Re: [Qemu-devel] [PATCH v6 12/16] qapi: Consistent generated code: prefer common labels

[Eric Blake]

On 09/29/2015 07:56 AM, Markus Armbruster wrote:
> Eric Blake <address@hidden> writes:
> 
>> We had some pointless differences in the generated code for visit,
>> command marshalling, and events; unifying them makes it easier for
>> future patches to consolidate to common helper functions.
>>
>> This is patch 3/4, focusing on naming the goto labels 'out' (not
>> 'clean') and 'out_obj' (not 'out_end').  Additionally, the generator
>> was inconsistent on whether labels had a leading space; prefer the
>> emacs style of including the space, so that only function names and
>> comments occur in column 1 of the generated file [HACKING gives no
>> preference, and both styles occur elsewhere in qemu.git].
> 
> Your Emacs prefers this?  Mine doesn't.

I'm using Dan's patch "Add .dir-locals.el file to configure emacs coding
style"; which has been claimed to be on the trivial patch queue
(although I haven't seen a trivial pull request since that date, nor can
I find that queue pushed anywhere public):
https://lists.gnu.org/archive/html/qemu-devel/2015-09/msg04148.html

> 
> If I read CC Mode correctly, only style "gnu" does this.  I use
> "stroustrup" for QEMU, because it's very close.  It doesn't do this.

Dan's patch uses "K&R" rather than "stroustrup", although the two are
similar.  Digging deeper, I see that my buffer when visiting a .c in
qemu sets c-label-minimum-indentation to 1, and attributes it to
set-from-style; but I'm not sure why I'm getting it, as I agree with
your reading of c-style-alist that c-gnu-impose-minimum (which is what
typically sets c-label-minimum-indentation to 1) is only tied to "gnu"
mode.  So I'm lost as to why I'm seeing different behavior.

> 
> As far as I can see, the only label we generate indented before this
> patch is qapi-event.py's clean.
> 
> When in doubt, avoid the churn: we rename that one anyway, so let's drop
> the space there, and leave the others alone.

Sure, I can go that way for less churn, and since I can't figure out why
it's getting set for me.

> 
>> No change in semantics to the generated code.
>>
>> Signed-off-by: Eric Blake <address@hidden>
>>
>> ---
>> v6: split 9/46 into four patches, update docs where they are affected,
>> pick consistent spacing of label
>>
>> Signed-off-by: Eric Blake <address@hidden>

Oh weird - I didn't notice the double s-o-b, but the second one gets
lost during 'git am'.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature