[DEV] Bump Emacs requirement to 24.4?

[DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Hello,

Just to be sure, can we require Emacs 24.4 for development version
(a.k.a. Org 8.4)? As a data point, Debian stable provides it.

Also, what is the status of XEmacs support? AFAIU Org 8.3 doesn't build
on XEmacs but no one is complaining. We may as well drop it and ignore
most of "org-compat.el".

WDYT?


Regards,

-- 
Nicolas Goaziou                                                0x80A93738

Re: [DEV] Bump Emacs requirement to 24.4?

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 516 bytes --]

+1

On Wed, Aug 5, 2015 at 3:38 PM Nicolas Goaziou <mail@nicolasgoaziou.fr>
wrote:

> Hello,
>
> Just to be sure, can we require Emacs 24.4 for development version
> (a.k.a. Org 8.4)? As a data point, Debian stable provides it.
>
> Also, what is the status of XEmacs support? AFAIU Org 8.3 doesn't build
> on XEmacs but no one is complaining. We may as well drop it and ignore
> most of "org-compat.el".
>
> WDYT?
>
>
> Regards,
>
> --
> Nicolas Goaziou                                                0x80A93738
>
>

[-- Attachment #2: Type: text/html, Size: 858 bytes --]

Re: [DEV] Bump Emacs requirement to 24.4?

[Suvayu Ali]

On Wed, Aug 05, 2015 at 09:39:19PM +0200, Nicolas Goaziou wrote:
> Hello,
> 
> Just to be sure, can we require Emacs 24.4 for development version
> (a.k.a. Org 8.4)? As a data point, Debian stable provides it.
> 
> Also, what is the status of XEmacs support? AFAIU Org 8.3 doesn't build
> on XEmacs but no one is complaining. We may as well drop it and ignore
> most of "org-compat.el".

Agreed to both :).

-- 
Suvayu

Open source is the future. It sets us free.

Re: [DEV] Bump Emacs requirement to 24.4?

[Rasmus]

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> Just to be sure, can we require Emacs 24.4 for development version
> (a.k.a. Org 8.4)? As a data point, Debian stable provides it.


Are the any differences between 24.4 and 24.5 that we care about?

wrt the distros I care about:

- Arch and Fedora are on both on 24.5.
- CentOS7 is still 24.3.  So is Slackware.

Rasmus

-- 
m-mm-mmm-mmmm bacon!

Re: [DEV] Bump Emacs requirement to 24.4?

[Robert Klein]

On Wed, 05 Aug 2015 23:51:32 +0200
Rasmus <rasmus@gmx.us> wrote:

> Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:
> 
> > Just to be sure, can we require Emacs 24.4 for development version
> > (a.k.a. Org 8.4)? As a data point, Debian stable provides it.
> 
> 
> Are the any differences between 24.4 and 24.5 that we care about?
> 
> wrt the distros I care about:
> 
> - Arch and Fedora are on both on 24.5.
> - CentOS7 is still 24.3.  So is Slackware.

 SUSE Linux Enterprise 12 and openSUSE 13.2 also have 24.3.  The
 package for future openSUSE:42 is 24.3, too.

Best regards
Robert

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Robert Klein <RoKlein@roklein.de> writes:

> On Wed, 05 Aug 2015 23:51:32 +0200
> Rasmus <rasmus@gmx.us> wrote:
>
>> Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:
>> 
>> > Just to be sure, can we require Emacs 24.4 for development version
>> > (a.k.a. Org 8.4)? As a data point, Debian stable provides it.
>> 
>> 
>> Are the any differences between 24.4 and 24.5 that we care about?
>> 
>> wrt the distros I care about:
>> 
>> - Arch and Fedora are on both on 24.5.
>> - CentOS7 is still 24.3.  So is Slackware.
>
>  SUSE Linux Enterprise 12 and openSUSE 13.2 also have 24.3.  The
>  package for future openSUSE:42 is 24.3, too.

Bumping requirements to Emacs 24.3 instead of 24.4 is no biggie. 24.4 is
just icing on the cake.

Regards,

Re: [DEV] Bump Emacs requirement to 24.4?

