Repin the C library, because a version requirement here is exact

[Sunrisepeak]

Measured against the index after
openkal-musl 0.6.0
was published — a manifest saying openkal-musl = "0.5.0" still resolves to
0.5.0:

kernel-abi        openkal        (openkal-linux@0.6.0, graph)
c-abi             musl           (openkal-musl@0.5.0, graph)

Nothing floats up. That is the property that keeps a release from moving under a
consumer that did not ask for it, and the consequence is that this package does
not receive a fix by having one published beneath it.

What it does not currently receive

0.6.0 answers three defects reported as
openkal-linux#13. Two are
reachable from these tools without any of them doing anything unusual:

• a caller that redirects a standard stream and then starts a program — with
  posix_spawn, with system, or with fork and execve — had the started
  program write to the stream the caller had been started with;
• abort did not end the program. musl's raise is tkill, the port had no
  case for it, so abort fell through to a_crash() — hlt on x86_64 — and an
  assertion failure reported a segmentation fault.

xargs and sh are the tools here that start programs, and the whole suite is
run by a harness that reads what each tool wrote.

Already measured

This repository's CI substitutes openkal-musl's default branch working tree
for the version named here, so it has been building against 0.6.0's content
since that merged. Dispatched after the merge: green, all 97 tools.

What this change adds is that a consumer resolving sbase from the index gets
the same C library the CI has been testing with.

Nothing else changes. 0.1.2 → 0.1.3.