Speedup canned small-icu build

configure

@@ -952,7 +952,10 @@ def configure_intl(o):
           print '\n ** Corrupted ZIP? Delete %s to retry download.\n' % targetfile
     return None
   icu_config = {
-    'variables': {}
+    'variables': {
+      'icu_small_canned': b(False),
+      'icu_need_swap': b(True),
+    }
   }
   icu_config_name = 'icu_config.gypi'
   def write_config(data, name):
@@ -1046,7 +1049,7 @@ def configure_intl(o):
 
   if (o['variables']['icu_small'] == b(True)) and using_default_locales and (not with_icu_source) and canned_icu_available:
     # OK- we can use the canned ICU.
-    icu_config['variables']['icu_small_canned'] = 1
+    icu_config['variables']['icu_small_canned'] = b(True)
     icu_full_path = canned_icu_dir
 
   # --with-icu-source processing
@@ -1144,6 +1147,9 @@ def configure_intl(o):
   # this is the input '.dat' file to use .. icudt*.dat
   # may be little-endian if from a icu-project.org tarball
   o['variables']['icu_data_in'] = icu_data_in
+  if icu_data_in.endswith('%s.dat' % icu_endianness):
+    # we may still swap, but mark that we can skip it if need be
+    icu_config['variables']['icu_need_swap'] = b(False)
   # this is the icudt*.dat file which node will be using (platform endianness)
   o['variables']['icu_data_file'] = icu_data_file
   if not os.path.isfile(icu_data_path):

tools/icu/icu-generic.gyp

@@ -261,7 +261,8 @@
                               '-F', 'icu_small.json',
                               '-O', 'icudt<(icu_ver_major)<(icu_endianness).dat',
                               '-v',
-                              '-L', '<(icu_locales)'],
+                              '-L', '<(icu_locales)',
+                              '-K', '<(icu_small_canned)'],
                 },
                 {
                   # build final .dat -> .obj
@@ -325,7 +326,17 @@
               # link against stub data (as primary data)
               # then, use icupkg and genccode to rebuild small data
               'dependencies': [ 'icustubdata', 'genccode#host', 'icupkg#host', 'genrb#host', 'iculslocs#host',
-                               'icu_implementation', 'icu_uconfig' ],
+                                'icu_implementation', 'icu_uconfig' ],
+              'conditions': [
+                ['icu_small_canned == "true"', {
+                  # with canned data, don’t need genrb or iculslocs
+                  'dependencies!': [ 'genrb#host', 'iculslocs#host' ],
+                }],
+                ['icu_need_swap == "false" and icu_small_canned == "true"', {
+                  # canned data, so don’t need to swap
+                  'dependencies!': [ 'icupkg#host' ],
+                }],
+              ],
               'export_dependent_settings': [ 'icustubdata' ],
               'actions': [
                 {
@@ -342,7 +353,8 @@
                               '-F', 'icu_small.json',
                               '-O', 'icudt<(icu_ver_major)<(icu_endianness).dat',
                               '-v',
-                              '-L', '<(icu_locales)'],
+                              '-L', '<(icu_locales)',
+                              '-K', '<(icu_small_canned)'],
                 }, {
                   # rename to get the final entrypoint name right
                    'action_name': 'rename',
@@ -512,7 +524,25 @@
       'conditions': [
         ['OS == "solaris"', {
           'defines': [ '_XOPEN_SOURCE_EXTENDED=0' ]
-        }]
+        }],
+        ['OS != "win" and icu_need_swap == "false" and icu_small_canned == "true"', {
+          # we are ONLY building genccode.
+          # pare down the dependencies dramatically.
+          'sources!': [ '<@(icu_src_i18n)',
+                        '<@(icu_src_io)', ],
+          'defines!': [ 'UCONFIG_NO_BREAK_ITERATION=0' ],
+          'defines': [
+            #disables the 'year' stuff
+            'UCONFIG_NO_FORMATTING=1',
+            'UCONFIG_NO_REGULAR_EXPRESSIONS=1',
+            'UCONFIG_NO_NORMALIZATION=1',
+            'UCONFIG_NO_SERVICE=1',
+            'UCONFIG_NO_IDNA=1',
+            # should be ok to disable conversion in this case
+            # .. but http://bugs.icu-project.org/trac/ticket/11046
+            #'UCONFIG_NO_CONVERSION=1',
+        ]
+        }],
       ],
       'direct_dependent_settings': {
         'include_dirs': [
@@ -581,6 +611,10 @@
         '<@(icu_src_genccode)',
         'no-op.cc',
       ],
+      'defines': [
+        #disables the 'current year' stuff
+        'UCONFIG_NO_FORMATTING=1',
+      ]
     },
   ],
 }

tools/icu/icutrim.py

@@ -1,6 +1,6 @@
 #!/usr/bin/python
 #
-# Copyright (C) 2014 IBM Corporation and Others. All Rights Reserved.
+# Copyright (C) 2014,2016 IBM Corporation and Others. All Rights Reserved.
 #
 # @author Steven R. Loomis <srl@icu-project.org>
 #
@@ -65,6 +65,10 @@
                     action="count",
                     default=0)
 
+parser.add_option("-K","--canned",
+                    action="count",
+                    default=0)
+
 parser.add_option('-L',"--locales",
                   action="store",
                   dest="locales",
@@ -170,7 +174,14 @@ def runcmd(tool, cmd, doContinue=False):
 ## The first letter of endian_letter will be 'b' or 'l' for big or little
 endian_letter = options.endian[0]
 
-runcmd("icupkg", "-t%s %s %s""" % (endian_letter, options.datfile, outfile))
+if ( options.datfile.endswith("%s.dat" % endian_letter) ):
+    shutil.copyfile(options.datfile, outfile)
+else:
+    runcmd("icupkg", "-t%s %s %s""" % (endian_letter, options.datfile, outfile))
+
+if ( options.canned >  0 ):
+    print "Canned small ICU - no need to actually trim."
+    sys.exit(0)
 
 ## STEP 2 - get listing
 listfile = os.path.join(options.tmpdir,"icudata.lst")