performance improvements

[evgenigourvitch]

Hello guys,

We are using this kindly shared package and since we are using it in heavy loaded server environment (~400K requests/sec) each millisecond counts, so I added some flexibility to
the current implementation without affecting the base capabilities.
It allows to detect only selected parts (for example not to parse Device if not needed, like in our use-case).
I've also added some sorting capability of the regexp arrays to allow to reshuffle them into descending order by the most successful matches.

May be you will find it usable.
Best regards and thank you for the incredible package!

[elsigh]

Can you try the perf benchmarks with this change? See #26

[evgenigourvitch]

=== RUN TestOSParsing
--- PASS: TestOSParsing (0.04s)
=== RUN TestUAParsing
=== RUN TestDeviceParsing
=== RUN TestGenericParseMethodConcurrency
--- PASS: TestGenericParseMethodConcurrency (0.51s)
--- PASS: TestDeviceParsing (9.01s)
--- PASS: TestUAParsing (14.64s)
PASS
**

BenchmarkParser-4 50 22327841 ns/op
BenchmarkParserWithOptions-4 100 12900935 ns/op

**
ok github.com/streamrail/uap-go.evg/uaparser 17.813s

[elsigh]

Is this PR the source of Issue #33?

[evgenigourvitch]

Actually, I can see no issue, but as designed behaviour, like I discribed
in the issue's comment.
Thanks.

On Wed, Aug 10, 2016 at 7:37 PM, Lindsey Simon [email protected]
wrote:

Is this PR the source of Issue #33
#33?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#32 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/ATeefGDKFJAdFdtQf4ipVaxJHZN1qPrJks5qef5ggaJpZM4JUnKZ
.