Add big endian support

scripts/gen_datS.py

+#!/usr/bin/env python
+
+import binascii
+import sys
+
+if len(sys.argv)<3:
+        print "Provide input and output file names as arguments to use the script!"
+else:
+        input_file = sys.argv[1];
+        output_file = sys.argv[2];
+
+        input = open(input_file, 'rb').read()
+
+        output = open(output_file, 'w')
+
+        output.write(".globl icudt54_dat\n")

[jungshik at Google, 2016/05/12 08:52:52]

It'd be better to take a major version number as an input parameter (54, 56, 57,
58 .... ) or read off the header portion from ${ICU_SRC}/linux/icudtl_dat.S 

Given that this  script can be used for icudtl_dat.S (instead of checking that
in as we do now) as well as icudtb_dat.S, passing the major version number in
the command line should be fine. 

[jungshik at Google, 2016/05/12 17:27:15]

Or, a bit hacky approach of reading the major version number from
source/common/unicode/uvernum.h or README.chromium or from icudt[bl].dat   

strings common/icudtl.dat  | grep icudt | head -1 | sed
's/icudt\([5-9][0-9]\)[bl].*$/\1/'

Perhaps, the last is the most self-contained because this script is taking
icudt[bl].dat as an input. 

+        output.write("\t.section .note.GNU-stack,\"\",%progbits\n")
+        output.write("\t.section .rodata\n")
+        output.write("\t.balign 16\n")
+        output.write("#ifdef U_HIDE_DATA_SYMBOL\n")
+        output.write("\t.hidden icudt54_dat\n")
+        output.write("#endif\n")
+        output.write("\t.type icudt54_dat,%object\n")
+        output.write("icudt54_dat:\n")
+
+        split = [binascii.hexlify(input[i:i+4]).upper().lstrip('0') for i in range(0, len(input), 4)]
+
+        for i in range(len(split)):
+                if (len(split[i])==0):
+                        split[i]='0'
+                elif (len(split[i])==1):
+                        if not any((c in '123456789') for c in split[i]):
+                                split[i]='0x0'+split[i]
+                elif (len(split[i])%2==1):
+                        split[i]='0x0'+split[i]
+                else:
+                        split[i]='0x'+split[i]
+
+        for i in range(len(split)):

[jungshik at Google, 2016/05/12 08:52:52]

Can't you just combine this for-loop with the previous one? 

+                if (i%32==0):
+                        output.write("\n.long ")
+                else:
+                        output.write(",")
+                output.write(split[i])
+
+        output.write("\n")
+        output.close()