unofficial mirror of emacs-devel@gnu.org 
 help / color / mirror / code / Atom feed
From: joaotavora@gmail.com (João Távora)
To: Mark Oteiza <mvoteiza@udel.edu>
Cc: jwiegley@gmail.com, emacs-devel@gnu.org,
	"Simen Heggestøyl" <simenheg@gmail.com>,
	dgutov@yandex.ru, "Steve Purcell" <steve@sanityinc.com>,
	sdl.web@gmail.com, monnier@iro.umontreal.ca
Subject: Re: Flymake refactored
Date: Thu, 05 Oct 2017 11:57:32 +0100	[thread overview]
Message-ID: <87shex276r.fsf@gmail.com> (raw)
In-Reply-To: <87mv566yjx.fsf@udel.edu> (Mark Oteiza's message of "Wed, 04 Oct 2017 23:52:50 -0400")

Mark Oteiza <mvoteiza@udel.edu> writes:

> Modeled after your example and bits from flycheck[0] and syntastic[1], I
> have attached humble beginnings to a clang checker (errors only).

Looks great.

> Something is probably very broken, as using it to check Emacs C at this
> point triggers the following:
>
>   funcall-interactively: binding stack not balanced (serious byte compiler bug)

Sounds serious indeed. I've seen this too, once or twice, no idea how to
debug it.

> I also seem to have problems using the column number (match-string 2)
> and so it is left unused.

The problem is that you are calling flymake-diag-region in the wrong
buffer. It has to be the source buffer, so in your case you need a
(with-current-buffer source ...) around it.

But that inconvenience has been fixed in the very latest code emacs-26
branch. In that version, you just pass 'source' (the c/c++ buffer) to
flymake-diag-region.

>              :sentinel
>              (lambda (p _ev)

One of the things you must do here is check if your process is obsolete,
i.e. if Flymake decided to launch another one in the meantime. A good
way to do this is to check if 'p' is 'eq' to the buffer-local value of
clang-flymake--procress.

>                       for (beg . end) = (flymake-diag-region
>                                          (string-to-number
>                                          (match-string 1)))

This is the bit where you would pass 'source' to flymake-diag-region

I've built a backend very similar to yours but base on gcc (clang is
500MB and no time for that right now). Have a look, below my sig.

I've also noticed, there's a lot of repetition building up in these
examples. Later it's probably useful to invent an abstraction that hides
it away.

João

;;; test-gcc-backend.el --- naive gcc Flymake backend -*- lexical-binding: t; -*-

(defvar-local gcc--flymake-proc nil)

(defvar gcc-program "gcc")

(defun gcc-flymake (report-fn &rest _args)
  (unless (executable-find gcc-program)
    (error "Cannot find a suitable gcc"))
  (when (process-live-p gcc--flymake-proc)
    (kill-process gcc--flymake-proc))
  (let ((source (current-buffer)))
    (save-restriction
      (widen)
      (setq gcc--flymake-proc
            (make-process
             :name "gcc-flymake"
             :buffer (generate-new-buffer "*gcc-flymake*")
             :command `(,gcc-program
                        "-fsyntax-only" "-Wextra" "-Wall"
                        "-fno-diagnostics-show-option"
                        "-x" "c" "-")
             :noquery t :connection-type 'pipe
             :sentinel
             (lambda (p _ev)
               (unwind-protect
                   (when (eq p gcc--flymake-proc)
                     (with-current-buffer (process-buffer p)
                       (goto-char (point-min))
                       (cl-loop
                        while (search-forward-regexp
		               "^<stdin>:\\([0-9]+\\):\\([0-9]+\\): \\(.*\\): \\(.*\\)$"
		               nil t)
                        for msg = (match-string 4)
                        for (beg . end) = (flymake-diag-region source
                                                               (string-to-number (match-string 1))
                                                               (string-to-number (match-string 2)))
                        for type = (assoc-default (match-string 3)
                                                  '(("error" . :error)
                                                    ("note" . :note)
                                                    ("warning" . :warning))
                                                  #'string-match)
                        collect (flymake-make-diagnostic source beg end type msg)
                        into diags
                        finally (funcall report-fn diags))))
                 (display-buffer (process-buffer p)) ; use this instead of the next one for debug
                 ;; (kill-buffer (process-buffer p))
                 )
               )))
      (process-send-region gcc--flymake-proc (point-min) (point-max))
      (process-send-eof gcc--flymake-proc))))







  reply	other threads:[~2017-10-05 10:57 UTC|newest]

