Re: Embedded list selection with ido-completing-read.

[Hongyi Zhao]

On Sun, Oct 24, 2021 at 10:28 AM Hongyi Zhao <hongyi.zhao@gmail.com> wrote:
> BTW, my ultimate purpose it to implement the in-buffer company-mode
> [1] completion support for QE-modes [2], which is the main reason why
> I ask this question. And I’ve got some useful advice here [3-5].
>
> [1] https://github.com/company-mode/company-mode
> [2] http://pwtk.ijs.si/qe-modes.html
> [3] 
> https://github.com/company-mode/company-mode/discussions/1247#discussioncomment-1525332
> [4] 
> https://github.com/company-mode/company-mode/blob/master/CONTRIBUTING.md#guidelines-for-third-party-packages
> [5] 
> https://github.com/company-mode/company-mode/blob/master/CONTRIBUTING.md#backend-integration
>
> It seems that I still have a long way to go to achieve this goal.

I describe the outline of my final goal as follows [1]:

The ultimate purpose of [`QE-modes`](http://pwtk.ijs.si/qe-modes.html)
package is to facilitate the generation of [`Quantum
ESPRESSO`](http://www.quantum-espresso.org/) input files. Therefore,
the content to be automatically completed here may not be the same as
ordinary automatic completion. Let me use [the following
example](https://github.com/QEF/q-e/blob/master/test-suite/cp_h2o/h2o-mt-blyp-1.in)
to illustrate some of the technical difficulties that need to be
resolved.

```
&control
   title = ' Water Molecule ',
   calculation = 'cp',
   restart_mode = 'from_scratch',
   ndr = 51,
   ndw = 51,
   nstep  = 100,
   iprint = 100,
   isave  = 100,
   tstress = .TRUE.,
   tprnfor = .TRUE.,
   dt    = 5.0d0,
   etot_conv_thr = 1.d-9,
   ekin_conv_thr = 1.d-4,
   prefix = 'h2o'
   verbosity = 'medium'
/
&system
   ibrav = 14,
   ce
   celldm(1) = 12.0,
   celldm(2) = 1.0,
   celldm(3) = 1.0,
   celldm(4) = 0.0,
   celldm(5) = 0.0,
   celldm(6) = 0.0,
   nat  = 3,
   ntyp = 2,
   nbnd = 4,
   ecutwfc = 80.0,
/
&electrons
   emass = 400.d0,
   emass_cutoff = 2.5d0,
   electron_dynamics = 'damp',
   electron_damping = 0.2
/
&ions
   orthogonalization = 'ortho',
   ion_dynamics = 'none',
   ion_radius(1) = 0.8d0,
   ion_radius(2) = 0.8d0,
/
ATOMIC_SPECIES
   O 16.0d0 O.blyp-mt.UPF
   H 1.00d0 H.blyp-vbc.UPF
ATOMIC_POSITIONS (bohr)
   O     0.0099    0.0099    0.0000  0 0 0
   H     1.8325   -0.2243   -0.0001  1 1 1
   H    -0.2243    1.8325    0.0002  1 1 1
```
The input file specification of the above example is located
[here](https://www.quantum-espresso.org/Doc/INPUT_CP.html). The main
problems I want to solve are:

1. When I try to input one namelist, hinting all possible namelists at
the point with drop-down frame; once I select the namelist, say,
`&CONTROL`, then insert it corresponding to the specification as
follows, and put the point in the middle of it, denoted by `|`:

```
&control
|
/
```
2. When I try to input one CARD, hinting all possible CARDs at the
point with drop-down frame; once I select the CARD, say,
`ATOMIC_SPECIES`, then insert it corresponding to the specification as
follows, and put the point in the middle of it, denoted by `|`:

```
ATOMIC_SPECIES
|
```
3. When I try to input one variable with the point located in certain
namelist, hinting all possible variables belonging to the given
namelist at the point with drop-down frame. Say, for the namelist
`&control`, only [the following
variables](https://www.quantum-espresso.org/Doc/INPUT_CP.html#idm4)
should be used to construct the candidates list:

`     calculation | title | verbosity | isave | restart_mode | nstep |
iprint | tstress | tprnfor | dt | outdir | saverho | prefix | ndr |
ndw | tabps | max_seconds | etot_conv_thr | forc_conv_thr |
ekin_conv_thr | disk_io | memory | pseudo_dir | tefield `

4. After I have completed an input of one CARD, inserting an example
usage of the CARD, say, the following one:

```
ATOMIC_POSITIONS (bohr)
   O     0.0099    0.0099    0.0000  0 0 0
```


[1] 
https://github.com/company-mode/company-mode/discussions/1247#discussioncomment-1526576

HZ