake#L113 > if (CMAKE_SYSTEM_NAME MATCHES "Linux") BTW, the following code just use STREQUAL. https://github.com/WebKit/webkit/blob/master/Source/cmake/OptionsCommon.cmake#L22 > ${CMAKE_SYSTEM_NAME} STREQUAL "Linux" How does the variable CMAKE_SYSTEM_NAME look like? Can it be "FooBar Linux"?
(In reply to Fujii Hironori from comment #4) > Comment on attachment 347875 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=347875&action=review > > > Source/WTF/wtf/PlatformJSCOnly.cmake:59 > > +elseif (CMAKE_SYSTEM_NAME MATCHES ".*Linux") > > You don't need '.*'. > > https://github.com/WebKit/webkit/blob/master/Source/cmake/OptionsGTK. > cmake#L113 > > > if (CMAKE_SYSTEM_NAME MATCHES "Linux") > > BTW, the following code just use STREQUAL. > > https://github.com/WebKit/webkit/blob/master/Source/cmake/OptionsCommon. > cmake#L22 > > > ${CMAKE_SYSTEM_NAME} STREQUAL "Linux" > > How does the variable CMAKE_SYSTEM_NAME look like? Can it be "FooBar Linux"? From https://github.com/WebKit/webkit/blob/master/Source/ThirdParty/openvr/CMakeLists.txt#L23 > if(CMAKE_SYSTEM_NAME MATCHES ".*Linux") https://cmake.org/cmake/help/latest/variable/CMAKE_SYSTEM_NAME.html doesn't explicitly state that it is just Linux. and it sounds like you could technically set it when cross compiling to whatever you like. Since its JSCOnly its likely that there are out of trunk ports that might set something weird, but I don't have a strong opinion here.
Created attachment 347893 [details] Patch Updated based on review comments
Comment on attachment 347893 [details] Patch Clearing flags on attachment: 347893 Committed r235215: <https://trac.webkit.org/changeset/235215>
All reviewed patches have been landed. Closing bug.
<rdar://problem/43630726>
A future change where there would be a generic bmalloc implementation snuck in the last patch. It needs to be removed.
Created attachment 347938 [details] Build fix
Comment on attachment 347938 [details] Build fix Clearing flags on attachment: 347938 Committed r235235: <https://trac.webkit.org/changeset/235235>