[Robert Klein]

On Thu, 06 Aug 2015 10:03:17 +0200
Nicolas Goaziou <mail@nicolasgoaziou.fr> wrote:

> Robert Klein <RoKlein@roklein.de> writes:
> 
> > On Wed, 05 Aug 2015 23:51:32 +0200
> > Rasmus <rasmus@gmx.us> wrote:
> >
> >> Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:
> >> 
> >> > Just to be sure, can we require Emacs 24.4 for development
> >> > version (a.k.a. Org 8.4)? As a data point, Debian stable
> >> > provides it.
> >> 
> >> 
> >> Are the any differences between 24.4 and 24.5 that we care about?
> >> 
> >> wrt the distros I care about:
> >> 
> >> - Arch and Fedora are on both on 24.5.
> >> - CentOS7 is still 24.3.  So is Slackware.
> >
> >  SUSE Linux Enterprise 12 and openSUSE 13.2 also have 24.3.  The
> >  package for future openSUSE:42 is 24.3, too.
> 
> Bumping requirements to Emacs 24.3 instead of 24.4 is no biggie. 24.4
> is just icing on the cake.


On second thought, go for it.  Org 8.3 runs on Emacs 24.3 and if I want
a newer Org I can also install my own Emacs.  (Just checked: I already
do so, though I forgot the reason. )

Best regards
Robert

Re: [DEV] Bump Emacs requirement to 24.4?

[Rasmus]

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> Bumping requirements to Emacs 24.3 instead of 24.4 is no biggie. 24.4 is
> just icing on the cake.

I don't care about 24.{3,4,5}.  In the end, on a very pragmatic level it's
a question of which machines are shown on David's buildbot...

Personally, I'd go for 24.5.  Perhaps that's too egoistic.

Rasmus

-- 
Tack, ni svenska vakttorn. Med plutonium tvingar vi dansken på knä!

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien Guerry]

Hi,

There is no rush for this change.

I will ask on the emacs-devel mailing list for advice on what seems
best to support.

Rasmus, can you revert the 24.3 requirement on org.el for now ?

Let's not make changes in the master branch that push the decision
when it is not taken yet.

Thanks,

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Hello,

Bastien Guerry <bzg@gnu.org> writes:

> There is no rush for this change.
>
> I will ask on the emacs-devel mailing list for advice on what seems
> best to support.
>
> Rasmus, can you revert the 24.3 requirement on org.el for now ?
>
> Let's not make changes in the master branch that push the decision
> when it is not taken yet.

Master branch is already 24.3+ (with compilation warnings, 24.4+ without
them): I committed "org-lint.el", moved some libraries to lexical
binding, etc.

Is there any reason to avoid requiring at least Emacs 24.3 in
development branch?


Regards,

-- 
Nicolas Goaziou

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien Guerry]

Hi Nicolas,

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> Master branch is already 24.3+ (with compilation warnings, 24.4+ without
> them): I committed "org-lint.el", moved some libraries to lexical
> binding, etc.

Yes, we should have waited for the decision to be taken for this.

> Is there any reason to avoid requiring at least Emacs 24.3 in
> development branch?

I don't know, but we have only little information on what impact such
a change will have.  Our best move for now is to ask on mailing lists,
collect information, then decide.

Also such a change needs to be advertized correctly, I'd rather have
it for Org 9.0 -- same for dropping XEmacs compatibility.

We don't need to complete the whole 8.x series and may jump to 9.0
soonish, but for the time being, I suggest you create a 9.0 branch
with the 24.3 requirement and changes that cannot go without it.

WDYT?

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Rasmus]

Bastien Guerry <bzg@gnu.org> writes:

> Hi Nicolas,
>
> Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:
>
>> Master branch is already 24.3+ (with compilation warnings, 24.4+ without
>> them): I committed "org-lint.el", moved some libraries to lexical
>> binding, etc.
>
> Yes, we should have waited for the decision to be taken for this.
>
>> Is there any reason to avoid requiring at least Emacs 24.3 in
>> development branch?
>
> I don't know, but we have only little information on what impact such
> a change will have.  Our best move for now is to ask on mailing lists,
> collect information, then decide.
>
> Also such a change needs to be advertized correctly, I'd rather have
> it for Org 9.0 -- same for dropping XEmacs compatibility.
>
> We don't need to complete the whole 8.x series and may jump to 9.0
> soonish, but for the time being, I suggest you create a 9.0 branch
> with the 24.3 requirement and changes that cannot go without it.
>
> WDYT?

But then what will master hold and what would be the point of it?  It
sounds like a way to rename "master" to "branch9.0"...

Also, Emacs 24.3 was released in March, 2013.  By the time the next Org is
released, it will be more than 3 years old.

