Hi Yasu, Thanks for reporting this. I believe the optimization we added for #gsub can also be used in #split. I see that you filed a ticket, we will follow up there. Laurent On Jan 16, 2011, at 10:12 AM, Yasu Imao wrote:
Hi again,
Another String performance issue. Tested on a text file with 4360 lines (\n) (different from the one I used with other tests). With these tests, 0.9 nightly is much much slower than 0.8, so I'll file a ticket for 0.9.
Test 1 - File.read("test.txt").split(/\n/) Test 2 - File.read("test.txt").split("\n") Test 1 - force_encoding - File.read("test.txt").force_encoding("UTF-16BE").split(/\n/) Test 2 - force_encoding - File.read("test.txt").force_encoding("UTF-16BE").split("\n".force_encoding("UTF-16BE"))
Test 3 - File.read("test.txt").lines {}
Ruby 1.8.7 Test 1 - 0.0081 0.0078 0.0086 Test 2 - 0.0078 0.0079 0.0079 Test 3 - 0.0058 0.0053 0.0065
Ruby 1.9.2 Test 1 - 0.0089 0.0090 0.0089 Test 2 - 0.0035 0.0035 0.0035 Test 3 - 0.0036 0.0035 0.0035
MacRuby 0.8 2011/01/16 Test 1 - 0.026 0.039 0.023 Test 2 - 0.019 0.018 0.018 Test 3 - 0.023 0.020 0.026
MacRuby 0.9 2011/01/16 Test 1 - 0.021 0.021 0.021 Test 2 - 23.51 23.55 23.50 Test 3 - 23.43 23.38 23.35
MacRuby 0.9 2011/01/16 - force_encoding Test 1 - 0.015 0.015 0.015 Test 2 - 0.011 0.011 0.011
_______________________________________________ MacRuby-devel mailing list MacRuby-devel@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macruby-devel