[PATCH] D124983: [HLSL} add -fcgl option flag.

[Xiang Li via Phabricator via cfe-commits]

python3kgae marked 2 inline comments as done.
python3kgae added inline comments.


================
Comment at: clang/include/clang/Driver/Options.td:6765
          "as_6_5, as_6_6, as_6_7">;
+def fcgl : DXCFlag<"fcgl">,
+  HelpText<"Generate high-level code only. Without any dxil related passes.">;
----------------
aaron.ballman wrote:
> beanz wrote:
> > aaron.ballman wrote:
> > > Any particular reason for this name? Naively, I have no idea what this does or how it relates to code generation.
> > I have no idea what this name is short for… code gen language?… likeness?… lobotomy?…
> > 
> > We have this flag in DXC, and it is a useful shorthand.
> > 
> > Don’t know if @python3kgae or @pow2clk know what the actual abbreviation is.
> Can we pick a more descriptive name, or would that make portability too much of a challenge? Could we have a more descriptive name and then have `-fcgl` be an alias for it, or is that overkill because this is a rarely-used option?
Sure. I'll add a more descriptive name and alias fcgl to it.

It is something like frontend codeGen listing if I remember correctly.


================
Comment at: clang/unittests/Driver/ToolChainTest.cpp:613
 
+TEST(DxcModeTest, ValidatorFCGL) {
+  IntrusiveRefCntPtr<DiagnosticIDs> DiagID(new DiagnosticIDs());
----------------
aaron.ballman wrote:
> Any reason not to use the typical lit tests for this? It seems like you should be able to add a Driver test that does `-###` and looks for the expected pass-down flags. And then you can add a -cc1 test to ensure that the `-fcgl` flag isn't supported there.
I'll use -### to add a lit test.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D124983/new/

https://reviews.llvm.org/D124983