org mode moves to GNU emacs core

org mode moves to GNU emacs core

[Uwe Brauer]

Hi

I am not sure whether I understand that discussion in emacs dev
correctly. Will orgmode be moved into the GNU emacs try as it was done
with gnus?

If so, how can I download and test for example the current master
branch.

I didn't mind that policy for gnus, since the development is rather
slow and install GNU emacs every 6 months, but org mode is different
beast.

Uwe Brauer 

Re: org mode moves to GNU emacs core

[Rasmus]

Uwe Brauer <oub@mat.ucm.es> writes:

> I am not sure whether I understand that discussion in emacs dev
> correctly. Will orgmode be moved into the GNU emacs try as it was done
> with gnus?

AFAIK, there are no such plans.  The *version* of Org shipped with Emacs
is being updated to v9, though.

But perhaps I missed a thread that explicitly talks about moving the
development center of Org...

Rasmus 

-- 
Governments should be afraid of their people

Re: org mode moves to GNU emacs core

[Uwe Brauer]

>>> "Rasmus" == Rasmus  <rasmus@gmx.us> writes:

   > Uwe Brauer <oub@mat.ucm.es> writes:
   >> I am not sure whether I understand that discussion in emacs dev
   >> correctly. Will orgmode be moved into the GNU emacs try as it was done
   >> with gnus?

   > AFAIK, there are no such plans.  The *version* of Org shipped with Emacs
   > is being updated to v9, though.

Thanks for clarifying.


   > But perhaps I missed a thread that explicitly talks about moving the
   > development center of Org...

More likely that I did not catch up the thread and panicked :-D

Re: org mode moves to GNU emacs core

[Bastien]

Hi Uwe,

Uwe Brauer <oub@mat.ucm.es> writes:

> I am not sure whether I understand that discussion in emacs dev
> correctly. Will orgmode be moved into the GNU emacs try as it was done
> with gnus?

for the record, I would be in favor of this.  Why?

- Less installation headaches
- Less maintainance and backward-compatibility headaches
- Possibility of having etc/TODO in Emacs using org-mode
- Attracting Org developers/contributors to Emacs repo

As a maintainer, I don't see any advantage of having Org
maintained as an ELPA package.

-- 
 Bastien

Re: org mode moves to GNU emacs core

[Uwe Brauer]

>>> "Bastien" == Bastien  <bzg@gnu.org> writes:

    > Hi Uwe,
    > Uwe Brauer <oub@mat.ucm.es> writes:

    >> I am not sure whether I understand that discussion in emacs dev
    >> correctly. Will orgmode be moved into the GNU emacs try as it was done
    >> with gnus?

    > for the record, I would be in favor of this.  Why?

    > - Less installation headaches
    > - Less maintainance and backward-compatibility headaches
    > - Possibility of having etc/TODO in Emacs using org-mode
    > - Attracting Org developers/contributors to Emacs repo

    > As a maintainer, I don't see any advantage of having Org
    > maintained as an ELPA package.

But the release cycles are very different, so in order to have always
the latest stable org package, I need to compile and install the whole
GNU emacs beast. I thought the whole idea of a package system is to
avoid this headache. 

Re: org mode moves to GNU emacs core

[Eric S Fraga]

[-- Attachment #1: Type: text/plain, Size: 601 bytes --]

On Monday,  3 Jul 2017 at 08:33, Uwe Brauer wrote:
> But the release cycles are very different, so in order to have always
> the latest stable org package, I need to compile and install the whole
> GNU emacs beast. I thought the whole idea of a package system is to
> avoid this headache. 

I agree.  I like tracking org development but would not be at all keen
on tracking emacs in the same way.  I was actually quite disappointed
when gnus went into emacs master as I can no longer easily play with any
experimental branches etc.

-- 
: Eric S Fraga via Emacs 26.0.50, Org release_9.0.9-551-g92e8c8

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 194 bytes --]

Re: org mode moves to GNU emacs core

[Alan Schmitt]

[-- Attachment #1: Type: text/plain, Size: 1090 bytes --]

"ESF" == Eric S Fraga <e.fraga@ucl.ac.uk> writes:

ESF> On Monday,  3 Jul 2017 at 08:33, Uwe Brauer wrote:
>> But the release cycles are very different, so in order to have always
>> the latest stable org package, I need to compile and install the whole
>> GNU emacs beast. I thought the whole idea of a package system is to
>> avoid this headache. 

ESF> I agree.  I like tracking org development but would not be at all keen
ESF> on tracking emacs in the same way.  I was actually quite disappointed
ESF> when gnus went into emacs master as I can no longer easily play with any
ESF> experimental branches etc.

Same here. I stopped trying to debug some gnus issues when I could no
longer easily compile and install new versions independently of emacs.
(I use a port of emacs for macOS, so I cannot simply recompile emacs. I
need to wait until the changes are incorporated in that port.) It would
be too bad if the same happened with org mode.

Alan

-- 
OpenPGP Key ID : 040D0A3B4ED2E5C7
Monthly Athmospheric CO₂, Mauna Loa Obs. 2017-05: 409.65, 2016-05: 407.70

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: org mode moves to GNU emacs core

[Phillip Lord]

Bastien <bzg@gnu.org> writes:

> Hi Uwe,
>
> Uwe Brauer <oub@mat.ucm.es> writes:
>
>> I am not sure whether I understand that discussion in emacs dev
>> correctly. Will orgmode be moved into the GNU emacs try as it was done
>> with gnus?
>
> for the record, I would be in favor of this.  Why?
>
> - Less installation headaches
> - Less maintainance and backward-compatibility headaches
> - Possibility of having etc/TODO in Emacs using org-mode
> - Attracting Org developers/contributors to Emacs repo
>
> As a maintainer, I don't see any advantage of having Org
> maintained as an ELPA package.


The advantage should be that it allows an independent upgrade cycle for
org from the long cycle of Emacs. I presume you do see this as an
advantage? The issue is, surely, that it's too much of a PITA for the
advantage that you gain?

If that's true, I'm interested in which bits are a PITA. Is it
fundementally because of ELPA? That is, the latest version of org-mode
has to support more than one version of Emacs? Or, is it having two
version repos, with all the merging?

Phil

Re: org mode moves to GNU emacs core

[Bastien Guerry]

Hi Philip,

phillip.lord@russet.org.uk (Phillip Lord) writes:

> I presume you do see this as an advantage? The issue is, surely,
> that it's too much of a PITA for the advantage that you gain?

Well, it's not really about PITA-or-not-PITA, it's just that I want
org-mode to be the default mode for some files in Emacs, and having
org-mode in Emacs' core is the most simple way to go for this.

Maintainers of projects like Gnus or CEDET don't want their code to
live outside of Emacs repo neither, so I guess simplicity is a big
win.

-- 
 Bastien

Re: org mode moves to GNU emacs core

[qTim Cross]

Just to throw my 2 cents in.

While I can understand the benefits of being able to easily install the
latest org package via elpa, I think there are some significant benefits
to org being a part of core Emacs.

I currently find three issues with the current situation which may be
somewhat resolved if org was part of core emacs.

1. Problems with mixed versions. Currently, Emacs has org 8.x included
in the distribution. This is despite 9.x being out before the release of
25.2. Something needs to be done to improve coordination and perhaps if
it was part of the core, this would be more likely. At any rate, the
current situation means you need to be very careful to ensure no org
feature is loaded before the ELPA package is loaded or you will get odd
behaviour and the symbol's value is void errors. 

2. If you just want to load the ELPA version of org (not
org-plus-contrib) it can be a real pain. You have to play around with
package lists to ensure you actually get the right one. This can be a
real hassle if you also use the use-package package as you will often
get the older version bundled with Emacs if you don't have your package
lists in the right order.

3. I would really like to see two completely separate packages rather
than having org and org-plus-contrib. Currently, if you have packages
which have org as a dependency and you have loaded org-plus-contrib
rather than just org, you will end up with both. Not a big issue, unless
your on a slow link as now you will download updates for both org and
org-plus-contrib. (there is no 'cleverness' with ELPA dependency
specifications - you cannot specify alternative dependencies).

A lot will depend on when org becomes part f core. The trick will be to
do it once development of org slows down. I've been using org for a long
time now and have noticed that the rate of new features being added has
slowed down. Much of the changes now is about improvement and refinement
of the code base. I would imagine that at some point, things will become
even more stable with fewer releases. This would be the point at which
it would make sense to bring into core.

The other advantage of being part of core is that updates and changes to
Emacs will be integrated into org much better. We won't see situations
where new versions of Emacs require a rush to update org. for the end
user, this should create a much more stable org environment.

Then of course, there will always be the option to run org straight from
the git repository for those who really want the latest version. I find
that once you have the path added to load-path, running from the git
repo is not much more effort than installing the latest ELPA package. 


Uwe Brauer writes:

>>>> "Bastien" == Bastien  <bzg@gnu.org> writes:
>
>     > Hi Uwe,
>     > Uwe Brauer <oub@mat.ucm.es> writes:
>
>     >> I am not sure whether I understand that discussion in emacs dev
>     >> correctly. Will orgmode be moved into the GNU emacs try as it was done
>     >> with gnus?
>
>     > for the record, I would be in favor of this.  Why?
>
>     > - Less installation headaches
>     > - Less maintainance and backward-compatibility headaches
>     > - Possibility of having etc/TODO in Emacs using org-mode
>     > - Attracting Org developers/contributors to Emacs repo
>
>     > As a maintainer, I don't see any advantage of having Org
>     > maintained as an ELPA package.
>
> But the release cycles are very different, so in order to have always
> the latest stable org package, I need to compile and install the whole
> GNU emacs beast. I thought the whole idea of a package system is to
> avoid this headache. 


-- 
Tim Cross

Re: org mode moves to GNU emacs core

[Eric S Fraga]

[-- Attachment #1: Type: text/plain, Size: 846 bytes --]

On Monday,  3 Jul 2017 at 12:40, Bastien Guerry wrote:

[...]

> Well, it's not really about PITA-or-not-PITA, it's just that I want
> org-mode to be the default mode for some files in Emacs, and having
> org-mode in Emacs' core is the most simple way to go for this.

This, I must admit, is quite a good reason.  I do think of org as an
integral part of Emacs these days (org is my default mode essentially)
so hadn't stopped to think that without it being formally part of Emacs,
Emacs cannot depend on it.

Okay.  I'm convinced.  :-)

Will have to update emacs-snapshot more often then... although
unfortunately not available for my wee Pandora as far as I can see.  And
the thought of building Emacs on the Pandora sends me into a catatonic
state (600 MHz arm5 processor).

-- 
: Eric S Fraga via Emacs 26.0.50, Org release_9.0.9-551-g92e8c8

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 194 bytes --]

Re: org mode moves to GNU emacs core

[Adonay Felipe Nogueira]

I have just thought of some more advantages and disadvantages too.

* Strengths

** One less split in development effort

This one explains itself. Also, for a related article against "forking",
"bundling", and "reinventing the wheel" practices, see
[[https://wingolog.org/archives/2015/11/09/embracing-conways-law]].

* Threats

** People who still refuse to contribute to GNU Emacs for some reason

I have seem people making completely separated projects from GNU Emacs
due to concerns of being required to assign copyright to the
FSF. However, first and foremost, I must add that I'm in favor of doing
such assignment, because FSF is a charity organization --- not just a
non-profit ---, and they, together with other organizations, do
community-oriented copyleft enforcement, not the
"average"/"extortion-based" copyleft enforcement.

One thing that I would like to test would be to ask the same people who
deny copyright assignments to the FSF: Would they assign copyright to
some organization if such organization were to offer them a 1 bi
job? Between the two scenarios, I don't see much difference, except that
FSF works for a free/libre and just/fair digital society, whereas the
"organization" works for its own interests and can kick all their
workers and these wouldn't be able to take their own copyrights/work
back.

-- 
- [[https://libreplanet.org/wiki/User:Adfeno]]
- Palestrante e consultor sobre /software/ livre (não confundir com
  gratis).
- "WhatsApp"? Ele não é livre, por isso não uso. Iguais a ele prefiro
  GNU Ring, ou Tox. Quer outras formas de contato? Adicione o vCard
  que está no endereço acima aos teus contatos.
- Pretende me enviar arquivos .doc, .ppt, .cdr, ou .mp3? OK, eu
  aceito, mas não repasso. Entrego apenas em formatos favoráveis ao
  /software/ livre. Favor entrar em contato em caso de dúvida.

Re: org mode moves to GNU emacs core

[Uwe Brauer]

>>> "qTim" == qTim Cross <theophilusx@gmail.com> writes:

   > Just to throw my 2 cents in.

   > 1. Problems with mixed versions. Currently, Emacs has org 8.x included
   > in the distribution. This is despite 9.x being out before the release of
   > 25.2. Something needs to be done to improve coordination and perhaps if
   > it was part of the core, this would be more likely. At any rate, the
   > current situation means you need to be very careful to ensure no org
   > feature is loaded before the ELPA package is loaded or you will get odd
   > behaviour and the symbol's value is void errors. 

   > 2. If you just want to load the ELPA version of org (not
   > org-plus-contrib) it can be a real pain. You have to play around with
   > package lists to ensure you actually get the right one. This can be a
   > real hassle if you also use the use-package package as you will often
   > get the older version bundled with Emacs if you don't have your package
   > lists in the right order.

   > 3. I would really like to see two completely separate packages rather
   > than having org and org-plus-contrib. Currently, if you have packages
   > which have org as a dependency and you have loaded org-plus-contrib
   > rather than just org, you will end up with both. Not a big issue, unless
   > your on a slow link as now you will download updates for both org and
   > org-plus-contrib. (there is no 'cleverness' with ELPA dependency
   > specifications - you cannot specify alternative dependencies).

But this critics could be applied to any emacs package and therefore to
the package system itself.

   > A lot will depend on when org becomes part f core. The trick will be to
   > do it once development of org slows down. I've been using org for a long
   > time now and have noticed that the rate of new features being added has
   > slowed down. Much of the changes now is about improvement and refinement
   > of the code base. I would imagine that at some point, things will become
   > even more stable with fewer releases. This would be the point at which
   > it would make sense to bring into core.

   > The other advantage of being part of core is that updates and changes to
   > Emacs will be integrated into org much better. We won't see situations
   > where new versions of Emacs require a rush to update org. for the end
   > user, this should create a much more stable org environment.

Well I update GNU emacs every 6 months it is not difficult but needs
considerable longer to compile and install than org mode.

   > Then of course, there will always be the option to run org straight from
   > the git repository for those who really want the latest version. I find
   > that once you have the path added to load-path, running from the git
   > repo is not much more effort than installing the latest ELPA package. 

I don't see how that would possible once it is integrated in GNU emacs
core, there will be no separate makefile or anything of that sort, but
maybe I am missing something.

Uwe 

Re: org mode moves to GNU emacs core

[Robert Horn]

Bastien Guerry writes:

> Hi Philip,
>
> phillip.lord@russet.org.uk (Phillip Lord) writes:
>
>> I presume you do see this as an advantage? The issue is, surely,
>> that it's too much of a PITA for the advantage that you gain?
>
> Well, it's not really about PITA-or-not-PITA, it's just that I want
> org-mode to be the default mode for some files in Emacs, and having
> org-mode in Emacs' core is the most simple way to go for this.
>

I just did a quick check of my git repositories for org-mode and emacs.
There is a significant difference in release cycle policies, and this
will affect users.  Emacs makes a release about once every 9 months,
usually a point release.  Major feature releases are less frequent.
Org-mode makes a release about once per month, also usually a point
release.

