Optimize CPP code generation.
Previously, we were limiting our calls to appendVAList to 512-byte
chunks to avoid a potential buffer overflow. This had two problems:
1 - it did not avoid the buffer overflow :(
see chromium:1092743
2 - every call to appendVAList is expensive; it incurs a resize of the
code buffer (alloc new, memcpy, free old)
This CL removes the 512-byte cap as the buffer overflow issue was
resolved at http://review.skia.org/297276
This CL also includes a few more minor improvements.
- `codeAppendf` now uses a raw string so the gencode is easier to read.
- Optimized `SkStringPrintf("%s", foo)` to `SkString(foo)`.
- Optimized `strA = strB` to `strA.swap(strB)` where safe to do so.
Change-Id: Ia0909a68719848dd2ca655066a9bc6929c8fd09f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/297358
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
diff --git a/src/gpu/gradients/generated/GrLinearGradientLayout.cpp b/src/gpu/gradients/generated/GrLinearGradientLayout.cpp
index 2d28b92..5bc1732 100644
--- a/src/gpu/gradients/generated/GrLinearGradientLayout.cpp
+++ b/src/gpu/gradients/generated/GrLinearGradientLayout.cpp
@@ -28,7 +28,9 @@
SkString sk_TransformedCoords2D_0 = fragBuilder->ensureCoords2D(
args.fTransformedCoords[0].fVaryingPoint, _outer.sampleMatrix());
fragBuilder->codeAppendf(
- "half t = half(%s.x) + 9.9999997473787516e-06;\n%s = half4(t, 1.0, 0.0, 0.0);\n",
+ R"SkSL(half t = half(%s.x) + 9.9999997473787516e-06;
+%s = half4(t, 1.0, 0.0, 0.0);
+)SkSL",
sk_TransformedCoords2D_0.c_str(), args.fOutputColor);
}