Rasmus


-- 
This message is brought to you by the department of redundant departments

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien Guerry]

Rasmus <rasmus@gmx.us> writes:

> But then what will master hold and what would be the point of it?  It
> sounds like a way to rename "master" to "branch9.0"...

The point is to allow committing changes that depend on this
requirement without forcing the requirement on master.

We can also simply revert those change and wait for the decision
to be taken.  This is a matter of waiting ~10 days I'd say.

> Also, Emacs 24.3 was released in March, 2013.  By the time the next Org is
> released, it will be more than 3 years old.

Com'on :)

I'm back for good and don't plan to wait years between releases.
I wish we can release 8.4 at the end of August and 9.0 in October.

My simple point is: let's get more information and let's take a
proper decision.  Let's not force the change.

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Bastien Guerry <bzg@gnu.org> writes:

> We can also simply revert those change and wait for the decision
> to be taken.  This is a matter of waiting ~10 days I'd say.

Revert any commit you want to your heart's content. Note however, that
I deleted my corresponding local branches, so you may want to do some
back-up prior to this.

>> Also, Emacs 24.3 was released in March, 2013.  By the time the next Org is
>> released, it will be more than 3 years old.
>
> Com'on :)
>
> I'm back for good and don't plan to wait years between releases.
> I wish we can release 8.4 at the end of August and 9.0 in October.

I'm lost here. There is only one change so far in development branch
that does not require Emacs 24.3+. What do you want to put in 8.4 that
cannot fit in 8.3.2 or 8.3.3?

Could you explain a bit what naming scheme you have in mind?

> My simple point is: let's get more information and let's take a
> proper decision.  Let's not force the change.

It's still not clear what information you're after. Switching to lexical
binding, counting on cl-lib.el, and so on, is the way to go. What do you
expect in delaying this change?


Regards,

Re: [DEV] Bump Emacs requirement to 24.4?

[Rasmus]

Bastien Guerry <bzg@gnu.org> writes:

> We can also simply revert those change and wait for the decision
> to be taken.  This is a matter of waiting ~10 days I'd say.
>
>> Also, Emacs 24.3 was released in March, 2013.  By the time the next Org is
>> released, it will be more than 3 years old.
>
> Com'on :)
>
> I'm back for good and don't plan to wait years between releases.
> I wish we can release 8.4 at the end of August and 9.0 in October.

It was not meant as 'finger-pointing'.  Slow releases are not necessarily
a bad thing.  Org is very important to the work of some folks, so a slow
cycle might be less disruptive.  OTOH, so could frequent releases...  I
don't know, really.

Achim Gratz <Stromeko@nexgo.de> writes:

> Nicolas Goaziou writes:
>> Just to be sure, can we require Emacs 24.4 for development version
>> (a.k.a. Org 8.4)? As a data point, Debian stable provides it.
>
> Debian Squeeze LTS or whatever they call it doesn't w/o backports.
> RHEL6 doesn't have it w/o epel (RHEL7 has 24.3 IIRC).
> RaspberryPi doesn't have it.

I have Emacs 24.4 on my Debian Wheezy which is pretty old.  Can ELPA serve
different versions based on the client?

> I'm still falling over Emacs 22 in various forms and Emacs 23, where it
> is standard is not always at the latest version (23.4).

If ELPA can (0) be used with Emacs-23 and (1) ELPA is smart enough to
serve the right version of Org, I don't think this is a problem.

>> Also, what is the status of XEmacs support? AFAIU Org 8.3 doesn't build
>> on XEmacs but no one is complaining. We may as well drop it and ignore
>> most of "org-compat.el".
>
> I've been doing a lot of this compat stuff, but I gave up since I
> couldn't get ERT to work on XEmacs.  Org did build (with lots of errors)
> until some point and it was at least superficially usable.  The two
> XEmacs users on this list have never responded to any requests for
> further testing.  So I guess that XEmacs can be considered
> unsupportable.

And THESE are the hidden dependency when targeting older Emacs.  We have
so many org-prefixed functions that have equivalents in cl-lib.

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:


>> We don't need to complete the whole 8.x series and may jump to 9.0
>> soonish, but for the time being, I suggest you create a 9.0 branch
>> with the 24.3 requirement and changes that cannot go without it.
>>
>> WDYT?
>
> I think it is a mistake. 
>
> Handling two development branches means people testing Org have to
> choose which branch to test. We don't have the manpower to waste testing
> capabilities like that.
>
> I also see no reason to write outdated code (e.g., new libraries without
> lexical binding) and update it later.

