[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[BLOG] #12 [[bbb:OrgMeetup]] on Wed, Oct 9, 19:00 UTC+3
From: |
Ihor Radchenko |
Subject: |
[BLOG] #12 [[bbb:OrgMeetup]] on Wed, Oct 9, 19:00 UTC+3 |
Date: |
Sat, 26 Oct 2024 09:50:23 +0000 |
[ This time, I will try to use Org mailing list tracker and its
functionality for "blog posts". Let's see if I can make this post into
RSS feeds that are automatically posted in #org-mode Matrix room. ]
Back to the meeting notes.
- As usual, we started from Sacha's News
https://sachachua.com/blog/2024/10/2024-10-07-emacs-news/
- For a bit of a twist, before the official start of the stream, I did
a bit of a screencast working on an Org mode bug report
- https://list.orgmode.org/orgmode/878qvbstna.fsf@gmail.com/
- The bug is about Texinfo export of variable definition:
: - Variable: my-name ::
- When variable "my-name" is exactly =nil=, Org mode errs
- The problem lies in the implementation detail of ox-texinfo and
in Org handling of attribute values
- ~org-texinfo--split-definition~ replaces "variable"
description lists with special blocks
#+begin_src emacs-lisp
(apply #'org-element-create 'special-block
(list :type cmd
:attr_texinfo (list (format ":options %s" args))
:post-blank (if contents 1 0))
(mapc #'org-element-extract contents))
#+end_src
- later, ~org-texinfo-special-block~ uses
~org-export-read-attribute~ to parse the =:options variable-name=
- The problem is that ~org-export-read-attribute~ treats ~nil~
specially, making =nil= value indistinguishable from the
absence of attribute
- It is not fully clear how to address the problem. There are multiple
options
1. ~org-texinfo--split-definition~ to not use ~:attr_texinfo~
and instead store the value directly + add special handling
in ~org-texinfo-special-block~. This will avoid invoking
~org-export-read-attribute~ altogether
2. Or I may try to address the limitation of
~org-export-read-attribute~, so that it becomes possible to
specify =nil= as a value of export attribute.
- This seems better overall (in a more global scope outside
of the bug report), but it is not clear what can be done
(removing special handling of nil will likely cause
regression and is thus not an option)
- [2024-10-26 Sat] I went with option (1)
- Chinmay saw the 33 stashes lying around in my local Org git repository
- Those are not all actual work though
- My workflow often involves jumping between different Org branches
and carrying over WIP changes. For example, I sometimes start to
write code on main, but then realize that it is more suitable for
bugfix (or vice versa)
- To carry the code around I often just use stashes:
1. stash current changes
2. switch branch
3. unstash
- I also tend to work on a number of small fixes in parallel,
sometimes leaving the changes to "cook" for a few days to avoid
overlooking obvious flaws (another day - another fresh look on my code)
- If I do see a flaw, I do not immediately delete a stash, but may
start changes afresh, so the stash is kept around in case if I
change my mind about the approach to handle one or another fix
- So, many stashes in my git repo are not really meaningful. I clean
them up from time to time
- While I was doing the bug fixing screencast, someone asked (by
voice, so I do not recall who it was) a bit of
an off-topic question - how I configure parenthesis to be
highlighted in different colors depending on the relative depth wrt
point
- I use =highlight-parentheses= package
- See
https://github.com/yantar92/emacs-config/blob/master/config.org#highlight-parentheses-in-code
- visuwesh asked karhink (co-author of the new latex preview branch)
about the new proposal to generalize the previews to work outside
Org mode - in any arbitrary major mode buffer
- https://list.orgmode.org/87edbhljr7.fsf@hyperspace
- Recently, Karthik shared a new prototype on the list
87a5fffmuc.fsf@gmail.com/">https://list.orgmode.org/orgmode/87a5fffmuc.fsf@gmail.com/
- Here is a video demo from the recent email
https://www.youtube.com/watch?v=u44X_th6_oY (watch from 9 minute mark)
- I did some initial feedback during and after the video and later
followed up with an actual reply on the list
https://list.orgmode.org/orgmode/87o73rkgja.fsf@localhost/
- Also, maybe it would be a nice talk for EmacsConf :)
- Chinmay asked about Emacs APAC meetup (suitable for Asia-Pacific/Europe time
zones)
- It is monthly, every forth Saturday, 1400 IST
