move REPL to stdlib #25544

KristofferC · 2018-01-13T15:49:45Z

This moved REPL to the stdlib. Since clearly Base needs to access the REPL to start it, I put a Ref where REPL puts itself in when initializing the module.

TODO:

Decide what to do with Base.active_repl and Base.active_repl_backend.
Decide if the "Interact with Julia" manual should be moved to the REPL standard library. Probably.
Should the file replutil.jl be renamed (it is still in Base)? It defines how e.g. errors are shown should not be REPL specific?

Comments?

JeffBezanson · 2018-01-13T18:59:32Z

Yes, I think replutil.jl should stay in Base and be renamed maybe showerror.jl. At the top there are a bunch of general non-error show methods that could be in show.jl or other files.

JeffBezanson · 2018-01-13T19:00:44Z

Hm, we already have arrayshow.jl and methodshow.jl, so I guess make that errorshow.jl.

KristofferC · 2018-01-14T12:14:25Z

Moved the REPL docs to the stdlib location. The part about colors is potentially a bit out of place but not sure if there is a better place to put it.

The last question, should Base.active_repl and Base.active_repl_backend be moved to the REPL module?

JeffBezanson · 2018-01-14T17:03:09Z

I think those can stay in Base; they're Base's way of getting a reference to the current repl, so they're mostly a feature of Base itself.

KristofferC · 2018-01-14T20:22:16Z

But moving atreplinit makes sense? Or is that also Base functionality?

KristofferC · 2018-01-15T13:00:23Z

Thinking about it, I sort of want to move back atreplinit to Base. It is not a function specific to the current REPL implementation. It is just stuff that runs when any REPL is loaded, could be the stdlib one, could be a package REPL. Opinions?

Although, it makes is a bit hard to figure out where to put the documentation for it.

JeffBezanson · 2018-01-15T15:12:43Z

That sounds fine to me.

KristofferC · 2018-01-19T13:59:15Z

Should be good to go assuming CI passes. Freebsd builder seemed to stall during building? @iblis17

iblislin · 2018-01-19T14:14:08Z

I think it's file testsuit again. I'm rebuilding it.

KristofferC · 2018-01-20T11:41:02Z

Rebased, will merge if CI pass because this picks up conflicts very fast.

rfourquet · 2018-01-20T14:49:53Z

base/client.jl

                    active_repl.history_file = history_file
                end
                # Make sure any displays pushed in .juliarc.jl ends up above the
-                # REPLDisplay
-                pushdisplay(REPL.REPLDisplay(active_repl))
+                # REPLREFDisplay


Seems like this one was changed accidentaly?

rfourquet · 2018-01-20T14:54:05Z

base/sysimg.jl

+Base.require(:Distributed)
+Base.require(:Printf)
+Base.require(:Future)
+Base.require(:Libdl)


Those 4 libs are arealy required ?

Rebase error, thanks.

rfourquet · 2018-01-20T15:02:55Z

test/compile.jl

@@ -220,7 +220,8 @@ try
            Dict(s => Base.module_uuid(Base.root_module(s)) for s in
                [:Base64, :CRC32c, :Dates, :DelimitedFiles, :Distributed, :FileWatching,
                 :Future, :IterativeEigensolvers, :Libdl, :LinearAlgebra, :Logging, :Mmap, :Printf,
-                 :Profile, :Random, :Serialization, :SharedArrays, :SparseArrays, :SuiteSparse, :Test, :Unicode]))
+                 :Profile, :Random, :Serialization, :SharedArrays, :SparseArrays, :SuiteSparse, :Test,
+                 :Unicode, :REPL]))


These module names are sorted alphabetically, it may not be so useful, but why not keep the order :)

I wanted to keep the precompile statements as close as possible to where they originally were.

rfourquet · 2018-01-20T15:05:13Z

base/client.jl

@@ -366,7 +373,11 @@ function __atreplinit(repl)
 end
 _atreplinit(repl) = invokelatest(__atreplinit, repl)

+# The REPL stdlib hooks into Base using this Ref
+const REPLREF = Ref{Module}()


I find the name REPLREF a bit confusing, this could suggest that it's a ref to a "REPL object" (not module), what about REPLMOD ? (not a big deal anyway).

I'll change it to REPL_MODULE_REF.

rfourquet · 2018-01-20T15:08:35Z

base/sysimg.jl

+    # PR 25544
+    @deprecate_binding REPL            root_module(:REPL)                 true ", run `using REPL` instead"
+    @deprecate_binding LineEdit        root_module(:REPL).LineEdit        true ", use `REPL.LineEdit` instead"
+    @deprecate_binding REPLCompletions root_module(:REPL).REPLCompletions true ", run `REPL.REPLCompletions` instead"


If REPLCompletions becomes a submodule of REPL, what about renaming it to REPL.Completions? REPL.REPLCompletions looks redundant (REPLCompletions may also be better if used without its prefix...).
Also, using is missing in the message: run `using REPL.REPLCompletions` instead"

rfourquet · 2018-01-20T15:10:43Z

base/sysimg.jl

+Base.require(:Printf)
+Base.require(:Future)
+Base.require(:Libdl)
+Base.require(:REPL)


This could as well be inserted according to the pre-existing alphabetical order

KristofferC · 2018-01-21T19:51:00Z

This PR has passed PR multiple times now (during 3 days) and just been rebased over and over. Last rebase was completely trivial so I'm just going to merge this.