I agree strongly with the above.

Rasmus

-- 
Dobbelt-A

Re: [DEV] Bump Emacs requirement to 24.4?

[Robert Horn]

Bastien Guerry writes:

> My simple point is: let's get more information and let's take a
> proper decision.  Let's not force the change.

I took a look at what is presently supported by Red Hat, Ubuntu, and
OpenSuSE in their long term support releases.  The results:

emacs 23.1 (RHEL 6.0, Ubuntu 12.04 LTS)
emacs 24.3 (OpenSUSE 13.x, RHEL 7, Ubuntu 14.04 LTS)
emacs 24.5 (OpenSUSE latest rolling, SUSE SLE-12)

In my opinion the goal of Org should be to run on the oldest version of
emacs that includes features needed by Org.  The long term support
versions are indicative of what we should expect from the
non-experimental users who just need Org to work.  They are not
exploratory development users.

Since three major distributions are still at emacs 24.3 for their most
recent long term support versions, that argues for Org not going beyond
24.3.  It's reasonable to expect the non-experimental not bleeding edge
users to be one the most recent long term support version.

-- 
Sent with my mu4e

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Hello,

Bastien Guerry <bzg@gnu.org> writes:

> Yes, we should have waited for the decision to be taken for this.

AFAIU, the decision was between 24.3 and 24.4. It isn't clear to me that
there was an hesitation about dropping Emacs 23 support.

> I don't know, but we have only little information on what impact such
> a change will have.  Our best move for now is to ask on mailing lists,
> collect information, then decide.

What impact do you think it can have? Users willing to use development
branch need Emacs 24.{3,4}. That's about it.

> Also such a change needs to be advertized correctly,

This is orthogonal to the issue. Once the desired minimal version is
settled, we will notify it in ORG-NEWS.

> I'd rather have it for Org 9.0 -- same for dropping XEmacs
> compatibility.

Your call.

> We don't need to complete the whole 8.x series and may jump to 9.0
> soonish, but for the time being, I suggest you create a 9.0 branch
> with the 24.3 requirement and changes that cannot go without it.
>
> WDYT?

I think it is a mistake. 

Handling two development branches means people testing Org have to
choose which branch to test. We don't have the manpower to waste testing
capabilities like that.

I also see no reason to write outdated code (e.g., new libraries without
lexical binding) and update it later.


Regards,

-- 
Nicolas Goaziou

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien Guerry]

Hi Nicolas,

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> AFAIU, the decision was between 24.3 and 24.4.

No.  You asked weather it was fine to require Emacs 24.4 in your
original message, then people pointed out that Emacs 24.3 was still
widely used, then you said it was fine to require Emacs 24.3, just
Emacs 24.4 was "icing on the cake".

So the questions are just :

- can we drop XEmacs support?  There seem to be a consensus about
  this (see Achim's message)

- what is the oldest version of Emacs that we should require for
  Org >8.3.1 ?  You proposed 24.4 or 24.3 and I want to check for
  possibly older versions.

FWIW, I'm all for Emacs 24.3, I just need time to think about it
and more information, such as what Robert just sent.

This is not a light decision, as we cannot move back, and I don't
see the need for rushing here.

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Bastien Guerry <bzg@gnu.org> writes:

> No.  You asked weather it was fine to require Emacs 24.4 in your
> original message, then people pointed out that Emacs 24.3 was still
> widely used, then you said it was fine to require Emacs 24.3, just
> Emacs 24.4 was "icing on the cake".

That's correct, excepted that you were the first to ask "why 24.4 and
not 24.3"? So I assumed it was between 24.3 and 24.4.

> This is not a light decision, as we cannot move back, and I don't
> see the need for rushing here.

There is no rushing, really. Emacs 24 was released in 2012, and lexical
binding is a huge step forward for developers. There is some serious
work involved to make the jump (along with cl-lib switch), so the sooner
we can start, the better.

Regards,

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien]

Hi Nicolas,

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

>> No.  You asked weather it was fine to require Emacs 24.4 in your
>> original message, then people pointed out that Emacs 24.3 was still
>> widely used, then you said it was fine to require Emacs 24.3, just
>> Emacs 24.4 was "icing on the cake".
>
> That's correct, excepted that you were the first to ask "why 24.4 and
> not 24.3"? So I assumed it was between 24.3 and 24.4.

