use the new signature of invoke

[yuyichao]

Replace all invoke(f, (...), ...) with invoke(f, Tuple{...}, ...)

If the CI fails, it's probably either a typo or a bug.

[tkelman]

we should probably have at least one or two tests of the fallback syntax for the sake of packages

[yuyichao]

@tkelman Done.

[tkelman]

from 32 bit build on travis:

exception on 8: ERROR: LoadError: TypeError: subtype: expected Type{T}, got Ptr{Any}
 in issubstate at ./inference.jl:1138
 in schanged at ./inference.jl:1149
 in stchanged at ./inference.jl:1171
 in typeinf_uncached at ./inference.jl:1628
 in typeinf at ./inference.jl:1304
 in typeinf at ./inference.jl:1254
 in abstract_call_gf at ./inference.jl:693
 in abstract_call at ./inference.jl:820
 in abstract_eval_call at ./inference.jl:901
 in abstract_eval at ./inference.jl:928
 in typeinf_uncached at ./inference.jl:1588
 in typeinf at ./inference.jl:1304
 in typeinf_ext at ./inference.jl:1248
 in show_list at ./show.jl:354
 in show_unquoted at ./show.jl:557
 in print at show.jl:241
 in print_to_string at ./string.jl:24
 in string at string.jl:33
 in default_handler at test.jl:25
 in do_test at test.jl:50
 in runtests at /tmp/julia/share/julia/test/testdefs.jl:77
 in anonymous at multi.jl:836
 in run_work_thunk at multi.jl:587
 in anonymous at task.jl:836
while loading ccall.jl, in expression starting on line 26
    From worker 3:       * path                 in   1.38 seconds
fatal error on 8: ERROR: cannot serialize a pointer
 in serialize at serialize.jl:65
 in serialize at serialize.jl:334 (repeats 2 times)
 in send_msg_ at ./multi.jl:174
 in send_msg_now at multi.jl:144
 in send_msg_now at multi.jl:75
 in deliver_result at multi.jl:760
 in anonymous at task.jl:837
Worker 8 terminated.
    From worker 8:       * ccall               ERROR (unhandled task failure): EOFError: read end of file
 in yieldto at task.jl:19
 in wait at ./task.jl:307
 in wait at ./task.jl:223
 in wait_full at ./multi.jl:570
 in remotecall_fetch at multi.jl:670
 in remotecall_fetch at multi.jl:675
 in anonymous at task.jl:1386

oh my... no idea whether it's related to this PR or pre-existing on master though

[yuyichao]

Ping

The failure seems to be disapeared after rebasing.......

Would be nice to have this so that the new signature can get more testing....

[tkelman]

This was the only one of your open PR's I felt remotely qualified to hit merge on. There are some really nasty bugs on 32 bit right now but I'm pretty sure this change can't make them any worse.