summaryrefslogtreecommitdiff
path: root/samples/BUILD.gn
diff options
context:
space:
mode:
authorTom Sepez <tsepez@chromium.org>2015-07-30 10:29:19 -0700
committerTom Sepez <tsepez@chromium.org>2015-07-30 10:29:19 -0700
commit22955d987b30056d787d5eeb1a8ac8a1277bf5e0 (patch)
treeb109acf2d1ba26d4676357e4952a4baabba2af26 /samples/BUILD.gn
parent452bafebc358d9a1d57266b31a5869584c0fbfb7 (diff)
downloadpdfium-22955d987b30056d787d5eeb1a8ac8a1277bf5e0.tar.xz
Use bundled freetype for the pdfium_test binary.
But use platform freetype for library itself according to the rules for the platform. This should greatly reduce per-platform diffs in the corpus tests, but requires that the corpus be rolled at the same time. When this rolls into chromium, its src/BUILD.gn will need to be updated to say third_party:fx_freetype instead of third_party:freetype. R=jam@chromium.org Review URL: https://codereview.chromium.org/1267493005 .
Diffstat (limited to 'samples/BUILD.gn')
-rw-r--r--samples/BUILD.gn4
1 files changed, 4 insertions, 0 deletions
diff --git a/samples/BUILD.gn b/samples/BUILD.gn
index 748a2af399..6f4518c17e 100644
--- a/samples/BUILD.gn
+++ b/samples/BUILD.gn
@@ -24,6 +24,10 @@ executable("pdfium_test") {
]
deps = [
"//third_party/pdfium",
+ # Regardless of whether the library ships against system freetype,
+ # always link this binary against the bundled one for consistency
+ # of results across platforms.
+ "//third_party/pdfium/third_party:fx_freetype",
"//v8:v8_libplatform",
":fx_lpng",
]