info: add --deps flag to --sizes for dependency footprint breakdown by EdenRochmanSharabi · Pull Request #22582 · Homebrew/brew

EdenRochmanSharabi · 2026-06-08T11:13:19Z

Have you followed the guidelines in our Contributing document?
Have you checked to ensure there aren't other open Pull Requests for the same change?
Have you added an explanation of what your changes do and why you'd like us to include them? Performance claims (e.g. "this is faster") must include Hyperfine benchmarks.
Have you written new tests (excluding integration tests) for your changes? Here's an example.
Have you successfully run brew lgtm (style, typechecking and tests) with your changes locally?

AI was used to generate or assist with generating this PR.

Git processing (branch management, rebasing, PR submission) was assisted by Claude. The implementation logic was written through manual analysis of the existing --sizes code path and Keg#runtime_dependencies.

Summary

Adds a --deps flag (depends on --sizes) to brew info that shows dependency size breakdown:

brew info --sizes --deps displays a table with Direct, Exclusive Deps, and Total Footprint columns for all installed formulae
brew info --sizes --deps <formula> shows detailed breakdown including exclusive and shared dependencies
brew info --sizes --deps -v <formula> shows the full dependency listing with sizes

This addresses the feedback from #22391 where @MikeMcQuaid suggested integrating footprint analysis into --sizes rather than creating a separate command. Replaces #22576 which couldn't be reopened after I mistakenly deleted the branch.

Notes

Tests and brew lgtm not yet run (will address once the approach is confirmed)
The --eval-all flag was marked as hidden/deprecated in the same commit; happy to split that into a separate PR if preferred

When used with --sizes, the new --deps flag shows the true disk cost of installed formulae by analyzing exclusive vs shared dependencies. This addresses the feedback from Homebrew#22391 to integrate footprint analysis directly into `brew info --sizes` rather than as a separate command.

MikeMcQuaid

You definitely promise you wrote 100% of the actual code by hand and just used Claude for analysis?

Need to slim this down a lot, it's a lot of code for a relatively simple feature. Also needs some (non-integration) tests.

Thanks!

EdenRochmanSharabi · 2026-06-08T12:11:04Z

Yes, I wrote all the code by hand. I used Claude to review it and suggest improvements, but I always rewrote those suggestions myself to avoid the endless slop you get from AI-generated code. That's why I wrote that I wrote the code by hand. Maybe my coding level isn't quite there yet for this project, sorry. I'd understand if you rejected it.

I'll slim it down and add tests.

MikeMcQuaid · 2026-06-08T12:44:10Z

Thanks for clarifying and opening the PR @EdenRochmanSharabi!

Inline build_reverse_dep_map, analyze_formula_footprint, print_footprint_table, and print_footprint_single into print_sizes_with_deps since each was only called once. Use Array() instead of is_a?(Array) guard for runtime_dependencies. Add four non-integration tests covering no-deps, exclusive deps, shared deps, and table output formats.

MikeMcQuaid · 2026-06-09T18:19:50Z

+          keg = installed_formula.any_installed_keg
+          next unless keg
+
+          Array(keg.runtime_dependencies).each do |dep|


Why not use installed_formula.runtime_dependencies?

Done, switched to formula.runtime_dependencies in both loops.

MikeMcQuaid · 2026-06-09T18:20:37Z

+            next unless full_name
+
+            dep_name = Utils.name_from_full_name(full_name)
+            (reverse_map[dep_name] ||= Set.new).add(installed_formula.name)


Suggested change

(reverse_map[dep_name] ||= Set.new).add(installed_formula.name)

reverse_map[dep_name] ||= Set.new

reverse_map[dep_name] << installed_formula.name

done, thanks!

MikeMcQuaid · 2026-06-09T18:20:48Z

+          exclusive_deps = T.let([], T::Array[T::Hash[Symbol, T.untyped]])
+          shared_deps = T.let([], T::Array[T::Hash[Symbol, T.untyped]])
+
+          Array(keg&.runtime_dependencies).each do |dep|


MikeMcQuaid · 2026-06-09T18:21:00Z

+          next unless keg
+
+          Array(keg.runtime_dependencies).each do |dep|
+            next unless dep.is_a?(Hash)


When is it not?

