On 02/05/2013 09:22 AM, Markus Armbruster wrote: > In particular, document the impact of our crappy UTF-8 handling on > reading. > > Now the QMP examples even work. > > Signed-off-by: Markus Armbruster <arm...@redhat.com> > --- > hmp-commands.hx | 20 ++++++++------------ > qapi-schema.json | 37 +++++++++++++++++++++++-------------- > qemu-char.c | 7 +++++++ > qmp-commands.hx | 37 +++++++++++++++++++------------------ > 4 files changed, 57 insertions(+), 44 deletions(-)
> +++ b/hmp-commands.hx
> @@ -844,16 +844,15 @@ ETEXI
> .name = "memchar_write",
> .args_type = "device:s,data:s",
> .params = "device data",
> - .help = "Provide writing interface for CirMemCharDriver. Write"
> - "'data' to it.",
> + .help = "Write to a memory character device",
> .mhandler.cmd = hmp_memchar_write,
> },
>
> STEXI
> @item memchar_write @var{device} @var{data}
> @findex memchar_write
> -Provide writing interface for CirMemCharDriver. Write @var{data}
> -to char device 'memory'.
> +Write @var{data} to memory character device @var{device}. @var{data}
> +must be an UTF-8 string.
s/an UTF-8/a UTF-8/
Blame stupid English rules for making that one tough. The rule of thumb
for leading 'h' and 'u' is to use 'a' when pronounced, 'an' when
relaxed, as in: 'a unicorn' vs. 'an umbrella'; 'a half hour' vs. 'an
hour'. In this case, the acronym is generally spoken you-tee-eff-eight,
so it gets 'a' since the 'u' was pronounced.
> +Read up to @var{size} bytes from memory character device @var{device}.
> +Bug: can screw up when the buffer contains invalid UTF-8 sequences,
> +NUL characters, after the ring buffer lost data, and when reading
> +stops because the size limit is reached.
Fair enough for 1.4 since we are in hard freeze; and hopefully 1.5 makes
life better.
Since the grammar fix is minor, feel free to keep this even if you respin:
Reviewed-by: Eric Blake <ebl...@redhat.com>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
