build: make custom cargo target dir work completely

[knutties]

Problem

The shared CARGO_TARGET_DIR approach was not possible because Leptos depended on files copied to the target directory relative to project root.

Solution

This PR enables symlinking the target in case the CARGO_TARGET_DIR is set to a different value.

Summary by CodeRabbit

• Chores
  • Refined build infrastructure configuration to improve development environment setup and process orchestration.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[semanticdiff-com]

Review changes with  

[coderabbitai]

Important

Review skipped

Auto incremental reviews are disabled on this repository.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

• 🔍 Trigger a full review

Walkthrough

Modified .gitignore to broaden target ignore rules and restore .zed entry. Introduced symlink-target Makefile target to create cargo target directory symlinks. Refactored build orchestration to route superposition process control through make targets instead of direct binary invocations.

Changes

Cohort / File(s)	Summary
Build Configuration .gitignore	Modified ignore rules: changed "target/" to "target" for broader matching; re-added ".zed" entry.
Build Orchestration makefile	Added symlink-target for cargo directory symlink creation; refactored superposition invocation flows through make targets instead of direct binary calls; added node_modules symlink linking in superposition target; adjusted kill command behavior.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Poem

🐰 A symlink here, a target there,
Make orchestrates with care,
No more binaries dashed about,
Just clean makefiles throughout,
Node_modules linked so tight,
Our build system's feeling right! 🎯

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title directly describes the main change: enabling a custom cargo target directory to work completely through symlinking.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix-custom-target-dir

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[ayushjain17]

this repeats db frontend superposition though

[knutties]

Ideally those should be no-ops right if already built. If not, they should be built right ?

In an ideal case all targets should be intelligent not to repeat if its dependencies did not change - but we do a lot of PHONY targets that some might be repeated and it is left to cargo to deal with it.

[knutties]

Alternatively, we can create a backend_run which does not build frontend.

Once we add FE tests it might become moot though.

[ayushjain17]

this would add frontend building as well

[knutties]

@ayushjain17 - can we address this separately - the dependencies are bit hay-wire. I promise I will refactor and fix the makefile :)

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@makefile`:
- Around line 180-183: The symlink-target Makefile recipe may create a
circular/self-referential symlink when CARGO_TARGET_DIR points to the default
./target; update the symlink-target logic to first resolve and compare the
absolute paths of $(CARGO_TARGET_DIR) and the literal target directory before
creating the link (i.e., get realpath of $(CARGO_TARGET_DIR) and of ./target or
use readlink -f) and only run ln -s when they differ and when target does not
already exist or is not a symlink; reference the symlink-target recipe and the
CARGO_TARGET_DIR variable when making this change.

🧹 Nitpick comments (1)

makefile (1)

234-246: Minor inefficiency: redundant dependency resolution.

The test target already depends on frontend and superposition, but $(MAKE) run & re-evaluates these PHONY targets. While functionally correct (builds are idempotent), this causes redundant work.

Consider extracting a lighter run-only target that skips rebuild deps, or refactoring so run doesn't rebuild what test already built.

[coderabbitai]

Actionable comments posted: 0

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

makefile (1)

186-188: ⚠️ Potential issue | 🟠 Major

Relative path for node_modules symlink may break with custom CARGO_TARGET_DIR.

The relative path ../node_modules assumes CARGO_TARGET_DIR is exactly one directory level below the project root. If CARGO_TARGET_DIR is set to an arbitrary location (e.g., /tmp/cargo-target or a shared cache), the symlink will point to the wrong location.

Consider using an absolute path:

🐛 Proposed fix using absolute path

 superposition: CARGO_FLAGS += --features='$(FEATURES)'
 superposition:
 	cargo build $(CARGO_FLAGS) --bin superposition
-	`@cd` $(CARGO_TARGET_DIR) && ln -s ../node_modules node_modules || true
+	`@cd` $(CARGO_TARGET_DIR) && ln -s $(CURDIR)/node_modules node_modules || true

$(CURDIR) is a built-in Make variable that holds the absolute path of the directory where Make was invoked.

🧹 Nitpick comments (1)

makefile (1)

180-184: Consider adding a guard against creating a self-referential symlink.

If CARGO_TARGET_DIR resolves to the project's own target directory (the default), this would attempt to create a symlink pointing to itself. While cargo metadata returns absolute paths, comparing the resolved paths would prevent unexpected behavior.

🔧 Proposed fix to add a guard

 symlink-target:
-	`@if` [ ! -e target ] && [ ! -L target ]; then \
+	`@if` [ ! -e target ] && [ ! -L target ] && [ "$$(cd $(CARGO_TARGET_DIR) && pwd -P)" != "$$(pwd -P)/target" ]; then \
 		ln -s $(CARGO_TARGET_DIR) target; \
 	fi

Alternatively, a simpler check using realpath if available:

 symlink-target:
-	`@if` [ ! -e target ] && [ ! -L target ]; then \
+	`@if` [ ! -e target ] && [ ! -L target ] && [ "$(CARGO_TARGET_DIR)" != "$$(pwd)/target" ]; then \
 		ln -s $(CARGO_TARGET_DIR) target; \
 	fi