Hello everyone,

The silos have thus far been assumed to be flat directories.  I just
made a change that should, in principle, allow them to be treated as
directory trees (i.e. they can have subdirectories).  Below is the
relevant entry from the manual:
https://protesilaos.com/emacs/denote#h:3aeacd4d-8cd2-49b7-bc49-15bbbf4281c0

If anyone here uses silos and has time to test this in a real-world
scenario, please let me know what you think.

    3.4.1. Determine levels of depth in a silo

    [ Part of 1.3.0-dev. ] 

    Silos have a directory-local variable that isolates them from the
    default value of the user option denote-directory. When the silo is a
    flat directory, Denote can easily know about it. Otherwise it needs to
    search the parents of the current directory in order to determine if
    the current file belongs to a silo. 

    To this end, the user can specify a natural number as the value of the
    user option denote-silo-nesting-depth. The default value is 1, meaning
    that Denote will check the current directory and its parent directory
    to determine if either of those paths constitute a silo. If they are,
    Denote will return the given path, else it will use the default
    (global) value of the denote-directory. 

    Maintain separate directory silos for notes: https://protesilaos.com/emacs/denote#h:15719799-a5ff-4e9a-9f10-4ca03ef8f6c5

    Levels of depth are needed to allow silos to have subdirectories.
    Otherwise they are assume to be flat. 

All the best,
Protesilaos (or simply "Prot")

-- 
Protesilaos Stavrou
https://protesilaos.com

Hello Prot,

On 2023-02-15 13:10, Protesilaos Stavrou <info@protesilaos.com> writes:

> The silos have thus far been assumed to be flat directories.  I just
> made a change that should, in principle, allow them to be treated as
> directory trees (i.e. they can have subdirectories).  Below is the
> relevant entry from the manual:
> https://protesilaos.com/emacs/denote#h:3aeacd4d-8cd2-49b7-bc49-15bbbf4281c0

It seems that these changes use `file-name-parent-directory'. This
function is not defined in my emacs (GNU Emacs 28.2). Is this a bug only
for me?

Best,

Alan

Hello again,

On 2023-02-17 09:13, Alan Schmitt <alan.schmitt@petitepomme.net> writes:

> It seems that these changes use `file-name-parent-directory'. This
> function is not defined in my emacs (GNU Emacs 28.2). Is this a bug only
> for me?

After some investigation, it seems that this function was introduced in
emacs 29.1. Maybe you could have denote be dependent on the compat.el
package?

https://elpa.gnu.org/devel/doc/compat.html#Added-Definitions-4

Best,

Alan

> From: Alan Schmitt <alan.schmitt@petitepomme.net>
> Date: Fri, 17 Feb 2023 09:55:15 +0100
>
> Hello again,

Hello Alan,

> On 2023-02-17 09:13, Alan Schmitt <alan.schmitt@petitepomme.net> writes:
>
>> It seems that these changes use `file-name-parent-directory'. This
>> function is not defined in my emacs (GNU Emacs 28.2). Is this a bug only
>> for me?
>
> After some investigation, it seems that this function was introduced in
> emacs 29.1. Maybe you could have denote be dependent on the compat.el
> package?
>
> https://elpa.gnu.org/devel/doc/compat.html#Added-Definitions-4

Indeed, that function is for Emacs 29.  My plan was to add compat as a
requirement, anyway, because I want to make Denote available for Emacs
27.1.  Will do it now and then eventually do the changes necessary for
the port to Emacs 27.1.

All the best,
Prot

-- 
Protesilaos Stavrou
https://protesilaos.com

Hello folks!

I have greatly simplified the way we check for a path to the root of a
silo.  This should, in principle, remove the performance penalty that
you experienced recently.  The user option for a depth level is no more.
This works with a standard function that finds the path of the
.dir-locals.  Please test Denote with the latest commit.  I will post
this message also on the GitHub mirror.

    commit d5cf35ec2e11fb2891f15e405644768105a6a119
    Author: Protesilaos Stavrou <info@protesilaos.com>
    Date:   Mon Feb 20 08:06:15 2023 +0200

        REMOVE recursive search for silo root; keep simpler check

        This was never a good idea but I wanted to see what the implications
        are.  We do not need this anymore due to commit 3d2e1d5 which
        simplifies 'denote--silo-p'.

        Thanks to relict007, Hilde Rhyne, Mirko Hernández, Noboru
        Ota (nobiot), Alan Schmitt, and hapst3r for their participation in the
        relevant discussions:

        - https://lists.sr.ht/~protesilaos/denote/%3C87fsb72nge.fsf%40protesilaos.com%3E
        - https://lists.sr.ht/~protesilaos/denote/%3C80CBB671-D812-4EA8-8C80-85F9F4144051%40disroot.org%3E
        - https://lists.sr.ht/~protesilaos/denote/%3C87pma6t59i.fsf%40kotlak.com%3E
        - https://github.com/protesilaos/denote/issues/129 (GitHub mirror)

     README.org | 26 --------------------------
     denote.el  | 41 +----------------------------------------
     2 files changed, 1 insertion(+), 66 deletions(-)

All the best,
Prot

-- 
Protesilaos Stavrou
https://protesilaos.com