[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v3 0/5] introduce QArray
From: |
Alex Bennée |
Subject: |
Re: [PATCH v3 0/5] introduce QArray |
Date: |
Tue, 28 Sep 2021 14:37:45 +0100 |
User-agent: |
mu4e 1.7.0; emacs 28.0.50 |
Christian Schoenebeck <qemu_oss@crudebyte.com> writes:
> On Montag, 27. September 2021 12:59:40 CEST Greg Kurz wrote:
>> On Mon, 27 Sep 2021 12:35:16 +0200
>>
>> Christian Schoenebeck <qemu_oss@crudebyte.com> wrote:
>> > On Dienstag, 31. August 2021 14:25:04 CEST Christian Schoenebeck wrote:
>> > > On Dienstag, 31. August 2021 13:58:02 CEST Greg Kurz wrote:
>> > > > On Thu, 26 Aug 2021 14:47:26 +0200
>> > > >
>> > > > Christian Schoenebeck <qemu_oss@crudebyte.com> wrote:
>> > > > > Patches 1 and 2 introduce include/qemu/qarray.h which implements a
>> > > > > deep
>> > > > > auto free mechanism for arrays. See commit log of patch 1 for a
>> > > > > detailed
>> > > > > explanation and motivation for introducing QArray.
>> > > > >
>> > > > > Patches 3..5 are provided (e.g. as example) for 9p being the first
>> > > > > user
>> > > > > of
>> > > > > this new QArray API. These particular patches 3..5 are rebased on my
>> > > > > current 9p queue:
>> > > > > https://github.com/cschoenebeck/qemu/commits/9p.next
>> > > >
>> > > > > which are basically just the following two queued patches:
>> > > > This looks nice indeed but I have the impression the same could be
>> > > > achieved using glib's g_autoptr framework with less code being added
>> > > > to QEMU (at the cost of being less generic maybe).
>> > >
>> > > I haven't seen a way doing this with glib, except of GArray which has
>> > > some
>> > > disadvantages. But who knows, maybe I was missing something.
>> >
>> > Ping
>> >
>> > Let's do this?
>>
>> Hi Christian,
>>
>> Sorry I don't have enough bandwidth to review or to look for an alternate
>> way... :-\ So I suggest you just go forward with this series. Hopefully
>> you can get some reviews from Markus and/or Richard.
>
> Ok, then I wait for few more days, and if there are no repsonses, nor vetos
> then I'll queue these patches anyway.
As far as I can make out the main argument for introducing a QEMU
specific array handler is to avoid needing to use g_array_index() to
reference the elements of the array. This in itself doesn't seem enough
justification to me.
I also see you handle deep frees which I admit is not something I've
really done with GArray as usually I'm using it when I want to have
everything local to each other.
>
> Best regards,
> Christian Schoenebeck
--
Alex Bennée