- See https://emacs-apac.gitlab.io/
- karthink asked about the performance of org-persist writing caches
to disk and how to debug it
- It can take time when there is a large number of caches
- For example, with my 20-30+Mbs of Org files loaded into Emacs,
it does take a few seconds to quit Emacs (not that I frequently
do it)
- One can use a profiler to check how long it takes for org-persist
to write things
- But you need a trick to prevent Emacs from actually exiting and
not showing the profiler data: make ~kill-emacs-hook~ throw an
error at the very end
: (add-hook 'kill-emacs-hook #'error 100)
- Then, M-x profiler-report after attempting to quit will work
- Later, the hook can be removed
: (remove-hook 'kill-emacs-hook #'error)
- In case of karthink, the possible problem is a sheer number of cache
entries as he is actively using latex previews that heavily make
use of the cache
- However, we need the actual profiler data to (1) confirm that
the problem is there; (2) figure out which part of org-persist
is slowing things down
- Chinmay asked about link previews in Org mode (similar to what many
instant messengers do)
- visuwesh suggested that creates previews for youtube links:
https://github.com/TobiasZawada/org-yt
- Of course, it is only specific to youtube links (special yt: link type)
- kathink's patch that adds an API to create arbitrary link previews
is a step towards having web (and other, like, say, pdf) link
previews
https://list.orgmode.org/orgmode/875xrqg6cb.fsf@gmail.com/
- The patch introduces :preview link parameter that will be used
by ~org-toggle-inline-images~
- The patch itself does not produce web link previews, but such
previews can be implemented using the proposed API
- There is even a demo previewing youtube links from unpublished
kathink's code
https://share.karthinks.com/org-link-preview-demo-2.mp4
- We then went on discussing how to generalize the previews to be
not tied to a specific website
- Chinmay suggested https://oembed.com/ API. It might be used to
retrieve previews from arbitrary websites
- I was not so optimistic about the idea because of my
experience using OpenGraph protocol to retrieve website
metadata in my Org capture scrambler package:
https://github.com/yantar92/org-capture-ref
- Even though there are standards to provide
website metadata (previews, authors, titles, dates, etc), they
are simply not followed equally by different websites
- ... and you are lucky if the protocol is simply not
supported. Sometimes, it is kind of supported, but the
returned data is a complete garbage
- So, to have some kind of generic previewer, there might be no
universal solution. We might have to go the route I took in
org-capture-ref: (1) push our luck with "standard" protocol;
(2) fall back to website-specific scramblers otherwise.
- It is actually doable, except some especially bad websites
that do not even provide metadata without running JS
- As long as we can download HTML to Emacs, it can be parsed
using built-in libxml parser and searched for website-specific
info
- John Wiegley shared his feature idea
- He wants to have a way to define multiple variants of the same
paragraph/block/heading/etc
#+begin_quote
The idea I was proposing earlier, org-layers, would give you
markup for specifying multiple versions of a region of text, and
both Org-mode and the HTML export would give you a simple
interface for changing the view both locally and globally.
#+end_quote
- The inspiration is from https://en.wikipedia.org/wiki/Project_Xanadu#History
- [2024-10-20 Sun] I did not look into that link during the
meetup, but now I am a bit confused. According to the wiki, the
major idea is transclusion, and some people (who probably looked
into the same wiki page), immediately mentioned
https://github.com/nobiot/org-transclusion and
https://elpa.gnu.org/packages/lentic.html
- daniel german shared his extension to org-transclusion:
including parts of code in other files (not Org) into Org
document:
https://github.com/dmgerman/org-transclude-fn
- The idea is slightly different from org-babel in a way that
babel is mostly from Org source to code, while
org-transclude-fn is opposite
- [2024-10-20 Sun] On the other hand, we have
~org-babel-detangle~ that can ingest code back into Org
files
- As deniel later replied, another difference is that
org-transclude-fn does not include actual text into Org
document. It is just to be displayed and read; not edited.