KristofferC added REPL Julia's REPL (Read Eval Print Loop) excision Removal of code from Base or the repository labels Jan 13, 2018

KristofferC force-pushed the kc/repl_stdlib branch 2 times, most recently from 1a992f3 to 177ec3f Compare January 13, 2018 15:52

KristofferC force-pushed the kc/repl_stdlib branch 6 times, most recently from 79ef208 to 30588cf Compare January 14, 2018 11:52

KristofferC force-pushed the kc/repl_stdlib branch from 587a250 to cceda76 Compare January 14, 2018 13:41

KristofferC force-pushed the kc/repl_stdlib branch from cceda76 to de75b19 Compare January 14, 2018 20:24

KristofferC force-pushed the kc/repl_stdlib branch from 7ac86b8 to cd6e93e Compare January 15, 2018 13:32

JeffBezanson added this to the 1.0 milestone Jan 17, 2018

JeffBezanson added the stdlib Julia's standard library label Jan 17, 2018

KristofferC force-pushed the kc/repl_stdlib branch 3 times, most recently from 4bf56e0 to ce916f9 Compare January 19, 2018 12:06

KristofferC changed the title ~~WIP: move REPL to stdlib~~ move REPL to stdlib Jan 19, 2018

KristofferC force-pushed the kc/repl_stdlib branch from ce916f9 to 23f83d5 Compare January 20, 2018 11:40

rfourquet reviewed Jan 20, 2018

View reviewed changes

KristofferC force-pushed the kc/repl_stdlib branch 3 times, most recently from 3bead2d to 78730ba Compare January 21, 2018 15:21

KristofferC added 2 commits January 21, 2018 20:16

move REPL to stdlib

73c989b

move around and rename some show related stuff

d69180c

KristofferC force-pushed the kc/repl_stdlib branch from 78730ba to d69180c Compare January 21, 2018 19:16

KristofferC merged commit 3843259 into master Jan 21, 2018

KristofferC deleted the kc/repl_stdlib branch January 21, 2018 19:51

KristofferC added the needs news A NEWS entry is required for this change label Jan 21, 2018

KristofferC removed the needs news A NEWS entry is required for this change label Nov 13, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

move REPL to stdlib #25544

move REPL to stdlib #25544

KristofferC commented Jan 13, 2018 •

edited

Loading

JeffBezanson commented Jan 13, 2018

JeffBezanson commented Jan 13, 2018

KristofferC commented Jan 14, 2018 •

edited

Loading

JeffBezanson commented Jan 14, 2018

KristofferC commented Jan 14, 2018

KristofferC commented Jan 15, 2018 •

edited

Loading

JeffBezanson commented Jan 15, 2018

KristofferC commented Jan 19, 2018

iblislin commented Jan 19, 2018

KristofferC commented Jan 20, 2018

rfourquet Jan 20, 2018

KristofferC Jan 20, 2018

rfourquet Jan 20, 2018

KristofferC Jan 20, 2018

rfourquet Jan 20, 2018

KristofferC Jan 21, 2018 •

edited

Loading

rfourquet Jan 20, 2018

KristofferC Jan 20, 2018

rfourquet Jan 20, 2018 •

edited

Loading

rfourquet Jan 20, 2018

KristofferC commented Jan 21, 2018

move REPL to stdlib #25544

move REPL to stdlib #25544

Conversation

KristofferC commented Jan 13, 2018 • edited Loading

JeffBezanson commented Jan 13, 2018

JeffBezanson commented Jan 13, 2018

KristofferC commented Jan 14, 2018 • edited Loading

JeffBezanson commented Jan 14, 2018

KristofferC commented Jan 14, 2018

KristofferC commented Jan 15, 2018 • edited Loading

JeffBezanson commented Jan 15, 2018

KristofferC commented Jan 19, 2018

iblislin commented Jan 19, 2018

KristofferC commented Jan 20, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

KristofferC Jan 21, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rfourquet Jan 20, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

KristofferC commented Jan 21, 2018

KristofferC commented Jan 13, 2018 •

edited

Loading

KristofferC commented Jan 14, 2018 •

edited

Loading

KristofferC commented Jan 15, 2018 •

edited

Loading

KristofferC Jan 21, 2018 •

edited

Loading

rfourquet Jan 20, 2018 •

edited

Loading