Dear all,
And, finally, the meetup notes.
TL;DR: updates in Org website; memory usage; denote-silo; searching
in Org; alphapapa; accidentally deleting hidden text; fancy Org
bullets; LLMs in Org buffers - gptel; gptel tools, agent-shell;
using embeddings/similarity to auto-insert links to Org notes;
importing Google Docs; custom music/video links; Org on mobile;
support of sentence-at-point in Org; improving org-habit;
Scrim and Captee (Mac); another meetup for bug fixing; shrface.
- Zororg asked whether Org website has been updated
- The clickable folding in the example of Org syntax has been
added years ago, after the latest major re-design, by Timothy
- Coincidently, I did actually change the website a bit
1. I changed names of the top menu
2. Added "Mobile/Web" link after recent discussion on Hacker News
- Zororg complained about high memory usage in Emacs
- He tried profiler (memory), but that profiler does not really
show memory allocations. According to Eli, it is "something else".
AFAIK, that profiler is not practically useful.
- Note that Emacs memory usage is usually high due to images/pdfs.
Image cache holds uncompressed images that tend to occupy huge
space
- M-: (clear-image-cache) clears this cache if one needs testing
- Another common pitfall is that Emacs often allocates large amount
of memory temporarily and malloc may or may not return it to OS.
Memory figures may be misleading. One can try M-x malloc-trim
- A useful diagnostic tool is M-x memory-report (built-in)
It shows memory usage from various objects in Elisp
- Zororg also suspected ~buffer-undo-list~, but ~undo-limit~ is rather
conservative by default, so undo lists are not usually very long.
- There is an actively developed alternative garbage collector for Emacs,
but it actually tends to take more memory, which is the cost of being
non-blocking. That's =feature/igc3= branch. They are really pushing
to make it work for the next Emacs release, although there is no
100% guarantee.
- Some recent discussions
https://yhetil.org/emacs-devel/[email protected]/
- I am also testing this branch right now. It is mostly stable,
although I did experience a few crashes... (one fixed now, another
being fixed)
- For traditional GC, I personally use the following to great effect
#+begin_src emacs-lisp
(setq gc-cons-percentage 0.2)
(setq gc-cons-threshold (* 200 1000 1000))
(add-hook
'after-init-hook
(lambda () (setq gc-cons-threshold (* 20 1000 1000))))
#+end_src
This is not random. Check out my old talk on EmacsConf
https://emacsconf.org/2023/talks/gc/
- George Jones showed his denote setup
- He wants to use separate "silos" for different sets of notes
- There is https://protesilaos.com/emacs/denote-silo, but George
cannot make it work with his setup, so he instead uses a custom
Elisp to set denote-directory (or similar) manually, via prompt
- It was not 100% clear what the problem is, but he showcased his
usage of denote, which was sooth enough to feel like it is a normal
Emacs workflow all us are used to when simply opening files.
- On the todo side, he separates todo tasks and the notes,
keeping all the todos in a separate diary file, linking to
notes as needed for context.
- His config:
https://codeberg.org/eludom/home.public.elisp-public/src/branch/master/elisp-public/gmj-denote-config.el
- start at =(defvar gmj/denote-directories= and go to the bottom
of file
- I suggested writing to Prots about the problem. He is usually
very responsive.
- codaman mentioned that denote excels at searching information
"Denote allows to type '/.' to see the folders in the minibuffer and
know that _tag or --filetitle will immediately find a file."
- This kind of search functionality can be indeed replicated outside denote
- One can, indeed, use ripgrep or similar for full-text search. Of course,
the advantage (or disadvantage, depending on use case) is that denote
limit the search to specific keywords/tags present in the title
- If one needs to search only specific parts of Org files specifically,
https://github.com/alphapapa/org-ql can do that - you can search by
tags, titles, outlines, etc. I personally, define a custom search
predicate searching for tags, outline paths, and "keywords" inside heading,
with keywords being words prefixed with hash: #keyword.
org-ql is very flexible.
https://github.com/yantar92/emacs-config/blob/master/config.org#live-helm-org-ql-search-in-org
- Morgan also asked about my use for sparse trees
- I almost never use them, although they come handy when I need
to have a quick overview of tasks under specific subtree
- The same can be done with org-agenda + subtree restriction,
but sparse tree is simply faster to type
- I sometimes use sparse tree when I am writing a long text
and sparkle inline todos inside. It gives an idea what to work on
- Another use case is to see only, say "WAITING" tasks - sparse
tree filtering by todo keyword comes handy. Although, for this
specific purpose, I have a custom org-sort setup
- Zororg asked about news from alphapapa recently
- I saw him replying to github issues, but he is generally
quite busy. But he is around.
- Zororg asked about accidentally deleting collapsed subtrees
- There is ~org-ctrl-k-protect-subtree~ and ~org-catch-invisible-edits~,
but they cannot help with custom commands, especially for modal
editing where non-standard command are sometimes rolled out.
- There is nothing more reliable than good old git versioning and
backups
- Zororg specifically struggles with activating a region and
<del> removing the whole thing
- This is controlled by ~delete-active-region~ - you can disable it.
[2026-04-05 Sun] Actually, let me steal it to my config.
- In theory, one may make Org buffers read-only, except when running
dedicated white-listed commands, but that's really awkward
- George Jones asked about my "fancy" bullets for headings.
I use the following symbols for bullets:
: ■
: •
: ⊢
: ⋮
: ⋱
: ⋮
: ⋱
: ⋮
: ⋱
- my setup is using pretty-symbols, but there is more specialized
https://github.com/minad/org-modern as well
- As it usually happens recently-ish, we turned to LLM discussions
- We have gptel, which is excelled when you need seamless *text*
integration. For extras, I personally have an elaborate tool
system, including self-invented agent skills (I did not know about
those when I made them for my config)
- gptel supports pretty much all LLMs - local, online, even the
latest Claude Opus 4.6
- Nick Anderson also mentioned agent-shell
https://github.com/xenodium/agent-shell that is equipped with
ready-made Emacs interaction for actual agents like Claude
- codaman asked about using LLM to summarize from all the notes
with specific tag
- I personally do not have such setup, although I have a custom tool
for LLM to search my notes by ID
- There is https://codeberg.org/bajsicki/gptel-got with matching
tools though
- some of the tools there directly utilize org-ql
- Also, check out https://github.com/karthink/gptel/wiki/Tools-collection
- In general, it is very easy to write new gptel tools
- Just normal Elisp, can be a few lines. Example:
#+begin_src emacs-lisp
(defun yant/find-related-files (&optional buffer)
"Return a list of files related to BUFFER."
(with-current-buffer (or buffer (current-buffer))
(or (project-files (project-current))
(buffer-file-name))))
(gptel-make-tool
:name "list_related_files"
:function #'yant/find-related-files
:description "Return a list of files that belong to the
project/repository BUFFER_NAME belongs to."
:args '(( :name "buffer_name"
:type string
:description "Name of the buffer"))
:confirm nil
:category "search")
(push (gptel-get-tool "list_related_files") gptel-tools)
#+end_src
- torusJKL mentioned agent-shell skill to select a region in Emacs file
(to get context)
- https://github.com/xenodium/emacs-skills/blob/main/skills/select/SKILL.md
- Note that gptel has this feature built-in, except that you select
the context manually
- I personally also have a tool for buffer introspection, including
getting specific line range in Emacs buffer
-
https://github.com/yantar92/emacs-config/blob/master/config.org#buffersfiles
- Sacha Chua has been experimenting with linking to bookmarks
- She has a custom version of =org-insert-link= that, when
used on region, automatically looks up a bookmark that
matches the region text. That way, she auto-magically gets
links without having to remember heading ID or URL.
- That reminded me about
https://list.orgmode.org/orgmode/Z_1VNDnVnIhWmv8P@false/
where Jean implemented RAG embeddings to all his notes.
Then, selected region can be matched against embeddings
automatically linking the most similar note. More advanced
version of the same idea - not simple word matching, but
embeddings (mind you - you do not need cloud LLMs for embeddings.