OK, I understand.  I was just curious about why 24.4 and not 24.3,
since 24.4 does not add that much from a developer's point of view.

>> This is not a light decision, as we cannot move back, and I don't
>> see the need for rushing here.
>
> There is no rushing, really. Emacs 24 was released in 2012, and lexical
> binding is a huge step forward for developers. There is some serious
> work involved to make the jump (along with cl-lib switch), so the sooner
> we can start, the better.

I don't mean the rushing in general, I meaning the rushing between
the feedback you asked for and the move to commit things that assume
the answer has been given.

RMS replied to the question on emacs-devel:

  http://article.gmane.org/gmane.emacs.devel/188810

He suggests to continue supporting Emacs 23.

So my suggestion still stands:

- let's keep master in the current compatibility state since the
  question you asked still needs to be answer (it's just 10 days
  since it was asked).

- let's use a dedicated branch for commits requiring Emacs 24.3+.

Maybe we will end up distributing Org 8.3+ for Emacs 23.1 and
Org 9+ for Emacs 24.3+ -- I don't know.

Let's take 10-15 days to build a strategy together, with feedbacks
from various sources and information we share.  We need this time
to consider the question seriously.  Please revert the changes in
master that assume Org needs Emacs 24.3+.

Thanks,

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Achim Gratz]

Bastien writes:
> So my suggestion still stands:
>
> - let's keep master in the current compatibility state since the
>   question you asked still needs to be answer (it's just 10 days
>   since it was asked).
>
> - let's use a dedicated branch for commits requiring Emacs 24.3+.

I'm with Nicolas on this point: if we end up de-supporting Emacs 23,
then the master branch is where that should happen.  Since it already
has happened and the general consensus seems to be that we should go
there eventually, just maybe not right now, I don't see why we suddenly
also need to re-define what master is about.

> Maybe we will end up distributing Org 8.3+ for Emacs 23.1 and
> Org 9+ for Emacs 24.3+ -- I don't know.

What the next release from master will be called is not yet decided.
But we can and should decide that maint does not de-support Emacs 23
until <whatever version> and only after proper announcement.  That might
ultimately end up needing a separate maint23 branch for cherry-picking
some bug-fixes after the development has swicthed to Emacs24.3+ on
maint, although I don't see where the developer resources would come
from to keep three branches in flight, especially when apparently some
developers don't have Emacs23 at their disposal (actually that is what I
have the RaspberryPi for, believe it or not).

> Let's take 10-15 days to build a strategy together, with feedbacks
> from various sources and information we share.  We need this time
> to consider the question seriously.  Please revert the changes in
> master that assume Org needs Emacs 24.3+.

There's no need to rush that decision either.  I am one of those who
still need to run on Emacs23 and I will just switch from master to maint
for that.  I don't see any pressing bugs or missing features there
either, so if I'd need to freeze Emacs23 to some specific point in the
history that would quite likely also work.  From what happened with
Emacs 22, I expect at least another two years before I can generally
assume availability of Emacs 24.3+ on all systems.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

DIY Stuff:
http://Synth.Stromeko.net/DIY.html

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien Guerry]

Hi Achim,

Achim Gratz <Stromeko@nexgo.de> writes:

> Since it already
> has happened and the general consensus seems to be that we should go
> there eventually, just maybe not right now, I don't see why we suddenly
> also need to re-define what master is about.

The redefinition already happened (i.e. the master branch is about
Org+Emacs 24.3+) and it happened before we could reach a consensus
about it, or simply take the time to really discuss it as we need.

I'm trying to find the best conditions to move forward here.

If the repository is in a state that pushes us toward one decision,
these are not the best conditions IMO.

So if we revert the changes, that's fine.  If not, that's fine too,
but we will have to decide things as if the commits where not done.

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Achim Gratz]

Bastien Guerry writes:
> The redefinition already happened (i.e. the master branch is about
> Org+Emacs 24.3+) and it happened before we could reach a consensus
> about it, or simply take the time to really discuss it as we need.
>
> I'm trying to find the best conditions to move forward here.

Well, I'm not sure about achieving optimum, but I suggest that reverting
those commits isn't going to help anybody.

> If the repository is in a state that pushes us toward one decision,
> these are not the best conditions IMO.
>
> So if we revert the changes, that's fine.  If not, that's fine too,
> but we will have to decide things as if the commits where not done.

