aboutsummaryrefslogtreecommitdiff
path: root/me.jl
blob: 20dc22498e6bdd553407563060e5fa9367c5c317 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
# Handers for execute_request and related messages, which are
# the core of the Jupyter protocol: execution of Julia code and
# returning results.

import Base.Libc: flush_cstdio
import Pkg

# global variable so that display can be done in the correct Msg context
execute_msg = Msg(["julia"], Dict("username"=>"jlkernel", "session"=>uuid4()), Dict())
# global variable tracking the number of bytes written in the current execution
# request
const stdio_bytes = Ref(0)

import REPL: helpmode

# use a global array to accumulate "payloads" for the execute_reply message
const execute_payloads = Dict[]

function execute_request(socket, msg)
    code = msg.content["code"]
    @vprintln("EXECUTING", code)
    global execute_msg = msg
    global n, In, Out, ans
    stdio_bytes[] = 0
    silent = msg.content["silent"]
    store_history = get(msg.content, "store_history", !silent)
    empty!(execute_payloads)

    if !silent
        n += 1
        send_ipython(publish[],
                     msg_pub(msg, "execute_input",
                             Dict("execution_count" => n,
                                          "code" => code)))
    end

    silent = silent || REPL.ends_with_semicolon(code)
    if store_history
        In[n] = code
    end

    # "; ..." cells are interpreted as shell commands for run
    code = replace(code, r"^\s*;.*$" =>
                   m -> string(replace(m, r"^\s*;" => "Base.repl_cmd(`"),
                               "`, stdout)"))


    # "] ..." cells are interpreted as pkg shell commands
    if occursin(r"^\].*$", code)
        code = "IJulia.Pkg.REPLMode.do_cmd(IJulia.minirepl[], \"" *
            escape_string(code[2:end]) * "\"; do_rethrow=true)"
    end

    # a cell beginning with "? ..." is interpreted as a help request
    hcode = replace(code, r"^\s*\?" => "")

    try
        for hook in preexecute_hooks
            invokelatest(hook)
        end


        ans = result = if hcode != code # help request
            Core.eval(Main, helpmode(hcode))
        else
            #run the code!
            occursin(magics_regex, code) && match(magics_regex, code).offset == 1 ? magics_help(code) :
                SOFTSCOPE[] ? softscope_include_string(current_module[], code, "In[$n]") :
                include_string(current_module[], code, "In[$n]")
        end

        if silent
            result = nothing
        elseif result !== nothing
            if store_history
                Out[n] = result
            end
        end

        user_expressions = Dict()
        for (v,ex) in msg.content["user_expressions"]
            try
                value = include_string(current_module[], ex)
                # Like the IPython reference implementation, we return
                # something that looks like a `display_data` but also has a
                # `status` field:
                # https://github.com/ipython/ipython/blob/master/IPython/core/interactiveshell.py#L2609-L2614
                user_expressions[v] = Dict("status" => "ok",
                                           "data" => display_dict(value),
                                           "metadata" => metadata(value))
            catch e
                # The format of user_expressions[v] is like `error` except that
                # it also has a `status` field:
                # https://jupyter-client.readthedocs.io/en/stable/messaging.html#execution-errors
                user_expressions[v] = Dict("status" => "error",
                                           error_content(e, catch_backtrace())...)
            end
        end

        for hook in postexecute_hooks
            invokelatest(hook)
        end

        # flush pending stdio
        flush_all()

        undisplay(result) # dequeue if needed, since we display result in pyout
        invokelatest(display) # flush pending display requests

        if result !== nothing
            result_metadata = invokelatest(metadata, result)
            result_data = invokelatest(display_dict, result)
            send_ipython(publish[],
                         msg_pub(msg, "execute_result",
                                 Dict("execution_count" => n,
                                              "metadata" => result_metadata,
                                              "data" => result_data)))

        end
        send_ipython(requests[],
                     msg_reply(msg, "execute_reply",
                               Dict("status" => "ok",
                                            "payload" => execute_payloads,
                                            "execution_count" => n,
                                            "user_expressions" => user_expressions)))
        empty!(execute_payloads)
    catch e
        bt = catch_backtrace()
        try
            # flush pending stdio
            flush_all()
            for hook in posterror_hooks
                invokelatest(hook)
            end
        catch
        end
        empty!(displayqueue) # discard pending display requests on an error
        content = error_content(e,bt)
        send_ipython(publish[], msg_pub(msg, "error", content))
        content["status"] = "error"
        content["execution_count"] = n
        send_ipython(requests[], msg_reply(msg, "execute_reply", content))
    end
end