Thread overview: 79+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-28 14:27 Flymake refactored João Távora
2017-09-28 19:52 ` Stefan Monnier
2017-09-29  0:22   ` João Távora
2017-09-29  3:11     ` Stefan Monnier
2017-10-01 16:52       ` João Távora
2017-10-01 20:50         ` Stefan Monnier
2017-10-02  1:01           ` João Távora
2017-10-02  3:12             ` Stefan Monnier
2017-10-03  0:33               ` João Távora
2017-10-03  1:09                 ` Stefan Monnier
2017-09-29 12:51   ` Dmitry Gutov
2017-09-29 14:55     ` Ted Zlatanov
2017-09-29 15:03       ` Dmitry Gutov
2017-09-29 16:26         ` Ted Zlatanov
2017-09-29 17:35           ` Dmitry Gutov
2017-09-29 17:56             ` Ted Zlatanov
2017-09-30 15:07               ` Dmitry Gutov
2017-09-30  7:55 ` Marcin Borkowski
2017-09-30 23:43   ` João Távora
2017-10-01  8:53     ` Marcin Borkowski
2017-10-01 11:54       ` Mark Oteiza
2017-10-04 17:37 ` Simen Heggestøyl
2017-10-05  2:08   ` João Távora
2017-10-05  3:52     ` Mark Oteiza
2017-10-05 10:57       ` João Távora [this message]
2017-10-05 13:11         ` Stefan Monnier
2017-10-05 14:45           ` João Távora
2017-10-05 23:01             ` João Távora
2017-10-05 21:22         ` Mark Oteiza
2017-10-05 23:05           ` João Távora
2017-10-06  3:35             ` Stefan Monnier
2017-10-06  7:09               ` Lele Gaifax
2017-10-06  8:14                 ` Eli Zaretskii
2017-10-06  8:19                   ` Lele Gaifax
2017-10-06  9:48                     ` Eli Zaretskii
2017-10-06  9:54                       ` Lele Gaifax
2017-10-06 13:04                 ` Mark Oteiza
2017-10-06 14:47                   ` Lele Gaifax
2017-10-06 15:21                     ` Mark Oteiza
2017-10-06 15:26                       ` Mark Oteiza
2017-10-06 15:28                       ` Lele Gaifax
2017-10-06 16:28                         ` João Távora
2017-10-06 19:24                           ` Lele Gaifax
2017-10-06 15:13               ` João Távora
2017-10-07 13:28                 ` Stefan Monnier
2017-10-07 13:44                   ` Eli Zaretskii
2017-10-07 14:40                     ` Lele Gaifax
2017-10-07 14:52                       ` Eli Zaretskii
2017-10-08  2:06                       ` Stefan Monnier
2017-10-08  9:32                         ` João Távora
2017-10-08 11:24                           ` Lele Gaifax
2017-10-08 14:17                           ` Stefan Monnier
2017-10-08 23:33                             ` João Távora
2017-10-09  3:01                               ` Stefan Monnier
2017-10-09 10:19                                 ` João Távora
2017-10-09 15:50                                   ` [SUSPECTED SPAM] " Stefan Monnier
2017-10-09 16:33                                   ` [PATCH] " Lele Gaifax
2017-10-07  6:31               ` Marcin Borkowski
2017-10-07 13:37                 ` Stefan Monnier
2017-10-07 16:48                   ` Marcin Borkowski
2017-10-06 12:54           ` John Wiegley
2017-10-06 15:17             ` Mark Oteiza
2017-10-06 16:04               ` João Távora
2017-10-06 21:22                 ` Mark Oteiza
2017-10-06 22:03                   ` João Távora
2017-10-07 13:31               ` Stefan Monnier
2017-10-07 16:02                 ` João Távora
2017-10-07 16:07               ` João Távora
2017-10-07 18:18                 ` Mark Oteiza
2017-10-08  9:06                   ` João Távora
2017-10-08 12:51                     ` Mark Oteiza
2017-10-08 23:21                       ` João Távora
2017-10-10 14:27                         ` Mark Oteiza
2017-10-10 15:20                           ` João Távora
2017-10-10 16:10                             ` Mark Oteiza
2017-10-05 11:28     ` Lele Gaifax
2017-10-05 15:12       ` Lele Gaifax
2017-10-10 10:40 ` Lele Gaifax
2017-10-10 12:27   ` João Távora

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://www.gnu.org/software/emacs/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87shex276r.fsf@gmail.com \
    --to=joaotavora@gmail.com \
    --cc=dgutov@yandex.ru \
    --cc=emacs-devel@gnu.org \
    --cc=jwiegley@gmail.com \
    --cc=monnier@iro.umontreal.ca \
    --cc=mvoteiza@udel.edu \
    --cc=sdl.web@gmail.com \
    --cc=simenheg@gmail.com \
    --cc=steve@sanityinc.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).