- org-transclusion, in contrast, allows editing the
"transcluded" text, so that edits are reflected in the
original buffer
- Then, we went further into the old discussion about Emacs
allowing to combine parts of multiple buffers, ideally with
per-region major modes
- One of the past discussions of the idea on emacs-devel:
https://yhetil.org/emacs-devel/87h8kou1c0.fsf@telefonica.net/
- And the most recent one (now active):
https://yhetil.org/emacs-devel/86ttdau8hj.fsf@gmx.net/T/#t
- ... similar discussions keep popping up on emacs-devel from
time to time (maybe every 1-2 years or so, AFAIR)
- Another analogy that immediately stroke my mind is the problem of
translations (of manuals or normal user documents)
- Past discussion on emacs-devel:
https://yhetil.org/emacs-devel/835y0b29rk.fsf@gnu.org/
- When creating a translation of slowly changing document like
manual, we cannot simply translate it all at once
- Parts of the document may be changed over time, often by
people who are not capable to keep all the available
translations up to date
- So, some system is necessary that keeps track of which parts
of the translated document are obsolete and which parts are
still up-to-date
- The classical solution for this task is https://po4a.org/index.php.en
- ... and it has no Emacs major mode
- So, I have been long playing around with the idea that Org mode
may be up to the task of maintaining the translations (of
manuals or just normal multi-lingual documents/books; inspired by
Juan Manuel MacĂas sharing a bilingual book written in Org mode:
https://list.orgmode.org/orgmode/87h70tyyiz.fsf@posteo.net/)
- What if we make use of affiliated keywords to mark some
paragraphs/drawers/blocks/headings as "translation"
- Such mark will also contain a link and hash of the "original"
paragraph/.../heading
- Then, depending on the #+LANGUAGE: keyword, we may choose to
export different variants of the text
- If the original source texts change, but the translations do
not, it should then be easily tracked, and we can provide
various ways to handle this, like: throw an error; mark the
exported obsolete translation somehow; put the original
instead of the translation; both; etc.
- We can go even further, and link the "translation" to, for
example, Elisp function docstring (or even body). Then, if the
docstring changes, we can automatically flag the related
portion of the manual obsolete. Think about all the extended
explanations about various Org mode commands - it is often
hard to remember keeping them in sync with changes in the
function code.
- karthink asked about an update on my big Org codebase refactoring
project
- There no progress since I started my new job in September (so I
have less time for Org mode now)
- I also need to sort out my copyright paperwork with the new
employer, which is taking some time. I am being cautious about
patches (especially non-trivial) until things are settled
- We had a moment of awkward silence and I decided to look into some
recent feature requests
- https://list.orgmode.org/orgmode/87frq13w48.fsf@localhost/
proposes to extend Org timers (M-x org-timer; [[info:org#Timers]])
- The idea is to allow chaining multiple timers one after another,
akin what people do with work/rest cycles in Pomodoro
- If anyone is interested in such features, please chime in on the
mailing list. Unless there are replies, I will have to refute
the proposal. (Mostly because it might be non-trivial code-wise
and I do not want to complicate code without popular demand for
this new feature.)
- Chinmay raised a topic of using transient in more places (including
Org mode)
- My stance of transient is that we should eventually switch all the
Org menus to use it.