The decision for the current master branch was de-facto already taken.
There was certainly a bit of a miscommunication, but Nicolas did those
commits in good faith.  Nothing is lost so far since maint is still
Emacs 23 compatible.  As long as it is kept that way, you still have all
options open.  So it doesn't preclude any new decisions you want to
ponder on.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptations for Waldorf Q V3.00R3 and Q+ V3.54R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien]

Hi Nicolas and all,

As Achim said, we may have been miscommunicating, especially on
whether it was about requiring Emacs 23+ or 24.3 vs 24.4.

Here is my decision on this issue:

- the Org 8.x series will be Emacs 23+ compatible.

- the Org 9.x series will be Emacs 24.3+ compatible.

Emacs 23 and XEmacs support will be officially dropped as of Org 9.0.

The maint branch continues to be used to work on minor releases, as it
has always been used.

Instead of reverting changes from the master branch (you clearly don't
want to do that, and I don't either), I suggest we create a new branch
called "org-8-master" for Org 8.4+, and continue to use master for Org
9.x+ (i.e. "major major releases").

This is temporary: once the 9.0 version is released, we can simply use
maint and master as before, and delete org-8-master.

The reason for this is that we need to make room for new features in
the 8.x series, so that these new features will be available to the
Emacs 23 users.

If we drop Emacs 23 support as of Org 8.4, we won't be able to add new
features (e.g. new export backends) for Emacs 23.

I recognize having the manpower to watch after those branches might be
an issue, but we can overcome it by calling for careful testing before
major releases.

Let me know what you think,

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[David Engster]

Bastien writes:
> Emacs 23 and XEmacs support will be officially dropped as of Org 9.0.

Support for XEmacs should be dropped right away; it would just state a
fact, as Org didn't even compile with XEmacs for quite some time now
(and nobody complained).

> I recognize having the manpower to watch after those branches might be
> an issue, but we can overcome it by calling for careful testing before
> major releases.

At least the Buildbot can test the 8.3 branch with older Emacsen. I
already changed that master is only checked with 24.3+.

-David

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien]

Hi David,

David Engster <deng@randomsample.de> writes:

> Bastien writes:
>> Emacs 23 and XEmacs support will be officially dropped as of Org 9.0.
>
> Support for XEmacs should be dropped right away; it would just state a
> fact, as Org didn't even compile with XEmacs for quite some time now
> (and nobody complained).

Yes.  By "officially", I mean : "all the code related to XEmacs
compatibility will be removed from Org codebase".  Of course we can
declare that XEmacs is not supported right now, but removing this code
takes some time.

>> I recognize having the manpower to watch after those branches might be
>> an issue, but we can overcome it by calling for careful testing before
>> major releases.
>
> At least the Buildbot can test the 8.3 branch with older Emacsen. I
> already changed that master is only checked with 24.3+.

That's great, thanks.

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Suvayu Ali]

Hi Bastien,

On Wed, Aug 19, 2015 at 01:01:29AM +0200, Bastien wrote:
> 
> The maint branch continues to be used to work on minor releases, as it
> has always been used.
> 
> Instead of reverting changes from the master branch (you clearly don't
> want to do that, and I don't either), I suggest we create a new branch
> called "org-8-master" for Org 8.4+, and continue to use master for Org
> 9.x+ (i.e. "major major releases").

I'm slightly confused as-to why this special org-8-master branch is
needed.  As far as I understood from Achim's message (2nd paragraph):

  <http://mid.gmane.org/87twrzkbow.fsf@Rainer.invalid>,

that so far maint supports Emacs 23.  Although this might have been a
coincidence, without a conscious decision.  It will continue supporting
Emacs 23 until Org 9.  Post Org 9, if required, one can choose to start
a new maint branch called maint23 for continued Emacs 23 support.

Given the above, I do not understand the need for a separate
org-8-master branch.  Is this a gap in communication, or am I missing
something here?

Cheers,

-- 
Suvayu

Open source is the future. It sets us free.

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien]

Hi Suvayu,

Suvayu Ali <fatkasuvayu+linux@gmail.com> writes:

> I'm slightly confused as-to why this special org-8-master branch is
> needed.  As far as I understood from Achim's message (2nd paragraph):
>
>   <http://mid.gmane.org/87twrzkbow.fsf@Rainer.invalid>,
>
> that so far maint supports Emacs 23.

Yes, maint supports Emacs 23.

But maint is for minor release, and I want a branch where we can work
on major releases (8.4, 8.5, etc.) that support Emacs 23 too.

