Add support for sourcing a Vim9 script from a buffer

[yegappan]

No description provided.

[codecov]

Codecov Report

Merging #9974 (fbb2eaf) into master (397a87a) will decrease coverage by 0.17%.
The diff coverage is 66.66%.

❗ Current head fbb2eaf differs from pull request most recent head 0800017. Consider uploading reports for the commit 0800017 to get more accurate results

@@            Coverage Diff             @@
##           master    #9974      +/-   ##
==========================================
- Coverage   81.91%   81.74%   -0.18%     
==========================================
  Files         167      167              
  Lines      187269   187229      -40     
  Branches    42201    42202       +1     
==========================================
- Hits       153409   153044     -365     
- Misses      21516    21813     +297     
- Partials    12344    12372      +28     

Flag	Coverage Δ
huge-clang-none	82.31% <67.64%> (-0.02%)	⬇️
huge-gcc-none	82.65% <66.17%> (-0.02%)	⬇️
huge-gcc-testgui	?
huge-gcc-unittests	2.01% <0.00%> (+<0.01%)	⬆️
linux	83.58% <66.66%> (-0.33%)	⬇️
mingw-x64-HUGE	0.00% <0.00%> (ø)
mingw-x64-HUGE-gui	77.15% <70.17%> (-0.01%)	⬇️
windows	75.93% <68.96%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/scriptfile.c	83.10% <66.17%> (-0.38%)	⬇️
src/ex_docmd.c	91.47% <100.00%> (-0.13%)	⬇️
src/gui_gtk_x11.c	51.59% <0.00%> (-5.39%)	⬇️
src/gui_gtk.c	24.23% <0.00%> (-3.62%)	⬇️
src/sound.c	64.57% <0.00%> (-1.72%)	⬇️
src/hardcopy.c	77.49% <0.00%> (-1.62%)	⬇️
src/gui_xim.c	22.38% <0.00%> (-1.20%)	⬇️
src/os_unix.c	67.35% <0.00%> (-1.08%)	⬇️
src/gui.c	73.16% <0.00%> (-0.88%)	⬇️
src/highlight.c	81.25% <0.00%> (-0.83%)	⬇️
... and 33 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 397a87a...0800017. Read the comment docs.

[habamax]

Can it source the following:

vim9script

def Hello()
   var myList = ['multi',
            'lined',
            'list'
            ]
   echo myList[0]
enddef

nnoremap <F5> <ScriptCmd>Hello()<CR>

• :4,7source
• :11source -- would proper script context be used here?
  ?

[brammool]

We can keep using sourcing_a_script(), hide the details of how that is decided.

I don't think we should try to be clever and detect the kind of script. If someone sources some sequence of buffer lines we can't really know whether they are legacy or Vim9 script. They could even be some lines in a help file. We can require using "vim9cmd source". Unless "vim9script" is the first command in the range of lines.

We also need to define what happens when sourcing from the same buffer again. I suppose this would work like "vim9script noclear". That way you could, for example, source a function again after modifying it.

[vim-ml]

Hi Bram,

On Sun, Mar 20, 2022 at 3:43 AM Bram Moolenaar ***@***.***> wrote: We can keep using sourcing_a_script(), hide the details of how that is decided.

I have updated the PR to restore this function.

I don't think we should try to be clever and detect the kind of script. If someone sources some sequence of buffer lines we can't really know whether they are legacy or Vim9 script. They could even be some lines in a help file. We can require using "vim9cmd source". Unless "vim9script" is the first command in the range of lines.

Currently the code relies on the first sourced line to detect the script type and doesn't try to detect the type of script.

We also need to define what happens when sourcing from the same buffer again. I suppose this would work like "vim9script noclear". That way you could, for example, source a function again after modifying it. Yes. I have added a test for sourcing a Vim9 script with a function

multiple times. - Yegappan

[brammool]

There is a remark in the help about using the same SID when sourcing multiple times.
It would be good to say a few things about what happens when redefining a function. I suppose that works.
But other items from the script are still there, thus it would work like "vim9script noclear".
As is explained at "vim9-reload".

When sourcing just a few lines, e.g. from a help buffer, "vim9cmd source" should work.
Please add a test for that.

[Shane-XB-Qian]

#9967 // let me add a refer/remark to there to show the diff with legacy one.

…

-- shane.xb.qian

[vim-ml]

Hi,

On Sun, Mar 20, 2022 at 1:19 AM Maxim Kim ***@***.***> wrote: Can it source the following: vim9script def Hello() var myList = ['multi', 'lined', 'list' ] echo myList[0] enddef nnoremap <F5> <ScriptCmd>Hello()<CR> :4,7source :11source -- would proper script context be used here? ?

To source a Vim9 script from a buffer, it must start with "vim9script" as the first line. For a legacy script, you can source any range of lines. - Yegappan

[habamax]

To source a Vim9 script from a buffer, it must start with "vim9script" as the first line. For a legacy script, you can source any range of lines.

I have this operator to run both vimscript and vim9script ranges:

# source vimscript (operator)
def SourceVim(...args: list<any>): any
    if len(args) == 0
        &opfunc = matchstr(expand('<stack>'), '[^. ]*\ze[')
        return 'g@'
    endif
    if getline(1) =~ '^vim9script$'
        vim9cmd :'[,']source
    else
        :'[,']source
    endif
    return ''
enddef
nnoremap <silent> <expr> <space>v SourceVim()
xnoremap <silent> <expr> <space>v SourceVim()
nnoremap <silent> <expr> <space>vv SourceVim() .. '_'

So <space>vv to eval current line, <space>vip to eval current para, etc.