Turn nsIDocument::mCharacterSet into mozilla::NotNull<const mozilla::Encoding*> and change GetDocumentCharacterSet() and SetDocumentCharacterSet() accordingly.

Comment on attachment 8881094 [details] Bug 1373984 - Turn nsIDocument::mCharacterSet into mozilla::NotNull<const mozilla::Encoding*>. https://reviewboard.mozilla.org/r/152408/#review157480 ::: dom/base/nsGlobalWindow.cpp:13401 (Diff revision 1) > // > // Note the algorithm to get the base URI should match the one > // used to actually kick off the load in nsWindowWatcher.cpp. > nsCOMPtr<nsIDocument> doc = sourceWindow->GetDoc(); > nsIURI* baseURI = nullptr; > nsAutoCString charset(NS_LITERAL_CSTRING("UTF-8")); // default to utf-8 No need to use a string here anymore, AFAICT. ::: dom/html/nsHTMLDocument.cpp:311 (Diff revision 1) > > - if(requestCharsetSource <= aCharsetSource) > + if (requestCharsetSource <= aCharsetSource) > return; > > - if(NS_SUCCEEDED(rv) && IsAsciiCompatible(requestCharset)) { > + if (NS_SUCCEEDED(rv)) { > + auto encoding = Encoding::ForLabelNoReplacement(requestCharset); Why not `ForName`? ::: dom/html/nsHTMLDocument.cpp:349 (Diff revision 1) > } > > if(NS_SUCCEEDED(rv) && > !forceCharsetFromDocShell.IsEmpty() && > IsAsciiCompatible(forceCharsetFromDocShell)) { > - aCharset = forceCharsetFromDocShell; > + auto encoding = Encoding::ForLabelNoReplacement(forceCharsetFromDocShell); Hasn't this already gone through label resolution on the doc shell level? ::: editor/libeditor/EditorBase.cpp:1195 (Diff revision 1) > { > nsCOMPtr<nsIDocument> document = GetDocument(); > if (NS_WARN_IF(!document)) { > return NS_ERROR_UNEXPECTED; > } > - document->SetDocumentCharacterSet(characterSet); > + auto encoding = Encoding::ForLabelNoReplacement(characterSet); Why not `ForName`? ::: parser/html/nsHtml5MetaScanner.cpp:80 (Diff revision 1) > , strBuf(jArray<char16_t, int32_t>::newJArray(36)) > , content(nullptr) > , charset(nullptr) > , httpEquivState(HTTP_EQUIV_NOT_SEEN) > , treeBuilder(tb) > + , mEncoding(nullptr) Please file a follow-up to make sure this doesn't get overwritten the next time the parser is retranslated.

Comment on attachment 8881094 [details] Bug 1373984 - Turn nsIDocument::mCharacterSet into mozilla::NotNull<const mozilla::Encoding*>. https://reviewboard.mozilla.org/r/152408/#review157480 > No need to use a string here anymore, AFAICT. Good catch. Changed this to Encoding and UTF_8_ENCODING. > Why not `ForName`? Oh, .hintCharacterSet is already doing label resolution. Changed this to ForName. > Hasn't this already gone through label resolution on the doc shell level? I didn't read nsDocumentViewer::SetForceCharacterSet. Changed this to ForName. > Why not `ForName`? nsIEditor.documentCharacterSet is scriptable. If I change this to ForName, editor/libeditor/tests/test_documentCharacterSet.html will crash. It sets .documentCharacterSet from scripts and it will be used in the charset attribute of the meta element, so we need label resolution and we don't have to accept "replacement" here. I added a comment to crarify. > Please file a follow-up to make sure this doesn't get overwritten the next time the parser is retranslated. Filed bug 1376154.

Pushed by VYV03354@nifty.ne.jp: https://hg.mozilla.org/integration/autoland/rev/77af189b5c49 Turn nsIDocument::mCharacterSet into mozilla::NotNull<const mozilla::Encoding*>. r=hsivonen

Backed out for build bustage like https://treeherder.mozilla.org/logviewer.html#?job_id=109786760&repo=autoland https://hg.mozilla.org/integration/autoland/rev/ee1a057752b0

X86 calling conventions must die... https://treeherder.mozilla.org/#/jobs?repo=try&revision=0bdc873ffeb4712e26e3a8114d7154e4decea86a

Pushed by VYV03354@nifty.ne.jp: https://hg.mozilla.org/integration/autoland/rev/7235d05662b0 Turn nsIDocument::mCharacterSet into mozilla::NotNull<const mozilla::Encoding*>. r=hsivonen

