Added @softscope macro for ease of use

[dawbarton]

The title says it all really. Thanks for making this package.

[stevengj]

LGTM, once the merge conflict is fixed.

[dawbarton]

Great, I'll have a look at it tomorrow. (I might switch around how the escaping is handled since I'm not quite happy with it.)

[stevengj]

Isn't __module__ new in Julia 0.7 (JuliaLang/julia#22064)? SoftScope.jl has to support 0.6, where I guess you would use current_module().

[stevengj]

Needs a test.

needs test and 0.6 compatibility

[dawbarton]

Working on it; consider this a WIP for the moment.

[dawbarton]

OK, all done. Should be v0.6+ compatible and has tests.

[Datseris]

Sorry to invade this PR, but I think it is better to ask here than to open a new issue.
In JuliaLang/julia#28789 it was proposed that the following is possible:

julia> a = 0                                                                
0                                                                           
                                                                            
julia> for i = 1:10                                                         
         a += i                                                             
       end                                                                  
ERROR: UndefVarError: a not defined                                         
Stacktrace:                                                                 
 [1] top-level scope at .\REPL[2]:2 [inlined]                               
 [2] top-level scope at .\none:0                                            
                                                                            
julia> using SoftGlobalScope                                                
[ Info: Precompiling SoftGlobalScope [363c7d7e-a618-11e8-01c4-4f22c151e122] 
                                                                            
julia> for i = 1:10                                                         
         a += i                                                             
       end                                                                  
                                                                            
julia> a                                                                    
55  

i.e. the @softscope macro introduced here is applied everywhere by default. Is this still possible to be done?

[dawbarton]

Hi @Datseris, see #2 for that request. I looked into it further this afternoon and I have a few suggestions for a way forward (I'll write up in the next few days). In the meantime, I've updated the gist at https://gist.github.com/dawbarton/0388715fb56fb5cd05e0e4b12c322815 which implements my first attempt at the REPL hooks. Note the comment in the code if you use Revise! (Don't try putting the code in your startup.jl; it won't work there at the moment.)

[stevengj]

I modified the test to make sure it works with a variable that is defined in the @softscope block. I don't think it will at the moment— this is why softscope_include_string has to evaluate one statement at a time.

[dawbarton]

Two can play at that game! 😄

Try this with softscope_include_string -

softscope_include_string(Main, "if true ; ab = 0; for i = 1:10; ab += i ; end ; end")

I think we'll need to recurse into blocks and record new assignments rather than just trying to execute one block at a time (which will obviously fail here). I've got a stack of exam marking to do by close of play tomorrow so I won't be able to have a look before then but I'll have a go at it after that if you haven't already had a look at it.

[stevengj]

@dawbarton, yes, that occurred to me. It's not high on my priority list at the moment since I'm guessing that defining new global variables in blocks is not so common in interactive code, but we can definitely catch some common cases like x = .... assignments in begin and if blocks.

However, without that feature I suspect that @softscope is much less useful, since @softscope begin ... end is probably the most useful formulation.

(It will be possible to fool softscope with things like @eval, but that's even further down on my list of priorities.)

[dawbarton]

This should now all work correctly with #4 merged.