Bug 182274 – [CMake] Make WTF headers copies

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Don Olmstead]

WTF headers should be copied during the build to match the behavior of Apple ports.

[Don Olmstead]

Created attachment 332620 [details]
Patch

Trying the bots

[EWS Watchlist]

Attachment 332620 [details] did not pass style-queue:


ERROR: Source/WTF/wtf/CMakeLists.txt:236:  Alphabetical sorting problem. "UniStdExtras.h" should be before "UniqueRef.h".  [list/order] [5]
Total errors found: 1 in 9 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Don Olmstead]

Created attachment 332623 [details]
Patch

Hopefully make GTK and WPE happy

[Don Olmstead]

Comment on attachment 332623 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=332623&action=review

So this adds all header files explicitly to the forwarding headers and copies all of them across all CMake ports.

> Source/WTF/wtf/CMakeLists.txt:246
> +    WindowsExtras.h

All files within the root of Source/WTF/wtf are added into this WTF_HEADERS value.

There is another bug, https://bugs.webkit.org/show_bug.cgi?id=182275,  where some of these should really be moved into platform specific directories and included explicitly.

> Source/WebDriver/CMakeLists.txt:6
> +    "${FORWARDING_HEADERS_DIR}"

In here and in gtest the FORWARDING_HEADERS_DIR is added explicitly.

I'm not happy about this but currently WTF has both #include <> and #include "" within its headers. Adding the FORWARDING_HEADERS_DIR within the WTF_INCLUDE_DIRECTORIES would possibly result in some ambiguous includes. So I'm trying to avoid it. Not sure if there's another way in CMake to add a dependent include path that only affects dependent projects.

The real fix is to just flatten WTF but this is really a first step towards that.

[EWS Watchlist]

Comment on attachment 332623 [details]
Patch

Attachment 332623 [details] did not pass ios-sim-ews (ios-simulator-wk2):
Output: http://webkit-queues.webkit.org/results/6256939

New failing tests:
imported/w3c/web-platform-tests/service-workers/service-worker/navigation-redirect.https.html

[EWS Watchlist]

Created attachment 332628 [details]
Archive of layout-test-results from ews122 for ios-simulator-wk2

The attached test failures were seen while running run-webkit-tests on the ios-sim-ews.
Bot: ews122  Port: ios-simulator-wk2  Platform: Mac OS X 10.12.6

[Don Olmstead]

Comment on attachment 332623 [details]
Patch

Clearing flags on attachment: 332623

Committed r227845: <https://trac.webkit.org/changeset/227845>

[Don Olmstead]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/37037670>

[Csaba Osztrogonác]

(In reply to Don Olmstead from comment #7)
> Comment on attachment 332623 [details]
> Patch
> 
> Clearing flags on attachment: 332623
> 
> Committed r227845: <https://trac.webkit.org/changeset/227845>

FYI, it broke JSCOnly builds. See build.webkit.org for details.

[Don Olmstead]

Committed r227861: <https://trac.webkit.org/changeset/227861>

[Konstantin Tokarev]

Comment on attachment 332623 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=332623&action=review

> Source/WTF/wtf/CMakeLists.txt:-363
> -endif ()

This breaks build of non-Mac ports on macOS

[Konstantin Tokarev]

Committed r228007: <https://trac.webkit.org/changeset/228007>