The clean way would have been to use drop-emacs23 branch for working
on code that does not support Emacs 23, and continue to use master for
working on major releases, but since nobody wants to revert the code
on master, there is no harm in using "org-8-master" as the new master.

> Although this might have been a
> coincidence, without a conscious decision.  It will continue supporting
> Emacs 23 until Org 9.  Post Org 9, if required, one can choose to start
> a new maint branch called maint23 for continued Emacs 23 support.
>
> Given the above, I do not understand the need for a separate
> org-8-master branch.  Is this a gap in communication, or am I missing
> something here?

I hope my answer clarifies things.

Thanks,

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Suvayu Ali]

Hi Bastien,

On Wed, Aug 19, 2015 at 11:40:09AM +0200, Bastien wrote:
> Suvayu Ali <fatkasuvayu+linux@gmail.com> writes:
> 
> > I'm slightly confused as-to why this special org-8-master branch is
> > needed.  As far as I understood from Achim's message (2nd paragraph):
> >
> >   <http://mid.gmane.org/87twrzkbow.fsf@Rainer.invalid>,
> >
> > that so far maint supports Emacs 23.
> 
> Yes, maint supports Emacs 23.
> 
> But maint is for minor release, and I want a branch where we can work
> on major releases (8.4, 8.5, etc.) that support Emacs 23 too.

I think this was the bit of information I was missing.

Thanks,

-- 
Suvayu

Open source is the future. It sets us free.

Re: [DEV] Bump Emacs requirement to 24.4?

[Achim Gratz]

Bastien writes:
> Here is my decision on this issue:
>
> - the Org 8.x series will be Emacs 23+ compatible.

…and we should maybe do an 8.4 final release before it is frozen, but
not drag it along furhter like you suggested in emacs-devel.

> - the Org 9.x series will be Emacs 24.3+ compatible.
>
> Emacs 23 and XEmacs support will be officially dropped as of Org 9.0.

Org on XEmacs is practically dead already, so why do we need to pretend
it is still supported?  You can check when I did the last compatibility
fixes for it, after that nobody (not even any XEmacs user) has ever made
any mention of it working or not working even though there were numerous
changes that are unlikely to work or work correctly on XEmacs.

> The maint branch continues to be used to work on minor releases, as it
> has always been used.
>
> Instead of reverting changes from the master branch (you clearly don't
> want to do that, and I don't either), I suggest we create a new branch
> called "org-8-master" for Org 8.4+, and continue to use master for Org
> 9.x+ (i.e. "major major releases").

I'm not sure how you want to implement this in practise, but it looks
too complicated and error-prone.  My suggestion would be to keep the
maint branch compatible with Emacs23 until 8.4 and then (maybe) split
off a maint-23 branch for any bugfixes.  If a (new) maintainer springs
into action to backport features into this branch, then fine, but
otherwise it stops being connected to the development branches.

> This is temporary: once the 9.0 version is released, we can simply use
> maint and master as before, and delete org-8-master.

You mustn't delete public branches.

> The reason for this is that we need to make room for new features in
> the 8.x series, so that these new features will be available to the
> Emacs 23 users.
>
> If we drop Emacs 23 support as of Org 8.4, we won't be able to add new
> features (e.g. new export backends) for Emacs 23.

I don't think we should.  Emacs23 does not get any new features, then why
should Org on Emacs23 do?

> I recognize having the manpower to watch after those branches might be
> an issue, but we can overcome it by calling for careful testing before
> major releases.

See the above proposal which minimizes that impact.  Call for a new
maintainer for maint-23 and see if someone volunteers.  Otherwise just
freeze that branch and backport only fixes for really bad issues.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf microQ V2.22R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Re: [DEV] Bump Emacs requirement to 24.4?

[Rasmus]

Bastien <bzg@gnu.org> writes:

> Hi Nicolas and all,
>
> As Achim said, we may have been miscommunicating, especially on
> whether it was about requiring Emacs 23+ or 24.3 vs 24.4.
>
> Here is my decision on this issue:
>
> - the Org 8.x series will be Emacs 23+ compatible.
>
> - the Org 9.x series will be Emacs 24.3+ compatible.
>
> [...]
> 
> Let me know what you think,

It would seem to me, that the natural conclusion from this is simply to
EOL Org 8+ and move on to v9, no?

