unofficial mirror of emacs-devel@gnu.org 
 help / color / mirror / code / Atom feed
From: Alan Mackenzie <acm@muc.de>
To: emacs-devel@gnu.org
Subject: Correct byte compiler error/warning positions.  The solution!
Date: Fri, 26 Nov 2021 19:56:21 +0000	[thread overview]
Message-ID: <YaE75SCqaeK4fSLr@ACM> (raw)

Hello, Emacs.

A short while ago, I mused on a solution to the incorrect positions we
get in byte compiler warning/error messages.  My proposed solution
turned out to be impractical fantasy - it didn't and couldn't work,
because of the difficulties of getting macros to play nicely.

However, I've now hit on the answer.

My 2018 solution to the problem _was_ a solution.  It just ran somewhere
around 8% more slowly than the original it was branched from, hence was
deemed unacceptable by other hackers here.

The reason for the slowdown was the modification to the EQ macro in
lisp.h.  It became necessary to check whether "symbols with position"
were enabled whenever the first == comparison didn't match.

A _LOT_ of these comparisons were for NILP, which is just #defined as EQ
(arg, Qnil).  However if NILP were to be defined directly as arg ==
Qnil, it would bypass the overhead in EQ.  Also required here is for the
reader to handle a literal "nil" specially, and return the
non-positioned symbol `nil' for it, rather than a positioned `nil'.

I've just tried out the above on my 2018 proposed code.  On a benchmark,
the new attempt ran 7.3% faster than the 2018 version.  It thus seems
likely that this strategem will give correct positions in warning
messages without such an unacceptable slowdown.

I'm going to try it out on an up to date repository.

-- 
Alan Mackenzie (Nuremberg, Germany).



             reply	other threads:[~2021-11-26 19:56 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-26 19:56 Alan Mackenzie [this message]
2021-11-27  5:53 ` Correct byte compiler error/warning positions. The solution! Eli Zaretskii
2021-11-27  9:31   ` Alan Mackenzie
2021-11-27 10:07     ` Eli Zaretskii
2021-11-27 10:33       ` Alan Mackenzie
2021-11-27 10:51         ` Eli Zaretskii
2021-11-27 23:05           ` Alan Mackenzie
2021-11-28  7:25             ` Eli Zaretskii
2021-11-29 11:50               ` Alan Mackenzie
2021-11-29 12:45                 ` Eli Zaretskii
2021-11-29 19:39                   ` Alan Mackenzie
2021-12-01 15:58                     ` Alan Mackenzie
2021-12-01 16:49                       ` Eli Zaretskii
2021-12-01 16:58                         ` Alan Mackenzie
2021-12-01 17:04                           ` Lars Ingebrigtsen
2021-12-01 17:21                             ` Alan Mackenzie
2021-12-01 17:38                               ` Lars Ingebrigtsen
2021-12-01 20:28                                 ` Alan Mackenzie
2021-12-01 17:08                           ` Eli Zaretskii
2021-12-01 17:12                             ` Alan Mackenzie
2021-12-01 17:53                             ` Andrea Corallo
2021-12-01 17:57                               ` Eli Zaretskii
2021-12-02 11:21                               ` Alan Mackenzie
2021-12-02 16:31                                 ` Andrea Corallo
2021-12-02 20:35                                   ` Alan Mackenzie
2021-12-03 21:05                                     ` Alan Mackenzie
2021-12-04 19:22                                       ` Andrea Corallo
2021-12-04 19:39                                         ` Eli Zaretskii
2021-12-04 19:55                                           ` Andrea Corallo
2021-12-04 19:58                                             ` Eli Zaretskii
2021-12-04 20:06                                               ` Andrea Corallo
2021-12-14 14:29                                         ` Alan Mackenzie
2021-12-15  9:33                                           ` Andrea Corallo
2021-12-17 11:54                                             ` Alan Mackenzie
2021-12-20  8:24                                               ` Andrea Corallo
2021-12-21 17:48                                                 ` Alan Mackenzie
2021-11-29 13:24                 ` Robert Pluim
2021-11-29 19:16                   ` Alan Mackenzie
2021-11-30  9:52                     ` Robert Pluim
2021-11-28 20:15             ` Andrea Corallo
2021-12-01 16:18               ` Andrea Corallo
2021-12-01 16:46                 ` Alan Mackenzie

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=YaE75SCqaeK4fSLr@ACM \
    --to=acm@muc.de \
    --cc=emacs-devel@gnu.org \
    /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).