- We generally want to rely more on the existing Emacs features and
get rid of Org mode parts that are re-implementing the existing
functionality (or predate the time when relevant generic
functionality, like transient, have been added to Emacs)
- Similarly, we want to move parts of Org mode that are generic
enough into Emacs itself
- Several users raised concerns about using transient though
- Mostly along the lines that is still has problems like not
supporting what Emacs users are used to
- For example, ~C-h k~ does not work inside transients
- Transient menus are also hard to create dynamically: a critical
feature required for many Org mode menus
- visuwesh also linked to not-yet-closed bug report for transient
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=52058
- I myself faced a few (unreproducible) situations when transient
left my Emacs in unusable shape with key bindings completely
broken
- Having said the above, transient is still going to be an
improvement as Org menus are even worse in terms of supporting
~C-h k~ or even simply switching to the menu buffer (which, in
contrast, transient allows). After we solve the dynamic menu
building that is. [2024-10-25 Fri] But there is a hope:
https://list.orgmode.org/orgmode/8734m28l9a.fsf@gmail.com/
- As karthink pointed, a better documentation (user: and
discoverability!) would greatly benefit developing more
transient uses
- There is a good attempt to improve the docs in
https://github.com/psionic-k/transient-showcase; would be nice
if it were incorporated into upstream manual
- Also, karthink suggested that "The transient manual made sense
only after I read the EIEIO manual"
- Every instance of transient menu item is an instance of
EIEIO class
- Right after we discussed problems with transient, kickingvegas, a
big transient fan, joined
- Although, this time he did not talk about transient :)
- He presented his recent blog post on a small helper for writing Org table
formulas
http://yummymelon.com/devnull/referencing-org-table-cells-with-text-regions.html
- The problem post is trying to solve is inputting cell references into the
formulas
- Currently, one has to write the references manually, like =@13$7=
- This sometimes involves annoying counting of rows/columns
- Org does allow named rows/columns, which simplifies things,
but naming rows and columns require dedicated effort, which is
only worth it for really large and complex formulas
https://orgmode.org/manual/Advanced-features.html
- =C-c '= (~M-x org-table-edit-formulas~) helps somewhat by
highlighting the referenced table cell and by providing
=S-arrow= keys to modify reference at point interactively: move
reference up/down/left/right; but entering the reference
_initially_ still remains awkward
- kickingvegas used context menu approach to solve the problem
- He introduced a new context menu item, which copies cell
reference at point into the kill ring
- I suggested integrating the idea of getting cell reference via
mouse with ~org-table-edit-formulas~
- When inside edit buffer, mouse may automatically insert
cell/cell range reference at mouse into the formula editor
- More details in the followup discussion on the mailing list
95CE1447-FC08-431A-9CA1-83B4C3F77BA7@gmail.com/T/#t">https://list.orgmode.org/95CE1447-FC08-431A-9CA1-83B4C3F77BA7@gmail.com/T/#t
- visuwesh raised the topic for recent Org mode's support for M-x yank-media
- Our support revived interest in better implementations of M-x
yank-media on all the Emacs platforms
- Emacs does not support M-x yank-media on Windows well
- https://debbugs.gnu.org/cgi/bugreport.cgi?bug=71909 discusses how
to implement it considering Windows quirks (Windows clipboard is
not directly compatible with the notion of mime type used in
yank-media support for GNU/Linux)
- Beginning of the thread:
https://yhetil.org/emacs-bugs/tencent_7FA4E415A96083033C6BED7C7354AEE02505@qq.com/
- As a reminder, I now post links to meetup summaries and dates for
the upcoming meetups at https://orgmode.org/worg/orgmeetup.html
- This is in addition to Mastodon and the mailing list
- My new Mastodon server is https://fosstodon.org/@yantar92
:chat:
[17:31] Welcome to <b>[[bbb:OrgMeetup]]</b>!<br /><br />For help on using
BigBlueButton see these (short) <a href="https://www.bigbluebutton.org/html5"
target="_blank"><u>tutorial videos</u></a>.<br /><br />To join the audio bridge
click the phone button. Use a headset to avoid causing background noise for
others.<br /><br />This server is running <a
href="https://docs.bigbluebutton.org/" target="_blank"><u>BigBlueButton</u></a>.
[17:38] Ihor Radchenko : Hi
[17:38] Chinmay : Hi!
[17:38] Ihor Radchenko : We are starting in 20 minutes
[17:38] Chinmay : ye no problem
[17:39] Ihor Radchenko : Meanwhile, the latest Emacs News:
https://sachachua.com/blog/2024/10/2024-10-07-emacs-news/
[17:41] Chinmay : omg so many stashes - i need to start naming my stashes to
use them more
[17:49] Ihor Radchenko : highlight-parentheses
[17:59] visuwesh : actually, can you have a variable named nil?
