test-chg: use a different log to avoid flakyness

The test was deleting the log file to start anew. However a trailing working
process might still be alive at this time, and recreate the very same log on
exit.

We see the trace of such worker in the expected content of server.log (see the
trace modified by this patch). This is flaky because we don't know *when* the
worker will write to the file and there is a race with the `hg init cached`
command.

A much simpler and reliable way to start anew without having such race is… to
write to a different log file. No reuse → no conflict, no conflict → no race, no
race → no flakiness.

tests/test-chg.t

@@ -352,11 +352,10 b' remove foo'
 repository cache
 ----------------
 
-  $ rm log/server.log*
   $ cp $HGRCPATH.unconfigured $HGRCPATH
   $ cat <<'EOF' >> $HGRCPATH
   > [cmdserver]
-  > log = $TESTTMP/log/server.log
+  > log = $TESTTMP/log/server-cached.log
   > max-repo-cache = 1
   > track-log = command, repocache
   > EOF
@@ -420,9 +419,7 b' shut down servers and restore environmen'
 
 check server log:
 
-  $ cat log/server.log | filterlog
-  YYYY/MM/DD HH:MM:SS (PID)> worker process exited (pid=...)
-  YYYY/MM/DD HH:MM:SS (PID)> worker process exited (pid=...) (?)
+  $ cat log/server-cached.log | filterlog
   YYYY/MM/DD HH:MM:SS (PID)> init cached
   YYYY/MM/DD HH:MM:SS (PID)> id -R cached
   YYYY/MM/DD HH:MM:SS (PID)> loaded repo into cache: $TESTTMP/cached (in  ...s)