Fix #193, issue with makeform macro in v0.5

[ScottPJones]

The AST that it was inspecting had changed in v0.5 with recent changes, so that it didn't detect the in operators correctly.
I also fixed another issue that caused a warning in v0.5 (readall has been deprecated and renamed readstring)

[ScottPJones]

@timholy I noticed it looks like you have merge privileges on @dcjones' repos. Could you please take a look at this PR, and also the PRs #195 & #196? I'd been trying to get things working again in v0.5, I think a lot of people are using Gadfly and this is part of getting it working again. Thanks!

[aviks]

This looks OK to me. I'll merge this imminently unless there are any objections.

[aviks]

Actually, thinking about this a little more, shouldn't this do different things depending on Julia version? Rather than layering the new behaviour on top of the old? This might cause problems later?

Thanks @ScottPJones for tacking this...

[ScottPJones]

@aviks 😊 Thanks! I don't think this could cause a problem, because I believe the AST for parsing (x in xs) could only be in either the old form or the new form.

[timholy]

Thanks for looking at this @aviks. I am struggling with bandwidth issues (my own, not my ISP) at the moment, and I am really grateful for the help.

[timholy]

(but do let me know if you need me for some reason)

[aviks]

could only be in either the old form or the new form

Yes, indeed, but my worry is that the old form might be re-purposed in the future for some other use. In this particular case, that may be unlikely; but it seems better hygiene to protect this. no?

[ScottPJones]

OK, I'll try to narrow down just which of the many v0.5 commits changed the AST output, but that's going to be a good deal more work.
Would it be at all possible to merge this in, just to get things working again for people on v0.5, with a definite promise from me to submit a PR as soon as I've figured out which commit & version number to check for?
The other two PRs I submitted are needed independent of Julia version, BTW.
Thanks again, both of you!

[ScottPJones]

This indeed is because of JuliaLang/julia#15524, I just need to get the correct version number now to fix this as you requested @aviks, will be ready shortly.

[ScottPJones]

@aviks, all set now, using VERSION, as you asked for. Thanks for reviewing!

[aviks]

Thanks. I restarted the travis build of the last master commit, and that failed tests. This one passes. So looks good.