Sentence transformers can run locally on very old computers just
fine)
- Nick Anderson also remembered
https://github.com/brunoarine/org-similarity
- Sacha also mentioned her recent experiments related to interacting
with Google APIs. Strangely, not using Elisp, but in npm.
- I recently played with these things, and I figured two-factor
authentication (via existing Elisp library) + Google Docs API
in pure Elisp. That gotta be much better than npm!
https://sachachua.com/blog/2026/03/emacs-lisp-and-nodejs-getting-the-bolded-words-from-a-section-of-a-google-document/
- That's an Org dynamic block by the way on my side :)
- One of the Sacha's language learning workflow (obviously) leverages
Emacs, Org mode, and custom links. She can play the video/audio
right from inside Emacs and also export them nicely playable in
browser
- codaman asked about playing (Apple) music from inside Emacs
- Sasha simply uses mpv call in a custom follow function
https://sachachua.com/dotemacs/index.html#videos (my-org-video-follow)
- Sacha mentioned some cli tool to play Apple Music. As any cli,
it can be integrated as custom tool for link opening, not different
from mpv.
- There is also built-in =play-sound= is anyone needs it
- And we have plenty of music packages for Emacs, as one might expect
https://github.com/pft/mingus for example (I used it for some time),
or more recent https://github.com/xenodium/ready-player
- codaman asked about using Org mode on mobile
- Morgan suggested Orgro and Orgzly
- We also have many tools listed in https://orgmode.org/tools.html
- (and some Apple-related chat I can't comment on)
- As it usually goes when discussing mobile apps, we turned to
the question of file synchronization
- kickingvegas does it radically: run emacsclient right in his iphone
(using Tailscale/blink.sh) and connecting to Emacs running on his homecloud
- I personally just use syncthing + mainly edit inbox.org from mobile;
other files read-only
- Nick Anderson also uses syncthing + per-devide inbox.org files
- Naturally, people also talked termux (oh, these keyboards)
- ... which made Sacha Chua (who is playing A LOT with speech input
recently) think about doing whisper on the phone (which should be
totally doable, I even saw something on F-Droid)
- Morgan then commented that the latest Android has built-in debian
virtual machine
- I am personally getting more and more pushed away from
Google's version of Android. See
https://keepandroidopen.org/. The direction is clearly wrong
and not going to get better (just as with other Google
services over the last decade).
- Sacha also asked about getting Org supporting setence-at-point
- She used advice, but there is no reason Org should not support
it out of the box. We already do support parts of thingatpt.el
- Sacha later submitted a patch -
https://list.orgmode.org/orgmode/[email protected]/
- Morgan (one of active contributors) shared that he has been working
on improving org-habit state, which is indeed limited for now, with
many ideas that can be implemented.
- For example, there is interest to integrate habits with org-clock
to record progress not just by TODO/DONE or with numeric habits like
steps walked or pages read.
- There was a volunteer recently
https://list.orgmode.org/orgmode/[email protected]/ who
did not report back (alas). But maybe having a second person
can re-ignite the interest.
- kickingvegas asked his Org tools (Scrim and Captee - Mac) to be
listed on Org mode page: https://orgmode.org/tools.html
- They are ok as long as they are not non-free software
- Here is that page's source code
https://git.sr.ht/~bzg/orgweb/tree/master/item/tools.org
- kickingvegas later submitted a patch
https://list.orgmode.org/[email protected]/
- I also mentioned an idea that has been on the back of my mind
for some time now. I am thinking how to involve more contributors to
Org mode development, and one of the possibilities is helping people
pick up some easy tasks, so that they can later contribute something
more complex. An obvious choice for easy tasks is bugs (there is never
shortage of those in Org). However, not everyone can pick up a bug
and fix it upfront - some are vary of their Elisp knowledge, or do
not know where to start (we do have many instructions in
https://orgmode.org/worg/org-contribute.html, but long text are not
always something people are eager to read).
- The idea was a kind of meetup, but for people who are interested
to get started with bug fixes or contribution in general.
I can spend some time not fixing bugs myself but instead helping
others to do the fixes. That will (hopefully) get us more people
familiar with Org code base. Some of them may later become regular
contributors, making sure that the project does not rely on a small
team of contributors.