[18:00] visuwesh : ahhhh i see
[18:01] visuwesh : i see, i thought the report was about actually setting a
variable named nil
[18:02] visuwesh : ahhh okay
[18:04] visuwesh : a question for karthik: did the thread in emacs-orgmode
about adapting the new latex-preview code for auctex ever work out?
https://bbb.emacsverse.org/b/iho-h7r-qg8-led or will it have to be changed
compeltely now with the new wip emacs-wide intergration?
[18:04] visuwesh : https://list.orgmode.org/87edbhljr7.fsf@hyperspace
[18:05] visuwesh : got the link wrong, sry
[18:05] visuwesh : yep
[18:06] visuwesh : yes
[18:06] Dave Marquardt : yes
[18:06] karthink : The demo of latex is at around the 9 minute mmrk
[18:07] karthink : *demo of latex-mode
[18:07] Dave Marquardt : hehe
[18:08] karthink : @visuwesh the demo is based on prototype code that's not
very robust
[18:09] visuwesh : hmm, okay. i was modelling latex preview for my goldbook
interface after that thread but it just never worked and i never could work out
why
[18:09] visuwesh : i wanted to ask you via mail but i left it for later TM and
you released this video :)
[18:09] karthink : The changes required in org-latex-preview are not yet in the
dev branch
[18:10] visuwesh : was it ever expected to work before these changes though?
[18:10] karthink : Yes, the code provided by Tony Zorman should work with the
dev branch as is
[18:10] visuwesh : i used org-latex-preview-place with (BEG END VALUE) passed
to it
[18:10] visuwesh : hmm i see
[18:12] karthink : No I just made it to demo the possibility of latex previews
across Emacs
[18:13] Corwin Brust : gm :)
[18:13] karthink : Yeah it could be played at Emacsconf, I guess. In any case,
I need feedback on the API for plugging it into other major modes
[18:16] karthink : It will still `(require 'ox)` for the preamble, but yeah,
independent of Org-mode would be great.
[18:17] karthink : The other issue is that I want to avoid runtime dispatch on
the major-mode, because we run code in after-change-functions that needs to be
very fast
[18:17] Ihor Radchenko : 1. use thingatpt for boundaries
[18:17] Ihor Radchenko : 2. use custom variable/function for preamble
[18:18] karthink : I mean code like (funcall (plist-get major-mode ...))
[18:18] karthink : Yeah, the current prototype is using buffer-local variables
[18:19] karthink : Noted your point about providing defaults.
Okay, we can discuss this further in the thread? I think it's not relevant to
most people in the meetup
[18:20] karthink : No worries, please take your time
[18:24] Chinmay : is the emacs apac meetup every saturday?
[18:25] visuwesh : I believe it is monthly?
[18:25] Chinmay : ah
[18:25] visuwesh : happens around 2 pm or 2:30 pm IST iirc
[18:25] Chinmay : ye
[18:25] Chinmay : 2
[18:26] karthink : @Ihor that reminds me -- is it expected for Emacs to take
4-5 seconds to exit because of org-persist in kill-emacs-hook?
[18:27] karthink : It happens every time to me, I think it's because of the
large volume of latex previews that are cached
[18:28] karthink : How do you profile kill-emacs-hook?
[18:28] visuwesh : Can you not simplyprofile org-persist-write-all?
[18:30] Ihor Radchenko : (add-hook 'kill-emacs-hook #'error 100)
[18:30] Ihor Radchenko : This way, can profile killing emacs
[18:30] karthink : @Ihor, yeah, my experience is similar to yours just now
[18:31] Ihor Radchenko : error will prevent actual killing
[18:32] karthink : Okay, I'll find out what's happening
[18:33] karthink : Yeah, I'm reminded of the problem only when I quit Emacs,
which isn't very often
[18:39] Chinmay : has someone tried to add link previews to org mode?
[18:39] Chinmay : this thing https://imgur.com/ZqPjZtE
[18:43] visuwesh : https://github.com/TobiasZawada/org-yt this thing maybe?
