mirrors/jj
1
0
Fork 1
mirror of https://github.com/martinvonz/jj.git synced 2024-12-04 05:28:02 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 12
jj/testing/bench-revsets-git.txt

48 lines
1 KiB
Text
Raw Normal View History

cli: add a command for benchmarking many revsets This command is similar to Mercurial's revset benchmarking command. It lets you pass in a file containing revsets. I also included a file with some revsets to test on the git.git repo. I put it in `testing/`, which doesn't seem perfect. I'm happy to hear suggestions for better places, or we can move it later if we find a better place. Note that these tests don't clear caches between each run (or even between tests), so revsets that rely on filtering commit data that's not indexed appear faster than they typically are in reality.
2023-03-27 16:36:00 +00:00
# Revsets to pass to `jj bench revsets` on the Git
# Single tags
v1.0.0
v2.40.0
# Old history
:v1.0.0
..v1.0.0
# More history
:v2.40.0
..v2.40.0
# Only recent history
v2.39.0..v2.40.0
:v2.40.0 ~ :v2.39.0
v2.39.0:v2.40.0
# Tags and branches
tags()
branches()
# Intersection of range with a small subset
tags() & :v2.40.0
v2.39.0 & :v2.40.0
# Author and committer
author(peff)
committer(gitster)
# Intersection and union of large subsets
author(peff) & committer(gitster)
author(peff) | committer(gitster)
revset: do not always evaluate filter node to InternalRevset This basically removes hidden 'all() &' from union/negation of filters. To achieve that, I have two options: 1. add separate evaluation path (like the one this commit introduced), or 2. wrap "all()" revset to override predicate as Box::new(|_| true) function. I took the former since it's less ad-hoc. We can add an explicit RevsetExpression node to branch between evaluate() and evaluate_predicate(), but I don't think it would simplify the implementation at this point. We might need such node if we want to resolve "all()" at resolve_symbols(). It might be even better to extract a subset of RevsetExpression enum, which only contains evaluatable nodes. The cost of 'all() &' isn't significant for most filters. '~merges()' is the exception. For jj repo, revsets/:v0.3.0 & (author(martinvonz) | committer(martinvonz)) -------------------------------------------------------------- base 1.06 11.2±0.04m new 1.00 10.5±0.05m revsets/~merges() ----------------- base 1.69 750.0±8.47µ new 1.00 444.1±3.50µ
2023-04-01 07:41:47 +00:00
# Intersection of filter with a small subset
:v1.0.0 & (author(peff) & committer(gitster))
:v1.0.0 & (author(peff) | committer(gitster))
cli: add a command for benchmarking many revsets This command is similar to Mercurial's revset benchmarking command. It lets you pass in a file containing revsets. I also included a file with some revsets to test on the git.git repo. I put it in `testing/`, which doesn't seem perfect. I'm happy to hear suggestions for better places, or we can move it later if we find a better place. Note that these tests don't clear caches between each run (or even between tests), so revsets that rely on filtering commit data that's not indexed appear faster than they typically are in reality.
2023-03-27 16:36:00 +00:00
# Roots and heads of small subsets
roots(tags())
heads(tags())
# Roots and heads of large subsets
roots(author(peff))
heads(author(peff))
# Roots and heads of range
roots(:v2.40.0)
heads(:v2.40.0)
# Parents and children of small subset
tags()-
tags()+
revset: do not always evaluate filter node to InternalRevset This basically removes hidden 'all() &' from union/negation of filters. To achieve that, I have two options: 1. add separate evaluation path (like the one this commit introduced), or 2. wrap "all()" revset to override predicate as Box::new(|_| true) function. I took the former since it's less ad-hoc. We can add an explicit RevsetExpression node to branch between evaluate() and evaluate_predicate(), but I don't think it would simplify the implementation at this point. We might need such node if we want to resolve "all()" at resolve_symbols(). It might be even better to extract a subset of RevsetExpression enum, which only contains evaluatable nodes. The cost of 'all() &' isn't significant for most filters. '~merges()' is the exception. For jj repo, revsets/:v0.3.0 & (author(martinvonz) | committer(martinvonz)) -------------------------------------------------------------- base 1.06 11.2±0.04m new 1.00 10.5±0.05m revsets/~merges() ----------------- base 1.69 750.0±8.47µ new 1.00 444.1±3.50µ
2023-04-01 07:41:47 +00:00
# Filter that doesn't read commit object
merges()
~merges()
cli: add a command for benchmarking many revsets This command is similar to Mercurial's revset benchmarking command. It lets you pass in a file containing revsets. I also included a file with some revsets to test on the git.git repo. I put it in `testing/`, which doesn't seem perfect. I'm happy to hear suggestions for better places, or we can move it later if we find a better place. Note that these tests don't clear caches between each run (or even between tests), so revsets that rely on filtering commit data that's not indexed appear faster than they typically are in reality.
2023-03-27 16:36:00 +00:00
# Files are unbearably slow, so only filter within small set
file(Makefile) & v1.0.0..v1.2.0
Reference in a new issue Copy permalink