Rasmus

-- 
Send from my Emacs

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien]

Hi Rasmus,

Rasmus <rasmus@gmx.us> writes:

> It would seem to me, that the natural conclusion from this is simply to
> EOL Org 8+ and move on to v9, no?

(I'm not sure what you mean by "EOL".)

And surely I've not been clear (see my answer to Suvayu too):
the whole purpose of this three-branches temporary move is to be
able to continue adding new features to the 8.x series, so that
Emacs-23 users can enjoy those features.

So the next major version will be 8.4, and Emacs 23 compatible.

This does not prevent from working on the master branch and
develop things that are Emacs 24.3+ compatible.

I hope it makes more sense.

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Rasmus]

Hi,

Two reasons to at least jump to 24.4:

    - nadvice.el: we can clean up all the advising in the code and
      advocate add-function.  This is already employed in upstream Emacs.

    - prettify-symbols-mode:  Perhaps some entity display code can be
      dropped in favor of prettify-symbols-mode?  I don't know if this is
      feasible.

Rasmus

-- 
May contains speling mistake

Re: [DEV] Bump Emacs requirement to 24.4?

[Bastien Guerry]

Hi Nicolas,

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> Just to be sure, can we require Emacs 24.4 for development version
> (a.k.a. Org 8.4)? As a data point, Debian stable provides it.

Before discussion 24.4 vs say 24.3... why?
What compatibility code is a burden today?

> Also, what is the status of XEmacs support? AFAIU Org 8.3 doesn't build
> on XEmacs but no one is complaining. We may as well drop it and ignore
> most of "org-compat.el".

I'm all for dropping XEmacs support.

-- 
 Bastien

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Bastien Guerry <bzg@gnu.org> writes:

> Before discussion 24.4 vs say 24.3... why?
> What compatibility code is a burden today?

24.3 is the minimum required to drop `org-link-escape-browser', which is
currently buggy.

24.4 fixes `split-string' and introduces many changes at the lisp level
(variadic functions...).

24.5, OTOH, is very boring.


Regards,

Re: [DEV] Bump Emacs requirement to 24.4?

[Achim Gratz]

Nicolas Goaziou writes:
> Just to be sure, can we require Emacs 24.4 for development version
> (a.k.a. Org 8.4)? As a data point, Debian stable provides it.

Debian Squeeze LTS or whatever they call it doesn't w/o backports.
RHEL6 doesn't have it w/o epel (RHEL7 has 24.3 IIRC).
RaspberryPi doesn't have it.

I'm still falling over Emacs 22 in various forms and Emacs 23, where it
is standard is not always at the latest version (23.4).

> Also, what is the status of XEmacs support? AFAIU Org 8.3 doesn't build
> on XEmacs but no one is complaining. We may as well drop it and ignore
> most of "org-compat.el".

I've been doing a lot of this compat stuff, but I gave up since I
couldn't get ERT to work on XEmacs.  Org did build (with lots of errors)
until some point and it was at least superficially usable.  The two
XEmacs users on this list have never responded to any requests for
further testing.  So I guess that XEmacs can be considered
unsupportable.

If you were hoping to get rid of compat, then have a look at the warnings
on Emacs 25.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf Q+, Q and microQ:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds

Re: [DEV] Bump Emacs requirement to 24.4?

[Nicolas Goaziou]

Hello,

Achim Gratz <Stromeko@nexgo.de> writes:

> Debian Squeeze LTS or whatever they call it doesn't w/o backports.
> RHEL6 doesn't have it w/o epel (RHEL7 has 24.3 IIRC).
> RaspberryPi doesn't have it.

I don't think we have a lot of Org users on RaspberryPi.

> I'm still falling over Emacs 22 in various forms and Emacs 23, where it
> is standard is not always at the latest version (23.4).

So what? If you have to deal with Emacs 23, you can also live with Org
8.3. It is by no mean a sub-par release. It is also brand new.

> I've been doing a lot of this compat stuff, but I gave up since I
> couldn't get ERT to work on XEmacs.  Org did build (with lots of errors)
> until some point and it was at least superficially usable.  The two
> XEmacs users on this list have never responded to any requests for
> further testing.  So I guess that XEmacs can be considered
> unsupportable.

Unfortunately, this situation is not very good. We have compatibility
functions all over the place for zero benefit.

> If you were hoping to get rid of compat, then have a look at the warnings
> on Emacs 25.

I can only handle one fight at a time.


Regards,

-- 
Nicolas Goaziou