[18:44] Chinmay : i'll check this out, thanks
[18:44] visuwesh : but is there a general solution for this problem? im sort of
interested in this too
[18:45] Chinmay : iinw there's some sort of standard for this too?
[18:45] visuwesh : only thing i can remember is a PR for ement.el to add link
previews but i think that simply queries the matrix homeserver
[18:45] Chinmay : like on the web
[18:45] Chinmay : not emacs
[18:45] visuwesh : i see
[18:46] visuwesh : yes
[18:46] daniel german : There is a package called (require 'org-yt) that allows
to have previews of an youtube snapshot.
[18:46] visuwesh : we can't see your screen though
[18:46] karthink : @visuwesh what do you mean by a general solution?
[18:47] Chinmay : general links, not just yt i believe
[18:47] visuwesh : i presume the package works by querying youtube for the
thumbnail. i was thinking if there was a gen solution because matrix, discord,
etc. show a nice preview of links regardless of the website
[18:48] daniel german : yes, thi s is only for youtube. it creates a special
type of link and caches it
[18:48] daniel german : yt:video-url
[18:48] Chinmay : i see
[18:49] karthink : @Ihor there's the org-link-preview patch
[18:49] Ihor Radchenko : https://github.com/TobiasZawada/org-yt
[18:50] Chinmay : this is the standard https://oembed.com/
[18:50] karthink : https://list.orgmode.org/875xrqg6cb.fsf@gmail.com
[18:50] karthink : It's the :preview link parameter
[18:51] John Wiegley : Good morning (from California timezone)
[18:51] karthink : I think org-yt is advising org-display-inline-images (or
whatever it's called)
[18:53] karthink : https://share.karthinks.com/org-link-preview-demo-2.mp4
[18:54] karthink : It's just an image placed over the link with C-c C-x C-v
[18:55] Chinmay : ye
[18:56] karthink : It looks like oembed can also be used for inline link
previews in Org mode
[18:57] Chinmay : yes
[18:57] Chinmay : lmao
[18:58] Chinmay : i think twitter does its own thing
[18:59] Chinmay : og:image has a png link
[19:01] Chinmay : no idea
[19:02] Chinmay : nice nice
[19:03] visuwesh : Emacs can use libxml for parsing html and xml
[19:03] Chinmay : very cool
[19:03] visuwesh : there's an elisp library for parsing xml too iirc
[19:05] John Wiegley : I just had a feature idea I wanted to brainstorm
[19:06] John Wiegley : Project Xanadu
[19:06] Chinmay : ihor your voice is echoing
[19:06] daniel german : Ted Nelsons' hypertex
[19:06] daniel german : t
[19:07] daniel german : It might be an interesting additjion to org-transclude:
[19:07] daniel german : to have a parameter that indicates isi only the title
or the content of the transclusion are transcluded
[19:08] John Wiegley : I hadn't thought of tying it in with transclusion, but I
love that idea!
[19:08] John Wiegley : Both!
[19:08] John Wiegley : Translation of multiple languages, and transclusion of
title/content/etc
[19:08] karthink : What is org translation?
[19:09] John Wiegley : Yes, my microphone on these websites often produces
echo, my apologies.
[19:09] daniel german : org-transclude not org translation
[19:09] John Wiegley : https://github.com/nobiot/org-transclusion
[19:10] John Wiegley : I use Org transclusion so that I can give my employees
feedback in a meeting file, and then transclude that feedback into their
personnel file for the yearly review.
[19:10] daniel german : I am adding transclusion of a function, still in beta,
and it will be moved to Nobiot's repo as an module:
https://github.com/dmgerman/org-transclude-fn
[19:11] John Wiegley : Saves me from having to either centralize the feedback
in a single file, or having to chase links to read them all in one go.
[19:12] visuwesh : hmm how does this different from org-babel?
[19:12] Ihor Radchenko : https://elpa.gnu.org/packages/lentic.html
[19:13] daniel german : I'll respond:
[19:13] daniel german : it executes a function, and dynamically inserts the
result, but the result is not part of the actual buffer and it is read only.