- I am not sure how good the idea is though. Of course, I can try and
see, but that's partially about the organization. For example, I
tried to start this meetup earlier, but with a focus on people
reporting performance problems with Org in particular. That did
not work as well as what we have here in OrgMeetup.
- Sacha Chua suggested that Christian Tietze and Prots may
give useful insights as they are involved in coaching - something
at least a bit similar to the idea with helping people learn how
to contribute.
- I have also been playing with viewing web pages in Org-ish format
There is https://github.com/chenyanming/shrface that makes eww
render pages using foldable outlines, like in Org mode.
:chat:
[17:03] yantar92 is now the presenter
[17:03] [George Jones : VIEWER]: Are we having fun yet?
[17:05] [torusJKL : VIEWER]: Hi, first time here. Should I see or hear anything?
[17:05] [Zororg : VIEWER]: Did orgmode site have update recently?
https://orgmode.org/
Or am i seeing the example syntax being exactly how org mode operates in emacs?
toggle, link and so?
[17:08] [Zororg : VIEWER]: Nice
[17:09] [codaman : VIEWER]: This is my first time here. Are we allowed to ask
questions, or is their an agenda for today?
[17:10] [Zororg : VIEWER]: I use emacs git , sometimes i see high memory usage
and hard to figure whats causing this.
Even with profiler-report I see many times on completing-read or vertico.
But the culprit would be buffer-undo-list or something else.
even one time org-agenda was conusming high memory due to one org file..
[17:10] [George Jones : VIEWER]: I have a denote-silo like thing I created
recently I'd show, get feedback on (any denote users hers?)
[17:10] [George Jones : VIEWER]: here
[17:11] [codaman : VIEWER]: Yes, love Denote
[17:11] [torusJKL : VIEWER]: I'm a light Denote user.
[17:11] [codaman : VIEWER]: haven't used Silos yet; scared I will fragment
myself
[17:11] [yantar92 : MODERATOR]: https://protesilaos.com/emacs/denote-silo
[17:11] George Jones is now the presenter, set by yantar92
[17:12] [Zororg : VIEWER]: i want to use denote badly, but lazily just use
datetree single org file
[17:12] [codaman : VIEWER]: yes
[17:15] [codaman : VIEWER]: for Silos, are your org-agenda items in separate
views, one for each silo?
[17:16] [codaman : VIEWER]: so the diary is decoupled from the silos
[17:16] [codaman : VIEWER]: well, what about context?
[17:17] [codaman : VIEWER]: very clever, I want it
[17:18] [codaman : VIEWER]: Next question for everyone: I'm wondering, more
than ever, what is the current/popular/most effective AI solution. Is the
community leaning towards any tool more than others (gptel, agent-shell,
claude-code-ide, etc)?
[17:19] [torusJKL : VIEWER]: Would be great to have a look at the code @George
[17:20] yantar92 is now the presenter, set by yantar92
[17:21] [codaman : VIEWER]: for Denote, it's great for creating and finding
files. I had left it, but without it I would need to use something like
org-roam.
[17:22] [codaman : VIEWER]: Denote allows to type '/.' to see the folders in
the minibuffer and know that _tag or --filetitle will immediately find a file.
[17:22] [torusJKL : VIEWER]: George, when you create a new note do you create a
link first and than the new note or create first the note and than link it to
another Denote file?
[17:23] [codaman : VIEWER]: in that example
[17:23] [George Jones : VIEWER]: Code is here
https://codeberg.org/eludom/home.public.elisp-public/src/branch/master/elisp-public/gmj-denote-config.el
[17:24] [codaman : VIEWER]: the search shows all the files but you have to
browse to the list of files to select. With Denote, the completion exposes the
files and the file you want is already selected so you just press enter to
select it
[17:24] [George Jones : VIEWER]: Not isolated from my other denote config, so
start at (defvar gmj/denote-directories and go to the bottom of fille
[17:25] [codaman : VIEWER]: ahh
[17:25] [codaman : VIEWER]: ripgrep?
[17:25] [yantar92 : MODERATOR]: https://github.com/alphapapa/org-ql
[17:25] [Morgan : VIEWER]: do you use sparse trees?
