Variable assignment does not work in mirb

[mruby]

Variable assignment does not work in mirb

[bovi]

Hi pbosetti, yes I know about this issue. At the moment the evaluation is done inside of a proc object. So the scope is only per evaluation but don't least to the next evaluation. I'm wondering if there is a way to execute code in a global context. The code in questions is:

/* evaluate the bytecode */
mrb_return_value = mrb_run(mrb_interpreter,
  /* pass a proc for evaulation */
  mrb_proc_new(mrb_interpreter, mrb_interpreter->irep[byte_code]),
  mrb_top_self(mrb_interpreter));

Any ideas what function in mruby is allowing to execute code in the highest scope?

[pbosetti]

Hi Daniel,
Whoa, the problem looks interesting: mruby traded the global context in order to improve its use à la Lua, and looks like you found the first shortcoming...
I'll try to look at the problem. I think it's an important one, actually, if one plans to use mruby as an embedded interpreter...

Cheers,
Paolo

On May 14, 2012, at 10:53 PM, Daniel Bovensiepen wrote:

Hi pbosetti, yes I know about this issue. At the moment the evaluation is done inside of a proc object. So the scope is only per evaluation but don't least to the next evaluation. I'm wondering if there is a way to execute code in a global context. The code in questions is:

/* evaluate the bytecode */
mrb_return_value = mrb_run(mrb_interpreter,
 /* pass a proc for evaulation */
 mrb_proc_new(mrb_interpreter, mrb_interpreter->irep[byte_code]),
 mrb_top_self(mrb_interpreter));

Any ideas what function in mruby is allowing to execute code in the highest scope?

---

Reply to this email directly or view it on GitHub:
#133 (comment)

[pbosetti]

@bovi: I was looking at the mrb_f_require() in kernel.c: right now (9882fac) it is a dummy method always returning nil. I am not sure about how to implement this method (and looking at the corresponding one in CRuby source is not helping too much either). Nonetheless I am really curious to see how @matz will fill it, because I am pretty sure that this will give us a good example on how to add a bytecode to the top-level scope, don't you think?

[bovi]

@pbosetti actually I would believe mrb_f_eval_m() would be the optimal starting point if it would be available. At least this is how the IRB is usually implemented. But I'm wondering if @matz will actually allow eval inside of the final mruby. It would of course be awesome to have it on-demand and cut it out in case you need a small size mruby version but I guess that might be quite a task to do.

[matz]

I will fix this. I have to prepare a place to hold top-level local variable information.