refactor(ixgo): use aggregate package registration

[MeteorsLiu]

No description provided.

[codecov]

Codecov Report

❌ Patch coverage is 72.72727% with 6 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
cmd/llar/internal/install.go	72.72%	4 Missing and 2 partials ⚠️

📢 Thoughts on this report? Let us know!

[fennoai]

FennoAI Review — PR #158: refactor(ixgo): use aggregate package registration

Collapsing ~150 hand-maintained blank imports in internal/ixgo/pkgs.go into the single aggregate _ "github.com/goplus/ixgo/pkg" is a real maintainability win, and I verified the mechanics hold up locally:

• go build ./... → exit 0
• go mod tidy → no changes (go.mod/go.sum are tidy-clean)
• go mod verify → all modules verified
• Keeping net/rpc / net/rpc/jsonrpc separate is correct — the aggregate does not register them (confirmed against goplus/ixgo@v1.1.1/pkg/go124_pkgs.go).
• The testing/* exclusion is preserved — the aggregate does not include testing/* for go1.24.
• The 8 //go:generate qexp directives still map 1:1 to the local generated imports, so they are not stale.

The one substantive concern is a behavioral change hidden inside a "refactor", called out inline below. A few non-line-specific notes:

Implicit testing/* exclusion is now coincidental, not enforced. Previously the excluded packages were visible in the file. Under the aggregate, testing/* stays out only because the current upstream aggregate happens to omit it; a future ixgo bump could silently register it. Consider a small test that pins the expected registered set (asserts net/rpc* present, testing absent) so these invariants survive future dependency bumps.

Binary size / linker weight. The aggregate force-links every registered subpackage (160 for go1.24), including the newly-added debug/* readers (debug/dwarf, debug/elf, debug/macho, …) and their transitive deps. The runtime/startup delta is negligible for a CLI, but binary size will grow. A quick before/after ls -l on the built binary would confirm the increase is acceptable. (os/exec and syscall were already registered before this PR — pre-existing, not introduced here.)

Overall: the refactor is fine to land; please just confirm the debug/* re-inclusion is intended (and ideally reframe as more than a pure refactor, since it changes what interpreted code can reach).