$ ./Tools/Scripts/run-safari --debug
Starting SafariForWebKitDevelopment with DYLD_FRAMEWORK_PATH set to point to built WebKit in /Users/kbr/src/WebKit/WebKitBuild/Debug.
dyld: Symbol not found: _WKBundlePageCopyRenderLayerTree
Referenced from: /System/Library/PrivateFrameworks/Safari.framework/Versions/A/Safari
Expected in: /Users/kbr/src/WebKit/WebKitBuild/Debug/WebKit.framework/Versions/A/WebKit
Created attachment 405535 [details]
Comment on attachment 405535 [details]
This is fine:
- If these functions are linked but not called, you might want to put ASSERT_NOT_REACHED in them to help people understand they are intentionally left blank, or a comment, or both. I did that recently with some SPI that I had to leave around for the same reason.
- It’s also nice to leave out the argument names; if we complied with "unused argument" warnings, that would prevent them.
- Also, sometimes we move these things out of the SPI headers into an "old Safari depends on these" block, often not even in a header file, to help make sure no one accidentally adds new calls to the functions.
Not saying you need to do any of these things. Seems like you could land this just like it is.
Created attachment 405575 [details]
Patch for landing
(In reply to Darin Adler from comment #3)
> Comment on attachment 405535 [details]
Thanks for the review, and the suggestions.
> This is fine:
> - If these functions are linked but not called, you might want to put
> ASSERT_NOT_REACHED in them to help people understand they are intentionally
> left blank, or a comment, or both. I did that recently with some SPI that I
> had to leave around for the same reason.
I added comments to explain why the functions were intentionally left blank. Adding ASSERT_NOT_REACHED() caused the compiler to warn that the function is "noreturn".
Committed r265085: <https://trac.webkit.org/changeset/265085>
All reviewed patches have been landed. Closing bug and clearing flags on attachment 405575 [details].