Issue 3047 - Change default search engines

mozharness/abb/transform_locales.py

Index: mozharness/abb/transform_locales.py
diff --git a/mozharness/abb/transform_locales.py b/mozharness/abb/transform_locales.py
new file mode 100644
index 0000000000000000000000000000000000000000..650b4d5fb8c22eee7236b3cc1422a9d9dd206922
--- /dev/null
+++ b/mozharness/abb/transform_locales.py
@@ -0,0 +1,182 @@
+# This file is part of Adblock Plus
+# Copyright (C) 2006-2015 Eyeo GmbH
+#
+# Adblock Plus is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License version 3 as
+# published by the Free Software Foundation.
+#
+# Adblock Plus is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Adblock Plus.  If not, see <http://www.gnu.org/licenses/>.
+
+import os
+import re
+
+_LOCALE_RE = re.compile("^[a-z]{2}(-[A-Z]{2})?$")
+_SEARCH_PROPS_RE = re.compile("^browser\.search\."
+                              "(defaultenginename|order\.).*$")
+_SHORTNAME_RE = re.compile("^<ShortName>.*</ShortName>$")
+
+_SEARCHPLUGINS_PATH = os.path.join("browser", "searchplugins")
+_LIST_TXT_PATH = os.path.join(_SEARCHPLUGINS_PATH, "list.txt")
+_REGION_PROPS_PATH = os.path.join("browser", "region.properties")
+
+_DEFAULT_LOCALE = "en-US"
+
+_SEARCH_ENGINE_ORDER = {
+    "en-US": ["duckduckgo",
+              "yahoo",
+              "google",
+              "wikipedia",
+              "amazon"
+              ],
+    "zh-CN": ["baidu",
+              "duckduckgo",
+              "yahoo",
+              "google",
+              "wikipedia",
+              "amazon"
+              ]
+}
+
+
+def _get_shortname_from_id(needle, engine_ids, engine_names):
+    """Fuzzy finds needle in engine_ids and returns ShortName"""
+    regex = re.compile("^%s.*$" % needle)
+    for engine in engine_ids:
+        if regex.match(engine.lower()):

[Felix Dahlke, 2015/09/21 20:46:31]

Sorry for not realising this earlier, but wouldn't engine.startswith(needle) do
the trick here?

[René Jeschke, 2015/09/22 10:52:02]

Done.

+            return engine_names[engine]
+    return None
+
+
+def _write_lines(filename, lines):
+    """Writes lines into file appending \\n"""
+    with open(filename, "w") as fd:
+        for l in lines:
+            fd.write("%s\n" % l)
+
+
+def _transform_locale(locale, path, logger):
+    logger.info("Processing locale '%s'..." % locale)
+
+    # Check for list.txt existence
+    list_file_path = os.path.join(path, _LIST_TXT_PATH)
+    if not os.path.exists(list_file_path):
+        logger.fatal("Missing 'list.txt' for locale '%s'" % locale)
+
+    # Check for region.properties existence
+    region_file_path = os.path.join(path, _REGION_PROPS_PATH)
+    if not os.path.exists(region_file_path):
+        logger.fatal("Missing 'region.properties' for locale '%s'" % locale)
+
+    # Get whitelist and build regex
+    whitelist = _SEARCH_ENGINE_ORDER.get(locale,
+                                         _SEARCH_ENGINE_ORDER[_DEFAULT_LOCALE])
+    white_re = re.compile("^(%s).*$" % "|".join(whitelist))
+
+    # Read engine IDs from list.txt, discard engines not on whitelist
+    engine_ids = []
+    with open(list_file_path, "r") as fd:
+        for line in fd:
+            line = line.strip()
+            if len(line) > 0:
+                if white_re.match(line.lower()):

[Felix Dahlke, 2015/09/21 20:46:31]

I think we should do proper case insensitive matching here. The semantics differ
slightly from the current version, but to me it seems that we actually want a
plain old case insensitive comparison here anyway.

white_re = re.compile(..., re.IGNORECASE)
white_re.match(line)

[René Jeschke, 2015/09/22 10:52:02]

I threw out the case insensitive stuff completely, doesn't really make sense, as
also Mozilla does 'real matches' IIRC.

