builder: refactor job runner and use a shared semaphore across build jobs

Switching to a shared semaphore allows multi-build operations (compiler tests, package tests, etc.) to use the expected degree of parallelism efficiently.

While refactoring the job runner, the time complexity was also reduced from O(n^2) to O(n+m) (where n is the number of jobs, and m is the number of dependencies).

main.go

@@ -1237,7 +1237,7 @@ func main() {
 		PrintIR:         *printIR,
 		DumpSSA:         *dumpSSA,
 		VerifyIR:        *verifyIR,
-		Parallelism:     *parallelism,
+		Semaphore:       make(chan struct{}, *parallelism),
 		Debug:           !*nodebug,
 		PrintSizes:      *printSize,
 		PrintStacks:     *printStacks,