Reduce memory usage of .fetch_path
#55
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Merged
NickLaMuro
force-pushed
the
reduce_fetch_path_memory_usage
branch
from
b6d8121
to
77e7a48
Compare
By using recursion to loop through the keys for fetch_path, we are
creating new objects that are not needed every time with go further down
into the stack, because we are required to do a `args[1..-1].push`,
which will create a new array from the subset of the args, and pass it
into the next call of `.fetch_path`.
By not using recursion, we can completely remove any object allocations
that aren't necessary (besides the ones that are needed for the method's
spec), and still maintain the same functionality. The added benefit is
that this also now not prone to `stack level too deep` errors because it
no longer requires recursion to function.
Of note, I attempted to use both `.each_with_object` and `.inject`
instead of implementing something myself. But each had it's own
problems that were either broken or sub-optimal:
The `.each_with_object` implementation looked something like this:
args.each.with_index.each_with_object(self) do |(key, i), result|
Obviously not requiring the `result = self` line above it. This just failed to
function in some scenarios and I was unable to determine the cause.
The `.inject` implementation was similar to the above
`.each_with_object` implementation:
args.each.with_index.inject(self) do |result, (key, i)|
While actually functioning as expected, and just making sure to end the
block with a final call of `result`, the combination of having to do
`.each`, `.with_index`, and `.inject` to achieve having both a indexed
loop and a resulting object actually caused some extra objects to be
created.
Being explicit here and avoiding the ruby convenience methods seemed to
both prevent extraneous objects instantiation and provide the results
expected.
|
Checked commit NickLaMuro@77e7a48 with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0 |
|
🎊 |
Fryguy
added a commit
that referenced
this pull request
Aug 11, 2017
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Similar rational to #54
A note about the implementation
Of note, I attempted to use both
.each_with_objectand.injectinstead of implementing something myself. But each had it's own problems that were either broken or sub-optimal:.each_with_objectThe
.each_with_objectimplementation looked something like this:Obviously this allows us to omit the
result = selfline above it, but this implementation just failed to function in some scenarios and I was unable to determine the cause..injectThe
.injectimplementation was similar to the above.each_with_objectimplementation:While actually functioning as expected, and just making sure to end the block with a final call of
result, the combination of having to do.each,.with_index, and.injectto achieve having both a indexed loop and a resulting object actually caused some extra objects to becreated.
Being explicit here and avoiding the ruby convenience methods seemed to both prevent extraneous objects instantiation and provide the results expected.
Benchmarks
I used the existing benchmarks for #54 (an effort originating from ManageIQ/manageiq#15757) as a base, and applied these changes on top of them when doing the benchmarks to see the over all effect.
.inject).inject)Comparing the objects allocated by gem, we can see that
more_core_extensionsis the only one to change and improve between runs:Links
more_core_extensions: Reduce memory usage of.store_path#54