Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #31131 > unrolled thread

Re: __builtins__ thread-safe / __builtins__ as function?

Back to article view | Back to comp.lang.python

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: __builtins__ thread-safe / __builtins__ as function? Dave Angel <d@davea.name> - 2012-10-11 16:28 -0400

#31131 — Re: __builtins__ thread-safe / __builtins__ as function?

On 10/11/2012 10:16 AM, Juergen Bartholomae wrote:
> Hello.
> I have quite a peculiar problem.
>
> A little overview of our situation:
> Our program enables our users to write their own python code (which they
> use extensively).
> Unfortunately, (due to us actually encouraging this in an earlier release
> (!stupid!)) this meant that, in several cases, there are modules which use
> __builtins__ as a kind of global dictionary, where various variables are
> inserted, changed and read.
>
> Now, currently we are updating our program from a single-threaded to a
> multithreaded architecture.
> -> Of course, this means that the same module can now run in several
> threads at the same time, and, since __builtins__ is pretty much global in
> the system, they all share the same __builtins__.

And that's just the first problem.  Maybe you can just let your new
library revert to the old one unless the calling module specifically
activates the new features.

That way existing modules will continue to work the old way, and if they
want new features, they'll need to follow new constraints.  One of these
is no globals, including nothing added to __builtins__   But remember
that their own globals are just as problematic.


-- 

DaveA

[toc] | [standalone]

Back to top | Article view | comp.lang.python

csiph-web