Bug: Caching org-publish-find-date [8.2.10 (release_8.2.10 @ /opt/emacs/lisp/org/)]

Bug: Caching org-publish-find-date [8.2.10 (release_8.2.10 @ /opt/emacs/lisp/org/)]

[Lorenzo Bolla]

[-- Attachment #1: Type: text/plain, Size: 1780 bytes --]

Remember to cover the basics, that is, what you expected to happen and
what in fact did happen.  You don't know how to make a good report?  See

     http://orgmode.org/manual/Feedback.html#Feedback

Your bug report will be posted to the Org-mode mailing list.
------------------------------------------------------------------------

When using `org-publish-project` I noticed that generating a sitemap
sorted (anti-)chronologically is very slow. It turns out that the
slowness is due to the sorting of sitemap entries, which calls
`org-publish-find-date` during the comparison. But
`org-publish-find-date` is not cached therefore it is called over and
over for each file during the sorting process.
By contrast, `org-publish-find-title` is cached and sorting
alphabetically is faster.

To test this assumption, I've modified `ox-publish.el` to cache
   `org-publish-find-date`, too, obtaining a substantial speed-up:

   (defun org-publish-find-date (file)
     (or
      (org-publish-cache-get-file-property file :date nil t)
      (let ((date (org-publish-find-date-uncached file)))
        (org-publish-cache-set-file-property file :date date)
        date)))

   (defun org-publish-find-date-uncached (file)
     "Find the date of FILE in project.
   This function assumes FILE is either a directory or an Org file.
   If FILE is an Org file and provides a DATE keyword use it.  In
   any other case use the file system's modification time.  Return
   time in `current-time' format."
   ...

Is there a reason why we should not cache the date of a file, in the
same way as we cache its title?

Thanks,
Lorenzo

Emacs  : GNU Emacs 25.1.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.24.23)
 of 2016-11-01
Package: Org-mode version 8.2.10 (release_8.2.10 @ /opt/emacs/lisp/org/)

[-- Attachment #2: Type: text/html, Size: 2101 bytes --]

Re: Bug: Caching org-publish-find-date [8.2.10 (release_8.2.10 @ /opt/emacs/lisp/org/)]

[Nicolas Goaziou]

Hello,

Lorenzo Bolla <lbolla@gmail.com> writes:

> When using `org-publish-project` I noticed that generating a sitemap
> sorted (anti-)chronologically is very slow. It turns out that the
> slowness is due to the sorting of sitemap entries, which calls
> `org-publish-find-date` during the comparison. But
> `org-publish-find-date` is not cached therefore it is called over and
> over for each file during the sorting process.
> By contrast, `org-publish-find-title` is cached and sorting
> alphabetically is faster.
>
> To test this assumption, I've modified `ox-publish.el` to cache
>    `org-publish-find-date`, too, obtaining a substantial speed-up:
>
>    (defun org-publish-find-date (file)
>      (or
>       (org-publish-cache-get-file-property file :date nil t)
>       (let ((date (org-publish-find-date-uncached file)))
>         (org-publish-cache-set-file-property file :date date)
>         date)))
>
>    (defun org-publish-find-date-uncached (file)
>      "Find the date of FILE in project.
>    This function assumes FILE is either a directory or an Org file.
>    If FILE is an Org file and provides a DATE keyword use it.  In
>    any other case use the file system's modification time.  Return
>    time in `current-time' format."
>    ...
>
> Is there a reason why we should not cache the date of a file, in the
> same way as we cache its title?

That's post-mature optimization. When only cache results of a function
when it is reported as horribly slow.

I added a cache for `org-publish-find-date' in master branch. Thank you.

Regards,

-- 
Nicolas Goaziou