I think that switching to the emacs cycle would be perceived as making
org-mode far less responsive to problem reports and feature improvements.

There are ways that the git repositories and release policies can be
organized to enable more rapid response to minor bugs and small features
while still integrating into core emacs.  I think that you should figure
out a mutually acceptable means of maintaining the present rapid
responsiveness.  With a suitable structuring of make files, etc., you
can probably also deal with the performance issues associated with
building updated versions.  The emacs maintainers would have to agree.

It does call for a little more setup work, and probably a semi-permanent
branch structure in git to allow for org updates, while gaining most of
what you want.

It would also mean that those who want to stay on the leading edge of
org-mode would need to maintain git synchronization with emacs rather
than org-mode.  With good explanation and documentation that shouldn't
be too much of a problem.  I do it already on an ad-hoc basis because I
found elpa to be too problematic.

R Horn

Re: org mode moves to GNU emacs core

[Adonay Felipe Nogueira]

Also, all the issues related to updates and compatibility would be
solved if more poeple used GNU Guix package manager. This way, GNU Emacs
can integrate Org in its core, and use the "1 release per 9 months"
cycle, and also could have package recipes pointing to arbitrary commits
if there is a critical feature or critical bug.

Also, GNU Guix users can make and use their own package recipes with
ease.

Lastly, now a bit more technical: Among other things, GNU Guix project
tries to make package recipes in such a way that the package recipe
being created or edited avoids depending on specific versions of other
packages.

Re: org mode moves to GNU emacs core

[Nick Dokos]

Adonay Felipe Nogueira <adfeno@openmailbox.org> writes:

> I have just thought of some more advantages and disadvantages too.
>
> * Strengths
>
> ** One less split in development effort
>
> This one explains itself. Also, for a related article against "forking",
> "bundling", and "reinventing the wheel" practices, see
> [[https://wingolog.org/archives/2015/11/09/embracing-conways-law]].
>
> * Threats
>
> ** People who still refuse to contribute to GNU Emacs for some reason
>
> I have seem people making completely separated projects from GNU Emacs
> due to concerns of being required to assign copyright to the
> FSF. However, first and foremost, I must add that I'm in favor of doing
> such assignment, because FSF is a charity organization --- not just a
> non-profit ---, and they, together with other organizations, do
> community-oriented copyleft enforcement, not the
> "average"/"extortion-based" copyleft enforcement.
>

Whoa: some of what you describe exists (see e.g.
https://sfconservancy.org/blog/2016/jul/19/patrick-mchardy-gpl-enforcement/),
but it is a distant outlier. Calling it "average" is as inaccurate as
calling Pele an "average" footballer.

-- 
Nick

Re: org mode moves to GNU emacs core

[Adonay Felipe Nogueira]

Indeed I admit that I might have exagerated when I called it
"average". Please excuse my wrong doing. :)

Re: org mode moves to GNU emacs core

[Tim Cross]

Uwe Brauer writes:

>>>> "qTim" == qTim Cross <theophilusx@gmail.com> writes:
>
>    > Just to throw my 2 cents in.
>
>    > 1. Problems with mixed versions. Currently, Emacs has org 8.x included
>    > in the distribution. This is despite 9.x being out before the release of
>    > 25.2. Something needs to be done to improve coordination and perhaps if
>    > it was part of the core, this would be more likely. At any rate, the
>    > current situation means you need to be very careful to ensure no org
>    > feature is loaded before the ELPA package is loaded or you will get odd
>    > behaviour and the symbol's value is void errors.
>
>    > 2. If you just want to load the ELPA version of org (not
>    > org-plus-contrib) it can be a real pain. You have to play around with
>    > package lists to ensure you actually get the right one. This can be a
>    > real hassle if you also use the use-package package as you will often
>    > get the older version bundled with Emacs if you don't have your package
>    > lists in the right order.
>
>    > 3. I would really like to see two completely separate packages rather
>    > than having org and org-plus-contrib. Currently, if you have packages
>    > which have org as a dependency and you have loaded org-plus-contrib
>    > rather than just org, you will end up with both. Not a big issue, unless
>    > your on a slow link as now you will download updates for both org and
>    > org-plus-contrib. (there is no 'cleverness' with ELPA dependency
>    > specifications - you cannot specify alternative dependencies).
>
> But this critics could be applied to any emacs package and therefore to
> the package system itself.
>
Yes. It is a weakness of the package system not org. However, while I
have quite a few packages installed, org is the only one where I have
two versions installed at once.

>    > A lot will depend on when org becomes part f core. The trick will be to
>    > do it once development of org slows down. I've been using org for a long
>    > time now and have noticed that the rate of new features being added has
>    > slowed down. Much of the changes now is about improvement and refinement
>    > of the code base. I would imagine that at some point, things will become
>    > even more stable with fewer releases. This would be the point at which
>    > it would make sense to bring into core.
>
>    > The other advantage of being part of core is that updates and changes to
>    > Emacs will be integrated into org much better. We won't see situations
>    > where new versions of Emacs require a rush to update org. for the end
>    > user, this should create a much more stable org environment.
>
> Well I update GNU emacs every 6 months it is not difficult but needs
> considerable longer to compile and install than org mode.
>

I use to do that - actually, probably even more frequently. I still run
from my own build rather than distribution versions because they are
even further behind (generally). However, these days, I just run the
most recent Emacs release i.e. now 25.2

>    > Then of course, there will always be the option to run org straight from
>    > the git repository for those who really want the latest version. I find
>    > that once you have the path added to load-path, running from the git
>    > repo is not much more effort than installing the latest ELPA package.
>
> I don't see how that would possible once it is integrated in GNU emacs
> core, there will be no separate makefile or anything of that sort, but
> maybe I am missing something.
>

There is going to have to be a way for people to maintain and build org
independently. When you are maintaining Emacs, you don't want to have to
re-build the whole system every time you create a change. What you tend
to find is there are multiple Makefiles with a top level Emacs makefile
which calls sub-level makefiles as part of the build. It may be
necessary to modify configure or add a new option to build outside the
emacs tree, but that shouldn't be too difficult.

I should emphasise that while I agree org would be good as part of
Emacs' core, I don't think this should occur until org change velocity
has stabilised to a point where change velocity is lower than it is
now. At that point, there will be much less need to be running the most
recent snapshot.

Maybe my experience is very different. However, I found a lot more
motivation to go from org 7.x to 8.x than I did from 8.x to 9.x. In
fact, the only visible changes in 9.x I've had to deal with have been
about compatibility changes and minor bugs I've had to update
for/fix. I could still be running 8.x. The only reason I've updated to
9.x is to avoid issues with some of the contrib packages that have/may
have been updated to work with 9.x

The other point to keep in mind is that this change won't happen
quickly. It will take some time before this can occur and we probably
need to start thinking/talking about it now so that when the right time
arrives, things can move forward. 


--
Tim Cross

Re: org mode moves to GNU emacs core

[Bastien Guerry]

Hi Robert,

first of all, my bad: what I should have said in all these discussion
is that any decision regarding moving Org to Emacs' core won't happen
any time soon (I'd say two or three years).

Keeping Emacs master branch in sync with Org maint branch is not a
problem anymore, so the decision of whether Org should go into Emacs
core will not depend on this.

And by the time the decision will be made, I expect Emacs and Org
release cycles will come close.  In other words: when Org will be as
stable as Emacs and when most of Org's development will happen in its
external modules, it will be time to move Org's development to Emacs.