+                    engine_ids.append(line)
+                else:
+                    logger.info("Removing '%s'" % line)
+
+    # Make sure we still have search engines left
+    if len(engine_ids) == 0:
+        logger.fatal("No search engines left over for '%s'" % locale)
+
+    # 'Parse' XML to get matching 'ShortName' for all engine IDs
+    engine_names = {}
+    for eid in engine_ids:
+        xml_file_path = os.path.join(path, _SEARCHPLUGINS_PATH, "%s.xml" % eid)
+        if os.path.exists(xml_file_path):
+            short_name = None
+            with open(xml_file_path, "r") as fd:
+                for line in fd:
+                    line = line.strip()
+                    if _SHORTNAME_RE.match(line):
+                        short_name = line[11:-12].strip()

[Felix Dahlke, 2015/09/21 20:46:31]

Wow, just realised the hard coded offsets - that seems a bit fragile to me. Why
not just use a group to match the text content of the <ShortName> element?

[René Jeschke, 2015/09/22 10:52:02]

Done.

+
+            if not short_name:
+                logger.fatal("No ShortName defined for '%s' in '%s" %
+                             (eid, locale))
+            engine_names[eid] = short_name
+        else:
+            logger.fatal("XML definiton for '%s' in '%s' missing" %
+                         (eid, locale))
+
+    logger.info("Remaining engine IDs: %s" % ", ".join(engine_ids))
+
+    # Create search engine order with real engine names
+    engine_order = []
+    for eid in whitelist:
+        sn = _get_shortname_from_id(eid, engine_ids, engine_names)
+        if sn:
+            engine_order.append(sn)
+
+    logger.info("Resulting ordered list: %s" % (", ".join(engine_order)))
+
+    # Read region.properties and remove browser.search.* lines
+    props = []
+    with open(region_file_path, "r") as fd:
+        for line in fd:
+            line = line.rstrip("\r\n")
+            if not _SEARCH_PROPS_RE.match(line.strip()):
+                props.append(line)
+
+    # Append default search engine name
+    props.append("browser.search.defaultenginename=%s" % engine_order[0])
+
+    # Append search engine order
+    for i in range(0, min(3, len(engine_order))):
+        props.append("browser.search.order.%d=%s" % (i + 1, engine_order[i]))
+
+    # Write back list.txt
+    _write_lines(list_file_path, engine_ids)
+
+    # Write back region.properties
+    _write_lines(region_file_path, props)
+
+
+class MinimalLogger:
+    def info(self, s):
+        print "INFO: %s" % s
+
+    def error(self, s):
+        print "ERROR: %s" % s
+
+    def fatal(self, s):
+        print "FATAL: %s" % s
+        exit(1)
+
+
+def transform_locales(build_object, obj_dir):

[Felix Dahlke, 2015/09/21 20:46:31]

Maybe it's just me, but I think this would be more elegant:

def transform_locales(obj_dir, logger=MinimalLogger()):

[René Jeschke, 2015/09/22 10:52:02]

Done.

+    logger = build_object or MinimalLogger()
+
+    chrome_path = os.path.join(obj_dir, "dist", "bin", "chrome")
+    if not os.path.exists(chrome_path):
+        logger.fatal("'%s' does not exist" % obj_dir)
+
+    locales = []
+    for p in next(os.walk(chrome_path))[1]:
+        if _LOCALE_RE.match(p):
+            locales.append(p)
+    locales.sort()
+
+    logger.info("Found %d locales" % len(locales))
+
+    for locale in locales:
+        locale_path = os.path.join(chrome_path, locale, "locale", locale)
+        if os.path.exists(locale_path):
+            _transform_locale(locale, locale_path, logger)
+        else:
+            logger.error("Missing 'locale' folder for '%s'" % locale)
+