Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Gary Oberbrunner
Ha, I can hardly call myself a "core activist" anymore, but thanks. :-)

On Fri, Apr 21, 2017 at 1:28 PM, Russel Winder  wrote:

> On Fri, 2017-04-21 at 12:55 -0400, Gary Oberbrunner wrote:
> >
> […]
> >
> > I would be so happy about this. I would gladly volunteer to help. I'm
> > all
> > git+github for literally everything else, home & work.
>
> I think I have been tainted by Apache do-ocracy from association with
> Apache Groovy: to those that do is given the power to decide.
>
> Given that the two core activists want change form Mercurial to Git,
> then that change should happen.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>


-- 
Gary
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Russel Winder
On Fri, 2017-04-21 at 12:55 -0400, Gary Oberbrunner wrote:
> 
[…]
> 
> I would be so happy about this. I would gladly volunteer to help. I'm
> all
> git+github for literally everything else, home & work.

I think I have been tainted by Apache do-ocracy from association with
Apache Groovy: to those that do is given the power to decide.

Given that the two core activists want change form Mercurial to Git,
then that change should happen.

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Russel Winder
On Fri, 2017-04-21 at 09:41 -0700, Bill Deegan wrote:
> Russel,
> 
> At present I'm spending all my available energy and time on getting
> py2+py3
> working.
> Once that is done, then we can look at migrating the bug tracker, but
> not
> before.

Exactly, getting the Python 3 stuff finished clearly takes higher
priority than anything just now, it is the current sprint. I'll see if
I can take a look at more this weekend.

Next week is ACCU so I'll likely be unable to do anything constructive
on any project.  https://conference.accu.org/site/

> That said, I'm thinking post 3 a migration to github could have some
> real
> value.
> (And git, I'm tired of using HG on only this project)

I would find workflow very much easier using Git rather than Mercurial
for this project. The issue for me is in-repository transient branches
for short-lived feature branches.  With my knowledge of Mercurial,
whilst nice for individual use, is a pain for managing a mainline clone
and multiple feature branches.

I am happy using BitBucket, but would also be happy with GitHub, or
even GitLab.

> Also note that buildbot (our current ci http://buildbot.scons.org/ )
> is
> capable of pushing ci status to bitbucket (and github as
> well).  Travis-CI
> builds work for plain python builds, but very unlikely the
> VM/Containers
> the tests are run in will have many of the installed tools we use in
> many
> tests, and also most likely not have any MSVS versions installed..

I think this discussion should switch to an own thread. The advantage
of Travis-CI if we can get away with it is having various OS
installations. For the D tools, it is possible to create a build matrix
for all eight variations of D compiler.

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Bill Deegan
I guess the only question is what to do with the outstanding pull requests..

If we can leave them on bitbucket and migrate them after, then I could be
convinced to do the transition before 3.0..
(or as part of 3.0 release)

-Bill

On Fri, Apr 21, 2017 at 9:55 AM, Gary Oberbrunner 
wrote:

>
> On Fri, Apr 21, 2017 at 12:41 PM, Bill Deegan 
> wrote:
>
>> That said, I'm thinking post 3 a migration to github could have some real
>> value.
>> (And git, I'm tired of using HG on only this project)
>
>
> I would be so happy about this. I would gladly volunteer to help. I'm all
> git+github for literally everything else, home & work.
>
>
> --
> Gary
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Gary Oberbrunner
On Fri, Apr 21, 2017 at 12:41 PM, Bill Deegan 
wrote:

> That said, I'm thinking post 3 a migration to github could have some real
> value.
> (And git, I'm tired of using HG on only this project)


I would be so happy about this. I would gladly volunteer to help. I'm all
git+github for literally everything else, home & work.


-- 
Gary
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Bill Deegan
Russel,

At present I'm spending all my available energy and time on getting py2+py3
working.
Once that is done, then we can look at migrating the bug tracker, but not
before.

