Mitigate spoofing attempt using Latin letters.

components/url_formatter/top_domains/README

Index: components/url_formatter/top_domains/README
diff --git a/components/url_formatter/top_domains/README b/components/url_formatter/top_domains/README
new file mode 100644
index 0000000000000000000000000000000000000000..804f4e722899ff2b9db9674c9fc51c82f3902485
--- /dev/null
+++ b/components/url_formatter/top_domains/README
@@ -0,0 +1,23 @@
+* alexa_10k_domains.list
+  It is an input to make_top_domain_list and is made up of list of Alexa
+  top 10k domains (one per line).
+  It's derived from
+   src/tools/perf/page_sets/alexa1-10000-urls.json by running the following:

[ncarter (slow), 2017/04/20 22:26:59]

IIRC the alexa10000 from page_sets was almost five years old. If we're relying
on this for user protection, we'll need to set up a process to update the list.

+
+  grep http ../../../tools/perf/page_sets/alexa1-10000-urls.json | \
+  sed -r -e 's;^.*"https?://(.*)/".*$;\1;'  -e 's/www\.//'  | \
+  awk 'BEGIN {FS="."} { printf("%s%s\n", NF > 3 ? "#" : "", $0); } \
+       END {printf ("# for testing\ndigklmo68.com\ndigklmo68.co.uk\n");}' > \

[ncarter (slow), 2017/04/20 22:26:58]

This would probably be better as a python script, since sed is not part of our
developer toolchain.

+  alexa_10k_domains.list
+
+* alexa_10k_names_and_skeletons.gperf
+
+  It is generated by running make_top_domain_list and checked in.
+  No command line argument needs to be passed.
+
+  $ ninja -C $build_outdir make_top_domain_list
+  $ $build_outdir/make_top_domain_list
+
+  During a build, it is processed by base/dafsa/make_dafsa.py to generate
+  alexa_10k_names_and_skeletons-inc.cc that is included by
+  components/url_formatter/url_formatter.cc