iocmd.tf-22.2 leaks

$ make valgrind TESTFLAGS="-file ioCmd.test -match iocmd.tf-22.2"
...
==15293== 235,101 (40 direct, 235,061 indirect) bytes in 1 blocks are definitely lost in loss record 826 of 826
==15293== at 0x4A05E1C: malloc (vg_replace_malloc.c:195)
==15293== by 0x4BC08E: TclpAlloc (tclAlloc.c:705)
==15293== by 0x4CD692: Tcl_Alloc (tclCkalloc.c:1046)
==15293== by 0x55F9BA: AllocStringEntry (tclHash.c:836)
==15293== by 0x55F15C: CreateHashEntry (tclHash.c:361)
==15293== by 0x44CBED: Tcl_CreateNamespace (tclNamesp.c:778)
==15293== by 0x44EA42: TclGetNamespaceForQualName (tclNamesp.c:2274)
==15293== by 0x44EC1E: Tcl_FindNamespace (tclNamesp.c:2398)

But then.... add -verbose bpste to the testing....

$ make valgrind TESTFLAGS="-file ioCmd.test -verbose bpste -match iocmd.tf-22.2"
...
==15321== 235,133 (112 direct, 235,021 indirect) bytes in 1 blocks are definitely lost in loss record 825 of 825
==15321== at 0x4A05E1C: malloc (vg_replace_malloc.c:195)
==15321== by 0x4BC08E: TclpAlloc (tclAlloc.c:705)
==15321== by 0x4CD692: Tcl_Alloc (tclCkalloc.c:1046)
==15321== by 0x4BEBDA: Tcl_CreateCommand (tclBasic.c:2080)
==15321== by 0x40FA18: Tcltest_Init (tclTest.c:558)
==15321== by 0x42742F: NewTestThread (tclThreadTest.c:583)
==15321== by 0x544869: NewThreadProc (tclEvent.c:1565)
==15321== by 0x340960673C: start_thread (in /lib64/libpthread-2.5.so)

Something more than a simple leak lurking here it seems.