On 11/18/2011 06:32 AM, Stefan Berger wrote:
Use scripts for the renaming and cleaning up of chains. This allows us to get rid of some of the code that is only capable of renaming and removing chains whose names are hardcoded.
+static const char ebtables_script_func_collect_chains[] = + "collect_chains()\n" + "{\n" + " for tmp2 in $*; do\n" + " for tmp in $(%s -t %s -L $tmp2 | \\\n" + " sed -n \"/Bridge chain/,\\$ s/.*-j \\\\([%s]-.*\\\\)/\\\\1/p\");\n"
No need to change this unless you want, but this could be written in fewer bytes as: " sed -n '/Bridge chain/,$ s/.*-j \\([%s]-.*\\)/\\1/p')\n"
+ virBufferAsprintf(buf, NWFILTER_FUNC_SET_IFS); + virBufferAddLit(buf, "a=\"$(collect_chains "); + for (i = 0; chains[i] != 0; i++) { + PRINT_ROOT_CHAIN(rootchain, chains[i], ifname); + virBufferAsprintf(buf, "%s ", rootchain); + } + virBufferAddLit(buf, ")\"\n");
As written, you generate: a="$(collect_chains a b )" with an odd space before ). I also think the name $a is rather terse; it might be better to use something a bit less cryptic. What you have works, but you could also make this tweak if you'd like: virBufferAddLit(buf, "chains=\"$(collect_chains"); for ... virBufferAsprintf(buf, " %s", rootchain); virBufferAddLit(buf, ")\"\n"); ... virBufferAddLit(buf, "rm_chains $chains\n"); -- Eric Blake eblake@redhat.com +1-919-301-3266 Libvirt virtualization library http://libvirt.org