But all this is very hypothetical, right now very much "bikeshed" :)

Best,

-- 
 Bastien

Re: org mode moves to GNU emacs core

[Phillip Lord]

Bastien Guerry <bzg@gnu.org> writes:

> Hi Philip,
>
> phillip.lord@russet.org.uk (Phillip Lord) writes:
>
>> I presume you do see this as an advantage? The issue is, surely,
>> that it's too much of a PITA for the advantage that you gain?
>
> Well, it's not really about PITA-or-not-PITA, it's just that I want
> org-mode to be the default mode for some files in Emacs, and having
> org-mode in Emacs' core is the most simple way to go for this.
>
> Maintainers of projects like Gnus or CEDET don't want their code to
> live outside of Emacs repo neither, so I guess simplicity is a big
> win.


Yes, it would appear that way. But, at the moment, to be distributed
with Emacs you have to be in the emacs repo. We've discussed this
before, I know, but these two things are not an absolute
necessity. Having the Emacs build populate it's source tree with
packages is also possible.

Phil

Re: org mode moves to GNU emacs core

[Phillip Lord]

Tim Cross <theophilusx@gmail.com> writes:

> Just to throw my 2 cents in.
>
> While I can understand the benefits of being able to easily install the
> latest org package via elpa, I think there are some significant benefits
> to org being a part of core Emacs.
>
> I currently find three issues with the current situation which may be
> somewhat resolved if org was part of core emacs.
>
> 1. Problems with mixed versions. Currently, Emacs has org 8.x included
> in the distribution. This is despite 9.x being out before the release of
> 25.2. Something needs to be done to improve coordination and perhaps if
> it was part of the core, this would be more likely. At any rate, the
> current situation means you need to be very careful to ensure no org
> feature is loaded before the ELPA package is loaded or you will get odd
> behaviour and the symbol's value is void errors. 


I've argued on emacs-devel about this. The problem here is that the
distributed org-mode remains in the load path, so you are totally
dependent on load path shadowing to ensure the right package gets
loaded. This doesn't happen with ELPA packages since only the latest
gets added to the load path.

My solution to this would be to have Emacs use package.el to load
files in core as well as elsewhere. Then, when you installed org from
ELPA, package.el would remove the core installed files from the
load-path (or rather never add them). You'd need to restart Emacs after
installation, but otherwise the problem goes away.

Everybody else thought this was a bad idea!

Phil

Re: org mode moves to GNU emacs core

[Phillip Lord]

Tim Cross <theophilusx@gmail.com> writes:

>> I don't see how that would possible once it is integrated in GNU emacs
>> core, there will be no separate makefile or anything of that sort, but
>> maybe I am missing something.
>>
>
> There is going to have to be a way for people to maintain and build org
> independently. When you are maintaining Emacs, you don't want to have to
> re-build the whole system every time you create a change. What you tend
> to find is there are multiple Makefiles with a top level Emacs makefile
> which calls sub-level makefiles as part of the build. It may be
> necessary to modify configure or add a new option to build outside the
> emacs tree, but that shouldn't be too difficult.

Emacs builds all its lisp with a single Makefile, but the build is
incremental. So the rebuild is very quick. To be honest, even if you
modify the C layer, the dump is pretty fast.


> I should emphasise that while I agree org would be good as part of
> Emacs' core, I don't think this should occur until org change velocity
> has stabilised to a point where change velocity is lower than it is
> now. At that point, there will be much less need to be running the most
> recent snapshot.
>
> Maybe my experience is very different. However, I found a lot more
> motivation to go from org 7.x to 8.x than I did from 8.x to 9.x. In
> fact, the only visible changes in 9.x I've had to deal with have been
> about compatibility changes and minor bugs I've had to update
> for/fix. I could still be running 8.x. The only reason I've updated to
> 9.x is to avoid issues with some of the contrib packages that have/may
> have been updated to work with 9.x


This makes the assumption that all of org moves at the same speed. It
seems to me that the things currently in contrib still move fairly fast.

Phil