[17:26] [codaman : VIEWER]: can you show another example with the org ql search
to find the nodes?
[17:26] [Zororg : VIEWER]: wait.. how do you search with query?
which exact command?
org-ql-find?
[17:26] [George Jones : VIEWER]: torusJKL I usuallly just create notes and
link if it makes sense
[17:27] [Zororg : VIEWER]: Nowadays i'm finding literate config laggy to handle.
[17:27] [codaman : VIEWER]: this search seems very flexible
[17:27] [yantar92 : MODERATOR]: helm-org-ql
[17:28] [codaman : VIEWER]: so you do you care about searching for tags?
[17:29] [George Jones : VIEWER]: I'm interested in the question that was poted
about AI. I'm manly just using GPTel for dialogs and claude-code command line
for fixing my old hand-coded scripts, etc.
[17:29] [codaman : VIEWER]: I haven't used org-ql for searching files. I've
only used for searching tags. This is very nice. Very fast and flexible.
[17:29] [codaman : VIEWER]: Did you share your code for the search with
helm+orgql?
[17:31] [codaman : VIEWER]: I want to use your code as a headstart, then edit.
[17:31] [yantar92 : MODERATOR]:
https://github.com/yantar92/emacs-config/blob/master/config.org#live-helm-org-ql-search-in-org
[17:32] [Zororg : VIEWER]: Any activity from alphapapa these days?
used to see them alot in systemcrafters and reddit.
[17:33] [Morgan : VIEWER]: but sparse trees don't mix with your fancy org-ql
stuff?
[17:34] [Zororg : VIEWER]: how to avoid deleting content when heading is
collapsed?
is there a config?
I use it delete alot, but its easy to remove if not vc chunks
[17:35] [Morgan : VIEWER]: oh ok. i don't use wither at the moment so I was
just curious. thanks 😊
[17:35] [Morgan : VIEWER]: *either
[17:35] [George Jones : VIEWER]: what are the little dots betweeen headings?
[17:36] [yantar92 : MODERATOR]: org-fold-catch-invisible-edits-commands
[17:36] [Zororg : VIEWER]: thanks!
[17:37] [George Jones : VIEWER]: thanks
[17:37] [Zororg : VIEWER]: but if i mark from heading start (^) and do
Backspace, it still deletes... hehe
[17:37] [yantar92 : MODERATOR]: prettify-symbols-mode
[17:37] [yantar92 : MODERATOR]: built-in
[17:38] [Zororg : VIEWER]: i've delete-selection to t
[17:38] [yantar92 : MODERATOR]: delete-active-region
[17:40] [Morgan : VIEWER]: we seem to always have an AI discussion here 😜 are
you still interested in talking about AI stuff or are you getting sick of it?
[17:41] [Zororg : VIEWER]: i mean, what i see from btop
[17:41] [Zororg : VIEWER]: Sometimes from 400Mb goes to >1 GB
[17:42] [yantar92 : MODERATOR]: M-x memory-report
[17:43] [Zororg : VIEWER]: do you use undo-fu-session?
even buffer-undo-list makes laggy as well
[17:43] [yantar92 : MODERATOR]: M-x malloc-trim
[17:44] [Zororg : VIEWER]: I tried igc with pgtk, it took even more memory lol
[17:44] [Zororg : VIEWER]: still not merged into master yet?
[17:44] [yantar92 : MODERATOR]:
https://yhetil.org/emacs-devel/[email protected]/
[17:44] [yantar92 : MODERATOR]: IGC branch - new garbage collector (incremental)
[17:46] [yantar92 : MODERATOR]: clear-image-cache
[17:46] [Zororg : VIEWER]: Nice.
https://codeberg.org/ezemtsov/ewm/
[17:47] [yantar92 : MODERATOR]: (setq gc-cons-percentage 0.2)
(setq gc-cons-threshold (* 200 1000 1000))
(add-hook
'after-init-hook
(lambda () (setq gc-cons-threshold (* 20 1000 1000))))
[17:48] [Zororg : VIEWER]: is it suggested to run
M-x garbage-collect sometimes?
