commit a88af523c6af5a4467abe59fcd17662fce6f8eaf
author jar@chromium.org <jar@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> Fri Jun 05 01:34:53 2009
committer jar@chromium.org <jar@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> Fri Jun 05 01:34:53 2009
tree 53c9a3a3856570f496f4446f66680804e87ebb87
parent 53a3c8fec3cccf37f81e808e4b17188d9c53cd71

Reduce probability of SDCH holdback to 1%

Anticipating a broader test base, I wanted to reduce the percentage of
uses that were "holdback" (sdch not used when it was available).  This
should reduce the impact on users, and also be sufficient to get
verification statistics from a larger group.

r=huanr
Review URL: http://codereview.chromium.org/118287

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@17700 0039d316-1c4b-4281-b951-d872f2087c98

net/url_request/url_request_http_job.cc

diff --git a/net/url_request/url_request_http_job.cc b/net/url_request/url_request_http_job.cc
index 85cb219..6ad2dc55 100644
--- a/net/url_request/url_request_http_job.cc
+++ b/net/url_request/url_request_http_job.cc
@@ -605,8 +605,8 @@
       // eventually record statistics via either SDCH_EXPERIMENT_DECODE or
       // SDCH_EXPERIMENT_HOLDBACK, and we'll need some packet timing data.
       EnablePacketCounting(kSdchPacketHistogramCount);
-      if (base::RandDouble() < .5) {
-        sdch_test_control_ = true;
+      if (base::RandDouble() < .01) {
+        sdch_test_control_ = true;  // 1% probability.
         advertise_sdch = false;
       } else {
         sdch_test_activated_ = true;