[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 0/2] Trivial fixes for setup.rb usage.
From: |
James Rowe |
Subject: |
Re: [PATCH 0/2] Trivial fixes for setup.rb usage. |
Date: |
Mon, 2 Dec 2019 13:27:14 +0000 |
User-agent: |
NeoMutt/20180716 |
* Eric Wong (address@hidden) wrote:
> James Rowe <address@hidden> wrote:
> > Given that the `Config` name was removed so long ago, an alternative fix
> > might be to remove `setup.rb` altogether?
>
> RubyGems is a huge barrier to startup performance; so users
> should be able to opt-out of using gems. A painful thing is
> just having more gems installed (and even not using them)
> slows down startup.
Wow, I hadn’t realised how costly that was. As another data point for
your thoughts, my low-energy mobile Athlon reports:
▪ ~/P/dtas ψ:(master) ▶hyperfine --warmup 10 'dtas-ctl current'
Benchmark #1: dtas-ctl current
Time (mean ± σ): 236.0 ms ± 17.0 ms [User: 185.8 ms, System: 24.8 ms]
Range (min … max): 226.1 ms … 282.0 ms 11 runs
▪ ~/P/dtas ψ:(master) ▶hyperfine --warmup 10 'ruby --disable=g -Ilib
./bin/dtas-ctl current'
Benchmark #1: ruby --disable=g -Ilib ./bin/dtas-ctl current
Time (mean ± σ): 30.6 ms ± 6.3 ms [User: 21.1 ms, System: 4.6 ms]
Range (min … max): 27.0 ms … 63.9 ms 65 runs
On my system it makes little difference if you bypass gem’s script
wrapper unless you forcibly disable gems support like you suggest.
Thanks,
James
--
☒ Vote to destroy email signatures today!
signature.asc
Description: PGP signature