[17:48] [Morgan : VIEWER]: a while ago I remeber there was some "explain-pause"
library
[17:49] [yantar92 : MODERATOR]: https://emacsconf.org/2023/talks/gc/
[17:49] [Morgan : VIEWER]: I see. it became the problem it sought to overcome
[17:50] [Nick Anderson : VIEWER]: I like gptel, and im also really liking
agent-shell.
[17:50] [codaman : VIEWER]: *workflow
[17:51] [codaman : VIEWER]: gptel...does it support latest models? like Opus
4.6?
[17:52] [George Jones : VIEWER]: workflow++ what are people doing?
[17:52] [codaman : VIEWER]: ok, thanks
[17:54] [codaman : VIEWER]: does anyone have a workflow which uses AI to
synthesize notes from separate files? for example if I want AI to look at all
'happy' tags, then provide insights, how do you do that?
[17:55] [codaman : VIEWER]: the result would be something like, "you seem to be
happy on payday or when having your favorite food"
[17:56] [codaman : VIEWER]: The idea is to find notes from different files and
send only those notes (nodes/headings) to the AI to get insights
[17:56] [Nick Anderson : VIEWER]: I have processes for that, but they are
outside emacs. Basically inferecne each one with structured data output
"enrichment" and then you can query that and maybe inference against that to
get some narrowed categories
[17:58] [torusJKL : VIEWER]: Codaman, maybe use the Emacs SKILL from
Agent-Shell that does an eval and receives the relevant text.
[18:00] [codaman : VIEWER]: are you able to pull just content from search
results headings? instead of sending the whole project, send search results?
[18:01] [torusJKL : VIEWER]: If you can write a function that does that than
you should be able to eval it using the Emacs skill.
[18:01] [codaman : VIEWER]: Ok, I just recently installed agent-shell so I will
lookup how to do that. Thanks.
[18:02] [Nick Anderson : VIEWER]: For each result, I think you will want a
structured response, that you store, that perhaps you will store some
embeddings (for similarity match etc ...). And then you will have wild
categories all over the place unless you direct it. So, from the enrichments,
you can have llm to normalize the categories. then you can find things that are
similar to your happy tags and surface their summaries.
[18:03] [codaman : VIEWER]: My tags are predictable and disciplined. I use
groups for tags too.
[18:03] [Nick Anderson : VIEWER]: your tags are, but the enrichments that come
out of what you ask the llm to extract for you may not be ... unless you are
very good about providing the subcategories it should subdivide into.
[18:03] [codaman : VIEWER]: I can use org-ql to retrieve tags, but I am
wondering if someone has a workflow to then send that to gptel or somewhere.
[18:03] [Nick Anderson : VIEWER]: there is the gptel meta
[18:05] [Zororg : VIEWER]: alright, I've to go.
Thank you for the meet and discussion!
[18:05] [Nick Anderson : VIEWER]: trying to find the link i cant find the link
[18:05] [codaman : VIEWER]: My headings are disciplined too. Usually start
with VERB: and an emotion. For example: "'ATE: ice cream and loved it". The
heading also have a CREATED:: property. I do not have a separate diary. The
entire org files are the diary since all headings have a date.
[18:06] [Nick Anderson : VIEWER]: i thoguth so , but i cant find it
[18:06] [Nick Anderson : VIEWER]: it's another package
[18:07] [yantar92 : MODERATOR]: https://codeberg.org/bajsicki/gptel-got
[18:07] [yantar92 : MODERATOR]: that has a tool for tags
[18:07] [torusJKL : VIEWER]: Codaman, this skill selects a region of a file
using eval:
https://github.com/xenodium/emacs-skills/blob/main/skills/select/SKILL.md
[18:08] [codaman : VIEWER]: THANK YOU!!!!!
[18:10] [codaman : VIEWER]: This was very helpful. Thank you again.
[18:11] [Sacha Chua : VIEWER]: Ooh, I had fun making my org-insert-link dwim
function look up bookmarks in an Org file and automatically link the region to
the bookmark if matching
[18:12] [George Jones : VIEWER]: share?