[19:13] karthink : It would be great if Emacs provided C-level support for
including a part of one buffer in another. Would solve various issues,
including Org babel fontification, LSP support, "native" transclusion etc
[19:14] John Wiegley : @daniel It sounds like you could build a webserver on
that idea
[19:14] daniel german : :) once you can execute a function, the sky is the
limit :)
[19:15] visuwesh : if it is never going to clog up the result, that would be
nice yes
[19:15] visuwesh : it can be a bit inconvenient to switch to *Async Shell
Command Output* or the sync ersion of that
[19:16] John Wiegley : The idea I was proposing earlier, org-layers, would give
you markup for specifying multiple versions of a region of text, and both
Org-mode and the HTML export would give you a simple interface for changing the
view both locally and globally.
[19:17] Ihor Radchenko :
https://yhetil.org/emacs-devel/87h8kou1c0.fsf@telefonica.net/
Discussion about having text from multiple buffers in one buffer
[19:17] daniel german : One nice thing of org-transclude is that you can edit
the source text in the transcluding buffer. I highly recommend it.
[19:18] karthink : @John Would all versions be written to the file?
[19:18] John Wiegley : AYes
[19:19] John Wiegley : #+begin_layered
Version one (full text)
,#+next_layered
Version two (summary)
,#+next layered
Version three (translation?)
,#+end_layered
[19:21] John Wiegley : That hashing idea is really nice, Ihor, I could see this
being applicable to code comments in a source file too...
[19:21] John Wiegley : Like, the PR would not pass CI unless the comment had
been updated with the new hash, which requires revision of that text.
[19:25] karthink : @Ihor any update on your big Org refactoring effort?
[19:25] Ihor Radchenko : https://yhetil.org/emacs-devel/835y0b29rk.fsf@gnu.org/
Discussion about translations on Emacs devel
[19:25] Jeff Trull : Karthik claims to not be a programmer too ;)
[19:29] Ihor Radchenko :
https://list.orgmode.org/orgmode/87frq13w48.fsf@localhost/
[19:29] Ihor Radchenko : feature request: Chaining multiple org timers
[19:30] Chinmay : i love the transientification of everything
[19:30] John Wiegley : transient does seem like a great UI modality for many
things
[19:31] user : I wish transient played more nicely with "C-h k" etc. I don't
know if things have improved in that regard, but I recall going from transient
menu-item to function behind it was not easily discoverable
[19:31] karthink : It's hard to modify a transient menu though
[19:32] karthink : Compared to keymaps
[19:32] user : Fair, it may be better than Org menus
[19:32] visuwesh : I really hope the issues raised in
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=52058 will be resolved
[19:32] kickingvegas : hi folks - sorry to jump in late
[19:33] Jeff Trull : Perfect timing
[19:36] Ihor Radchenko :
https://list.orgmode.org/orgmode/877cbamq2q.fsf@gmail.com/
[19:36] Ihor Radchenko : feature: transient menu for opening citations
[19:37] karthink : Oh yeah, I've broken transient several times
[19:38] visuwesh : it i mostly a request for better documentation
[19:38] user : s/documentation/discoverability/
[19:38] visuwesh : and in-transient text that tries to help the user when
trying a transient for the first time
[19:38] visuwesh : documentation too, honestly
[19:39] Ihor Radchenko : https://github.com/psionic-k/transient-showcase
some extra docs
[19:39] Chinmay : gtg, bye!
[19:39] visuwesh : i find the documentation very opaque as a user
[19:39] user : You don't get the same thing
[19:40] karthink : The transient manual made sense only after I read the EIEIO
manual
[19:40] visuwesh : oh i didn't realise that repo contained user documentation.
i thought it was only s bunch of examples so never bothered to open it
[19:42] Jeff Trull : Karthink that's interesting what is the connection between
an OO library and a menu library
[19:43] karthink : I think the prefix and suffix distinction made sense when
transient was a smaller library -- since a transient-prefix is a generalization
of a prefix argument to an interactive elisp command
[19:43] karthink : @Jeff every transient menu item is an instance of an EIEIO
class
[19:44] Jeff Trull : Ahhh
[19:44] John Wiegley : Where was it posted?
