An Emacs 25 module for accessing PostgreSQL via the libpq client library.

Using libpq for client connections has various advantages over the wire-protocol speaking pure elisp implementations. For example, it has better performance and supports all features of the protocol like full TLS support and new authentication methods like scram-sha-256.

It doesn't expose many libpq features yet, but what's there should be crash-safe no matter what you do in the lisp world. I've been using it for three years now for reading mail through my custom Gnus backend without incidents. If you make it crash, please report.

• Basic usage : ELISP> (setq pq (pq:connectdb "dbname=andreas")) : # : ELISP> (pq:query pq "select version()") : ("PostgreSQL 9.6.7 on x86_64-pc-linux-gnu, compiled by gcc (Debian 6.3.0-18) 6.3.0 20170516, 64-bit") : ELISP> (pq:query pq "create table local_variables(name text, value text)") : nil : ELISP> (dolist (el (buffer-local-variables)) : (pq:query pq "insert into local_variables values ($1, $2)" : (car el) (cdr el))) : nil : ELISP> (pq:query pq "select name, length(value) from local_variables where value ~ 'mode'") : (["major-mode" 24] : ["change-major-mode-hook" 86] : ["hi-lock-mode-major-mode" 24] : ["eldoc-mode-major-mode" 24] : ["font-lock-major-mode" 24] : ["font-lock-mode-major-mode" 24])

• Error Handling =pq= raises SQL errors as error signal =pq:error=. This provides the [[https://www.postgresql.org/docs/current/errcodes-appendix.html][SQLSTATE]] error code in an additional string in the error data list. For example, you can reliably catch unique violations like this:

: (condition-case err (pq:query pq "insert into t values (666)") : (pq:error : (if (string= "23505" (nth 2 err)) : (progn : (message "Caught a unique violation")) : ;; re-throw anything else : (signal (car err) (cdr err)))))

• Conversion of data types from SQL to Emacs Lisp =pq= converts bigints and numerics your queries return to lisp floats because they don't fit into a lisp integer. This looses precision on big values. If you need the full precision, cast them to =text= and use, e.g., =calc-eval= to do arbitrary precision things with them. All other data types are returned as utf-8 strings.

• Conversion of data types from Emacs Lisp to SQL Strings and the query text itself is converted to utf-8 by the module interface. If this conversion fails, the behavior is undefined by the module interface. If you want to send strings that are not valid utf-8, you need to work around this. For example, I'm using code like the following to store raw bytes into a table with a =bytea= column:

: (pq:query con : "insert into t (blob) values (decode($1, 'base64'))" : (base64-encode-string my-random-bytes))

Any non-string parameter to pq:query is turned into an emacs string using =prin1-to-string= first. This works quite well to store arbitrary lisp data and read it back with =read=. All other aspects of =prin1-to-string= apply too. For example, when =print-length= or =print-level= are set to non-nil, these would be applied as well.

• Notifications After a [[https://www.postgresql.org/docs/current/sql-listen.html][=LISTEN=]] statement, PostgreSQL will deliver notifications asynchronously over the connection. Since the emacs-module interface does not allow for asynchronous callbacks, you have to check for these periodically after a =LISTEN= statement by calling =pq:notifies=. Calling it will not cause any traffic on the connection itself.

See the testsuite [[./test.el]] for more implemented features.

[[https://api.travis-ci.org/anse1/emacs-libpq.svg]]