qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH RFC] pyembed: integer python into QEMU


From: Stefan Weil
Subject: Re: [Qemu-devel] [PATCH RFC] pyembed: integer python into QEMU
Date: Mon, 16 Jan 2012 20:45:48 +0100
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.24) Gecko/20111108 Thunderbird/3.1.16

Am 16.01.2012 20:23, schrieb Anthony Liguori:
On 01/16/2012 12:35 PM, Alex Bradbury wrote:
On 16 January 2012 18:22, Anthony Liguori<address@hidden>  wrote:
This is something I started during 1.0-rc on a lark and spent some time last night actually making work. I'm sending it only to show that (1) it's possible and (2) to get some input about what other people think as a longer term
direction.

At the risk of starting a language flame-war, have you considered
embedding something like Lua?The source of the core Lua interpreter is
small enough that it could be imported into the Qemu repository.

Generally, I don't think importing external source code is a Good Idea. Part of the appeal to me about Python is the rich library that it brings in.

I'm less interested in Python for it's list comprehension syntax and more interested in it for it's config parsing library, RPC infrastructure, etc.

Plus, I'm reasonably confident that most QEMU developers have some experience with Python. I'd wager that very few people have any practical experience with LUA.

So in terms of embedding, I think Python is the only reasonable path forward (if we think we should even head in this direction).


5 years ago, I experimented with an embedded Ruby interpreter. Although it worked (at least to some degree), that's not a reasonable path. Standalone Ruby is a
good interpreter language, but embedding a Ruby interpreter is horrible.

Embedding TCL works well, but TCL is a little old fashioned, so that's also not a
reasonable path.

If we need an embedded interpreter, Python is a good choice, but Lua is also an excellent
embedded interpreter which would be useful for QEMU.

I'd prefer a competition with both of them instead of a single "only reasonable path". The overhead of supporting two interpreters won't be very high, and it will allow more
people to contribute scripts in their favorite language.

Cheers,
Stefan Weil




reply via email to

[Prev in Thread] Current Thread [Next in Thread]