Backed out for Android build bustage: https://hg.mozilla.org/integration/autoland/rev/e29483674e5a0f3e33bd41a24417bfd28a16861b Push with bustage: https://treeherder.mozilla.org/#/jobs?repo=autoland&revision=7235d05662b03741f84c383534a46fecc2075d58&filter-resultStatus=retry&filter-resultStatus=usercancel&filter-resultStatus=runnable&filter-resultStatus=testfailed&filter-resultStatus=busted&filter-resultStatus=exception Build log: https://treeherder.mozilla.org/logviewer.html#?job_id=109801919&repo=autoland [task 2017-06-25T16:24:36.816613Z] 16:24:36 INFO - /home/worker/workspace/build/src/sccache2/sccache /home/worker/workspace/build/src/android-ndk/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-g++ -std=gnu++11 -o Unified_cpp_gfx_angle0.o -c -I/home/worker/workspace/build/src/obj-firefox/dist/stl_wrappers -I/home/worker/workspace/build/src/obj-firefox/dist/system_wrappers -include /home/worker/workspace/build/src/config/gcc_hidden.h -DNDEBUG=1 -DTRIMMED=1 -D_CRT_SECURE_NO_DEPRECATE -D_HAS_EXCEPTIONS=0 -D_SECURE_SCL=0 -DANGLE_ENABLE_D3D9 -DANGLE_SKIP_DXGI_1_2_CHECK -DANGLE_COMPILE_OPTIMIZATION_LEVEL=D3DCOMPILE_OPTIMIZATION_LEVEL1 -DANGLE_NO_EXCEPTIONS -DGL_APICALL= -DGL_GLEXT_PROTOTYPES= -DEGLAPI= -DNOASSERT_UNIMPLEMENTED=0 -DANGLE_ENABLE_HLSL=1 -DANGLE_ENABLE_GLSL=1 -DANGLE_ENABLE_ESSL=1 -DANGLE_ENABLE_KEYEDMUTEX=1 -DSTATIC_EXPORTABLE_JS_API -DMOZ_HAS_MOZGLUE -DMOZILLA_INTERNAL_API -DIMPL_LIBXUL -I/home/worker/workspace/build/src/gfx/angle -I/home/worker/workspace/build/src/obj-firefox/gfx/angle -I/home/worker/workspace/build/src/gfx/angle/include -I/home/worker/workspace/build/src/gfx/angle/src -I/home/worker/workspace/build/src/gfx/angle/src/common/third_party/numerics -I/home/worker/workspace/build/src/obj-firefox/dist/include -I/home/worker/workspace/build/src/obj-firefox/dist/include/nspr -I/home/worker/workspace/build/src/obj-firefox/dist/include/nss -fPIC -DMOZILLA_CLIENT -include /home/worker/workspace/build/src/obj-firefox/mozilla-config.h -MD -MP -MF .deps/Unified_cpp_gfx_angle0.o.pp -idirafter /home/worker/workspace/build/src/android-ndk/platforms/android-9/arch-arm/usr/include -Wall -Wc++11-compat -Wempty-body -Wignored-qualifiers -Woverloaded-virtual -Wpointer-arith -Wsign-compare -Wtype-limits -Wunreachable-code -Wwrite-strings -Wno-invalid-offsetof -Wno-error=maybe-uninitialized -Wno-error=deprecated-declarations -Wno-error=array-bounds -Wno-error=free-nonheap-object -Wformat -fno-short-enums -fno-exceptions -march=armv7-a -mthumb -mfpu=vfp -mfloat-abi=softfp -mno-unaligned-access -I/home/worker/workspace/build/src/android-ndk/sources/cxx-stl/llvm-libc++/libcxx/include -I/home/worker/workspace/build/src/android-ndk/sources/android/support/include -I/home/worker/workspace/build/src/android-ndk/sources/cxx-stl/llvm-libc++abi/libcxxabi/include -fno-exceptions -fno-strict-aliasing -fno-rtti -ffunction-sections -fdata-sections -fno-exceptions -fno-math-errno -D_GLIBCXX_USE_CXX11_ABI=0 -pipe -g -freorder-blocks -fno-reorder-functions -Os -funwind-tables -Wno-attributes -Wno-shadow -Wno-sign-compare -Wno-unknown-pragmas -Wno-unreachable-code -Wno-shadow-compatible-local -Wno-shadow-local /home/worker/workspace/build/src/obj-firefox/gfx/angle/Unified_cpp_gfx_angle0.cpp [task 2017-06-25T16:24:36.831999Z] 16:24:36 INFO - cc1plus: error: to generate dependencies you must specify either -M or -MM [task 2017-06-25T16:24:36.832592Z] 16:24:36 INFO - /home/worker/workspace/build/src/config/rules.mk:1010: recipe for target 'Unified_cpp_parser_html0.o' failed [task 2017-06-25T16:24:36.833326Z] 16:24:36 INFO - gmake[5]: *** [Unified_cpp_parser_html0.o] Error 1 [task 2017-06-25T16:24:36.833686Z] 16:24:36 INFO - gmake[5]: *** Waiting for unfinished jobs....

Apparently Android builds dislikes using |nsIContentSink::Encoding;| and |nsIContentSink::NotNull;| https://treeherder.mozilla.org/#/jobs?repo=try&revision=a85ddf5755dedeb22a5447447933e8fe7a8ffe8e

Pushed by VYV03354@nifty.ne.jp: https://hg.mozilla.org/integration/autoland/rev/64412d8b6f4b Turn nsIDocument::mCharacterSet into mozilla::NotNull<const mozilla::Encoding*>. r=hsivonen

Were the BoxObject and ImageLayer changes intentionally part of the bustage fix to make the namespaces work? (The ImageLayer change doesn't looks like a mere bustage fix.)

(In reply to Henri Sivonen (:hsivonen) from comment #15) > Were the BoxObject and ImageLayer changes intentionally part of the bustage > fix to make the namespaces work? (The ImageLayer change doesn't looks like a > mere bustage fix.) The MozReview interdiff is fundamentally broken (bug 1233076). The only actual my change is: - using nsIContentSink::Encoding; - using nsIContentSink::NotNull; + using Encoding = mozilla::Encoding; + template <typename T> using NotNull = mozilla::NotNull<T>;