Now that I'm using formula.runtime_dependencies, the FormulaUnavailableError rescue shouldn't be necessary because the dependencies are resolved from the formula DSL. Should I remove it entirely and just keep the installed_kegs.empty? guard?

yes unless you can reproduce the error locally

MikeMcQuaid · 2026-06-09T18:21:28Z

+              exclusive_deps << { name: dep_name, size: dep_size }
+            else
+              also_needed_by = (dependents.to_a - [formula.name]).sort
+              shared_deps << { name: dep_name, size: dep_size, also_needed_by: }


T::Struct might be nicer than a hash here

Ok, done, thanks!

@EdenRochmanSharabi you don't seem to have pushed yet?

I haven't pushed yet. I'm not sure about the FormulaUnavailableError guard. Should I remove it and just keep the installed_kegs.empty? check, or keep it as a safety net?

remove for now. aim to simplify and minimise the amount of code as much as possible.

…r hashes, remove FormulaUnavailableError guard, simplify reverse map building

MikeMcQuaid

Hey @EdenRochmanSharabi. Thanks for the work so far. This isn't really making sense to me so far and even --sizes on testing doesn't make much sense either:

brew info ffmpeg-full --sizes or brew info ffmpeg-full --deps doesn't seem to work when ffmpeg-full isn't installed
brew info ffmpeg-full already displays the size when ffmpeg-full is installed
brew info ffmpeg-full lists the various dependencies

If anything I'm inclined to say we don't add this, we odeprecate --sizes (as it's broken enough to be pretty useless) and, if it doesn't negatively affect performance, we instead add size information into the dependencies section of brew info.

Alternatively, just fix --sizes to do the same brew info output as everything else but fetch/include/use size information for the package and all the deps.

MikeMcQuaid · 2026-06-19T15:18:10Z

+        switch "--deps",
+               depends_on:  "--sizes",
+               description: "Show dependency size breakdown with exclusive deps and total disk footprint."


Suggested change

switch "--deps",

depends_on: "--sizes",

description: "Show dependency size breakdown with exclusive deps and total disk footprint."

switch "--dependencies", "--deps",

depends_on: "--sizes",

description: "Show dependency size breakdown with exclusive dependencies and total disk footprint."

MikeMcQuaid · 2026-06-19T15:18:22Z

+            formulae = if args.no_named?
+              Formula.installed
+            else
+              args.named.to_formulae


This fails if passed any casks

MikeMcQuaid reviewed Jun 8, 2026

View reviewed changes

Comment thread Library/Homebrew/cmd/info.rb Outdated

Comment thread Library/Homebrew/cmd/info.rb Outdated

Comment thread Library/Homebrew/cmd/info.rb Outdated

Comment thread Library/Homebrew/cmd/info.rb Outdated

Comment thread Library/Homebrew/cmd/info.rb Outdated

EdenRochmanSharabi and others added 2 commits June 8, 2026 15:25

Merge branch 'main' into info-sizes-deps

0410aff

MikeMcQuaid reviewed Jun 9, 2026

View reviewed changes

EdenRochmanSharabi and others added 2 commits June 18, 2026 11:35

Merging the feedback: use Formula#runtime_dependencies, T::Struct ove…

d39ee13

…r hashes, remove FormulaUnavailableError guard, simplify reverse map building

Merge branch 'main' into info-sizes-deps

e7d32bc

MikeMcQuaid requested changes Jun 19, 2026

View reviewed changes

	(reverse_map[dep_name] \|\|= Set.new).add(installed_formula.name)
	reverse_map[dep_name] \|\|= Set.new
	reverse_map[dep_name] << installed_formula.name

Uh oh!

Conversation

EdenRochmanSharabi commented Jun 8, 2026

Summary

Notes

Uh oh!

MikeMcQuaid left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

EdenRochmanSharabi commented Jun 8, 2026

Uh oh!

MikeMcQuaid commented Jun 8, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

MikeMcQuaid left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants