bat release focus: performance

[sharkdp]

For the next major release of bat, we decided to focus on performance.

This issue serves as a forum for discussion and as a summary of tasks that can be worked on:

• Get Improved ANSI passthrough. #1596 by @eth-p merged. You can help by reviewing the code.
• Work on Improve bat startup speed #951
  • Get Reduce startup time in loop-through mode by 80%-90% #1747 by @Enselic merged
  • Fix Loading assets from the cache is significantly slower than from the binary #1753
  • Discuss / review the prototype suggested by @Enselic in this comment
• Work on bat loop-through speed, see Moving to the end of 6 Mb file takes forever #304
• Set up some tooling to improve bat benchmarking. See tests/benchmarks for a start.

Other performance-related tickets:

• bat freezes with extreme long lines (compact JSON) #1658
• Is it expected that Bat is rather slow on Windows? #1630
• bat slow when doing a git lg and using it as the pager compared to less #1147
• Slowness/choppy with specific Python syntax #1688

[federicotorrielli]

Any updates on bat using the splice?

[sharkdp]

Any updates on bat using the splice?

I'm assuming you are referring to https://endler.dev/2018/fastcat. It would have been nice to at least put a reference. If that's indeed your question, there is a ticket for that already: #304 (and #1112, specifically mentioning "splice"). It's also linked in the top-level comment.

[federicotorrielli]

Yes. I was referring to a future splice call implementation. Since those issue did not really have news, issue #194 was negative (but old) and the person who created #1112 did not make any comments on the issue #304 (that did not mention splice, btw), I was here to ask if we are seeing any new update on this or not. Thanks anyway!

[Enselic]

Not sure this issue is worth keeping open any longer? I'll go ahead and close and unpin it.

We could pin the remaining issues individually, but I'm not sure they are important enough to fix to justify getting such a prominent position.

For example #2085 would by my estimation provide more value to users to fix, so it would be higher up in my list of issues that deserves pinning.