[19:45] user : Every now and again I wish there was a primer for EIEIO and
cl-defstruct. Mostly because I don't use them often enough to ever remember
their conventions
[19:45] kickingvegas :
http://yummymelon.com/devnull/referencing-org-table-cells-with-text-regions.html
[19:45] user : primer as in cheatsheet as opposed to manual
[19:45] John Wiegley : Thank you! And nice to hear you as well.
[19:45] karthink : @Jeff This also means if you want a menu item with bespoke
behavior you have to define a new EIEIO class that implements it, just for one
item
[19:48] Ihor Radchenko : 3.5.10 Advanced features describes how to create named
rows/columns
[19:50] karthink : Not seeing KickingVegas' screen yet
[19:51] Dave Marquardt : Unless your screen is the "Welcome to BigBlueButton"
page
[19:53] karthink : @Ihor, what was the command you invoked to edit the tblfm
interactively?
[19:53] karthink : C-c '
[19:53] karthink : Got it, did you call it with the cursor in the #+TBLFM line?
[19:53] Ihor Radchenko : M-x org-table-edit-formulas
[19:54] karthink : Cool thank you
[19:55] Dave Marquardt : yes, I see Ihor's screen
[20:00] Ihor Radchenko : 42.9.2 Overlay Properties
[20:00] Ihor Radchenko : when in org-table-edit-formulas mode, you can create
overlay over the table and add 'keymap property to that overlay
[20:00] karthink : (info "(elisp) Overlay Properties")
[20:02] Ihor Radchenko : 22.7.5 Drag Events
[20:02] Ihor Radchenko : 22.7 Input Events
[20:03] John Wiegley : I agree, there's inertia for me too in naming cells
properly
[20:04] visuwesh : yank-media will be properly supported in Windows soon
enough:
https://yhetil.org/emacs-bugs/bafcfc5c-e9d5-402c-a6de-321d49229386@imayhem.com/
it would be nice to provide feedback on what file types will be useful for
handlers (currently images, text/html, filenames, audio and video are mentioned)
[20:05] visuwesh : it is about yank-media
[20:05] visuwesh : yea that was another bug report, with a proepr subject line
[20:05] visuwesh : yeah but they are trying to gather user feedback on useful
clipbaord items right now
[20:06] visuwesh : https://debbugs.gnu.org/cgi/bugreport.cgi?bug=71909
[20:07] visuwesh : what clipboard items would be useful for future yank-media
usage
[20:09] Ihor Radchenko : bug#71909: 30.0.60; Can not use yank-media for pasting
image from clipboad in org-mode on Windows platform
[20:09] Ihor Radchenko : actual thread title
[20:09] visuwesh :
https://yhetil.org/emacs-bugs/tencent_7FA4E415A96083033C6BED7C7354AEE02505@qq.com/
[20:11] visuwesh : i think in windows yo uhave to translate the clipboard data
so that it matches the mimetype string that is followed in linux
[20:11] visuwesh : so they will probably end up having a specific bunch of
supported types so it would be nice to alert the devs on what clipboard items
would be useful
[20:12] visuwesh : yea, it is the same on mac too. i think they have a very
limited num of supported clipboard items if im not wrong
[20:12] visuwesh : (at least on the ns port)
[20:12] visuwesh : yea, the pushed needed to be given c:
[20:13] daniel german : thank you.
[20:13] oylenshpeegul : Thanks, Ihor! Thanks, everybody!
[20:14] visuwesh : thank you all
[20:14] Ihor Radchenko : Meetup page and annoucements:
https://orgmode.org/worg/orgmeetup.html
Also, on Org mailing list and https://fosstodon.org/@yantar92
[20:14] kickingvegas : thanks all
[20:14] karthink : Thanks for the meetup
[20:15] John Wiegley : Good bye!
:end:
--
Ihor Radchenko // yantar92,
Org mode contributor,
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>