Basic SQL script support

[MMagueta]

Context

Most of the tools for interacting with Postgres on Emacs usually do not format the tables properly and lack that interactive feel of a Lisp Machine. I think this project is the perfect candidate for that!

Scope

Here I try to implement a script interface where you can either open or type your own SQL queries in a buffer, and eval specifically the ones you selected, without copying and pasting on the minibuffer, or having it all by itself on a buffer via e and E.

How it is implemented

I made a new mode called pgmacs-script-mode that is a minor mode of a buffer with its major mode as sql-mode. Then I made an action called pgmacs--script that passes over the local variables to the script buffer.

The main limitation I found to open arbitrary files would be passing over the local variables. Using (sql-mode) instead of (pgmacs-script-mode) in pgmacs--script simply overrides the connection, so it seems like I need an extension of pgmacs-mode. I also wanted to experiment with lsp-mode and how it behaves using the same connection.

Notes

• Please do provide criticism if the feature is desirable and if we should redesign the idea. I would be very eager to have it properly implemented;
• I still have not finished documenting as I don't know if the idea is sound.

[MMagueta]

@emarsden could you please share your input here? I would like to know if you want this feature in your package, and if so, what would be the way forward with this?

Also, great work so far. I am very glad you started this effort

[emarsden]

This looks like a useful feature, thanks.

I think you could implement pgmacs-run-sql-region and perhaps pgmacs-run-buffer-sql to also work in a random buffer (opened with find-file). Add a global variable pgmacs--connections to which the pgmacs--con buffer-local variable is added whenever a new PGmacs connection is created. In pgmacs-run-sql-region & friends, ask the user which connection to use if the list pgmacs--connections contains more than a single element.

Probably the underlying pgmacs-run-sql should behave in a nicer manner for SQL that returns an empty row set.

[MMagueta]

I haven't had much time to play around with this. I'm probably doing it on the weekend