[18:12] [Sacha Chua : VIEWER]: Don't have multiple matching bookmarks 😉
[18:12] Sacha Chua is now the presenter, set by yantar92
[18:14] [codaman : VIEWER]: Sacha do you also use denote link or some other
file link?
[18:14] [Nick Anderson : VIEWER]: there is this org-similarty thing too
https://github.com/brunoarine/org-similarity
[18:15] [George Jones : VIEWER]: related tangent, Sacha: how do you capture
your bookmarks
[18:15] [George Jones : VIEWER]: (I was just wrestling with org-protocol)
[18:18] [yantar92 : MODERATOR]:
https://github.com/yantar92/emacs-config/blob/master/config.org#dynamic-block-to-retrieve-google-docs-contents-in-org-format
[18:19] yantar92 is now the presenter, set by yantar92
[18:21] [George Jones : VIEWER]: I've long had ~/elisp-public/
[18:21] [George Jones : VIEWER]: general stuff stays in my config. Other
stuff gets farmed out to ~/elisp-public
[18:22] Sacha Chua is now the presenter, set by yantar92
[18:23] [yantar92 : MODERATOR]:
https://sachachua.com/blog/2026/03/emacs-lisp-and-nodejs-getting-the-bolded-words-from-a-section-of-a-google-document/
[18:23] [yantar92 : MODERATOR]: blog post about google api
[18:25] yantar92 is now the presenter, set by yantar92
[18:25] [codaman : VIEWER]: Sacha I use Apple Music. I have been trying to
figure out how to play songs from Emacs. I know I can use the URL on a
external browser. But this solution means I can play it from within Emacs,
right?
[18:25] [codaman : VIEWER]: (This is very nice by the way)
[18:26] [Sacha Chua : VIEWER]: Oh yeah, I just use mpv in my follow
[18:26] [yantar92 : MODERATOR]: play-sound
[18:26] [yantar92 : MODERATOR]: (function)
[18:27] [Sacha Chua : VIEWER]: https://sachachua.com/dotemacs/index.html#videos
[18:27] [Sacha Chua : VIEWER]: my-org-video-follow
[18:28] [Sacha Chua : VIEWER]: I don't know if it works with Apple Music, though
[18:28] [Sacha Chua : VIEWER]: Oooh, there seems to be an apple music cli player
[18:28] [Sacha Chua : VIEWER]: so you can probably wrap something around that
[18:30] [codaman : VIEWER]: One last question: Has anyone figured out a good
workflow for mobile org files? I know there are several apps, but are people
just exporting to markdown (Obsidian) for mobile notes?
[18:31] [Morgan : VIEWER]: orgro for viewing a single file. orgzly for agenda
view
[18:31] [codaman : VIEWER]: I am Apple, (I know, I know, I know!!!)
[18:31] [yantar92 : MODERATOR]: https://orgmode.org/tools.html
[18:33] [Morgan : VIEWER]: orgro has support for checkboxes so is pretty great
as a shopping list app
[18:33] [kickingvegas : VIEWER]: My personal approach is to avoid data sync and
use a terminal emulator to get at your org files through emacsclient
[18:34] [codaman : VIEWER]: I struggle to get Blink keybinding to work. I tried
but I cannot get the ipad file mappings to work like they do on my Doom Emacs.
(I use Tailscale with Blink)
[18:35] [Morgan : VIEWER]: i've heard horror stories of how bad it is to do
storage on apple. android seems to be pretty easy
[18:35] [Nick Anderson : VIEWER]: I use syncthing. To avoid sync issues, I have
one capture file per other device (like one for my phone, one for a tablet etc
... ) and then from my workstation i can easily grab content and pull into my
larger corpus. So i cant really access everythging from the mobile apps, but i
can capture which is what i mostly care about, and i can get to everything from
my workstation.
[18:36] [codaman : VIEWER]: I have no storage issues with Apple devices;
neither locally nor in iCloud.
[18:36] [codaman : VIEWER]: I prefer the Tailscale access though. Like
kickingvegas said, it's a lot easier to have a 'home cloud'
[18:37] [Sacha Chua : VIEWER]: I miss the swipy keyboard thing when I'm in
termux. It's sort of available, but you have to swipe the input to the side,
type, and then tell it to send it in.
[18:37] [Sacha Chua : VIEWER]: Oooh I should probably focus on getting some
kind of whisper speech recognition going.
[18:37] [Sacha Chua : VIEWER]: but typing is more discreet.
[18:39] [Nick Anderson : VIEWER]: Termux is great
[18:39] [Nick Anderson : VIEWER]: i use that for serious emacs on android
[18:39] [Morgan : VIEWER]: latest version of android has added a built-in
feature (no app download needed) to run a debian virtual machine
[18:40] [Morgan : VIEWER]: haven't tried it, but my termux is borked so it
might be the next thing i try
[18:40] [Sacha Chua : VIEWER]: Question: Is there a nice way to get
sentence-at-point to understand Org items?
[18:40] [Sacha Chua : VIEWER]: Right now I'm just overriding it with advice.
[18:40] [Sacha Chua : VIEWER]: This is a test:
had been using EWM, so did not wanted to be laggy
- The quick brown fox jumps over the lazy dog.
It fails if you call (sentence-at-point) from, say, fox
[18:41] [Sacha Chua : VIEWER]: Ah, good, there's a way to override it and Org
is already doing some of it, thanks
[18:42] [Sacha Chua : VIEWER]: Yup, it just needs an extra entry in there
[18:42] [Sacha Chua : VIEWER]: (thing-at-point-provider-alist)
[18:43] [Morgan : VIEWER]: i haven't looked into it but it seems to be entirely
seperate from android. no file access or anything. but it does have GUI access
[18:43] [yantar92 : MODERATOR]: https://keepandroidopen.org/
[18:46] [Morgan : VIEWER]: not really anything to report because i haven't
started but i was planning on factoring out the org-habit function that
determines if a habit is done in a user customizable function
[18:46] [Morgan : VIEWER]: we've had a couple discussions on people wanting to
extend org-habit, with like numeric habits and stuff and I wanted to extend it
to understand org-clock
[18:47] [yantar92 : MODERATOR]:
https://list.orgmode.org/orgmode/[email protected]/
[18:47] [yantar92 : MODERATOR]: person expressed interest to refactor org-habit
[18:47] [Morgan : VIEWER]: sounds good. i'll cc them on my work then
[18:51] [Morgan : VIEWER]: that sounds fun. i'd show up
[18:54] [kickingvegas : VIEWER]: Would like to request some tools of mine to be
added (Scrim, Captee)
[18:54] [Sacha Chua : VIEWER]: Christian Tietze - he has office hours
[18:54] [kickingvegas : VIEWER]: Also Journelly
[18:54] [codaman : VIEWER]: is Prot still in Cyprus? if so, has anyone heard if
he's safe?
[18:56] [kickingvegas : VIEWER]: Sorta related, but my recent Org support for
Casual uses org section predicates a lot to narrow down command choices
[18:59] [torusJKL : VIEWER]: https://xenodium.com/journelly-for-ios
[19:00] [yantar92 : MODERATOR]:
https://git.sr.ht/~bzg/orgweb/tree/master/item/tools.org
[19:01] [yantar92 : MODERATOR]: https://orgmode.org/tools.html
[19:01] [yantar92 : MODERATOR]: https://orgmode.org/tools.org
[19:02] [Nick Anderson : VIEWER]: org-download?
[19:02] [Nick Anderson : VIEWER]: org-web-tools?
[19:02] [Nick Anderson : VIEWER]: .ive used org-web-tools a lot to yank pages
to org
[19:03] [Nick Anderson : VIEWER]: the archive thing is nifty as well
[19:04] [yantar92 : MODERATOR]: https://github.com/chenyanming/shrface
[19:04] [yantar92 : MODERATOR]: view websites as Org files. Any website
[19:04] [yantar92 : MODERATOR]: from eww
[19:04] [Nick Anderson : VIEWER]: that one is new to me, thanks
[19:07] [Nick Anderson : VIEWER]: bye, thanks again
[19:07] [George Jones : VIEWER]: thanks
:end:
--
Ihor Radchenko // yantar92,
Org mode maintainer,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>