That said, I'm thinking post 3 a migration to github could have some real
value.
(And git, I'm tired of using HG on only this project)

Also note that buildbot (our current ci http://buildbot.scons.org/ ) is
capable of pushing ci status to bitbucket (and github as well).  Travis-CI
builds work for plain python builds, but very unlikely the VM/Containers
the tests are run in will have many of the installed tools we use in many
tests, and also most likely not have any MSVS versions installed..

-Bill

On Fri, Apr 21, 2017 at 2:57 AM, Russel Winder  wrote:

> On Wed, 2017-04-19 at 10:42 -0700, Bill Deegan wrote:
> > I appreciate your point of view.
>
> I think we all agree on the principle, but the details are getting in
> the way of practicality.
>
> > But for many years now it has been separate.
>
> And I suspect only the stalwarts and old hands even go there.
>
> > We will migrate at some point.
>
> I have to say, at this point in time, I have some doubts. I would like
> to be proved wrong.
>
> I think we have had two goes round this, and in the end no-one has had
> the energy to push past all the trivial niggles that people put forward
> to try and stiffle the move.
>
> > For the time being, it's really not such a huge burden to login to
> > tigris
> > to file a bug.
>
> For me arguably not. For D users used to GitHub, I suspect it may be. I
> can imagine them saying "The repo is on GitHub why the  do I have
> to go to some unheard of closing down system, sign up and login, to
> submit bug reports." Even were I to move the repo to BitBucket – and if
> I can work out how to get the equivalent of Travis-CI working from a
> Git repository on BitBucket, I'm happy to do that – the same argument
> regarding issues and Tigris would apply. I will though look to move to
> BitBucket on the assumption SCons is staying there – which seems
> sensible, at least in the short term.
>
> Clearly no-one has the energy or resource to be tackling the "transfer
> the issue handler" just now. But I can imagine the very act of
> suggesting we get stuck into this will bring up the arguments about how
> the BitBucket issue handler is rubbish, and all actual progress will be
> stiffled. What is needed here is action ignoring the words. Someone
> already had at least part of a "get stuff out of Tigris and put it in
> BitBucket". I suggest we update it, do any further tweaks, run it,
> switch to BitBucket and leave ourselves open to a better workflow. And
> ignore the moans from people who like the Tigris system.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-21 Thread Russel Winder
On Wed, 2017-04-19 at 10:42 -0700, Bill Deegan wrote:
> I appreciate your point of view.

I think we all agree on the principle, but the details are getting in
the way of practicality.

> But for many years now it has been separate.

And I suspect only the stalwarts and old hands even go there.

> We will migrate at some point.

I have to say, at this point in time, I have some doubts. I would like
to be proved wrong.

I think we have had two goes round this, and in the end no-one has had
the energy to push past all the trivial niggles that people put forward
to try and stiffle the move.

> For the time being, it's really not such a huge burden to login to
> tigris
> to file a bug.

For me arguably not. For D users used to GitHub, I suspect it may be. I
can imagine them saying "The repo is on GitHub why the  do I have
to go to some unheard of closing down system, sign up and login, to
submit bug reports." Even were I to move the repo to BitBucket – and if
I can work out how to get the equivalent of Travis-CI working from a
Git repository on BitBucket, I'm happy to do that – the same argument
regarding issues and Tigris would apply. I will though look to move to
BitBucket on the assumption SCons is staying there – which seems
sensible, at least in the short term.

Clearly no-one has the energy or resource to be tackling the "transfer
the issue handler" just now. But I can imagine the very act of
suggesting we get stuck into this will bring up the arguments about how
the BitBucket issue handler is rubbish, and all actual progress will be
stiffled. What is needed here is action ignoring the words. Someone
already had at least part of a "get stuff out of Tigris and put it in
BitBucket". I suggest we update it, do any further tweaks, run it,
switch to BitBucket and leave ourselves open to a better workflow. And
ignore the moans from people who like the Tigris system.

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-19 Thread Bill Deegan
I appreciate your point of view.
But for many years now it has been separate.
We will migrate at some point.

For the time being, it's really not such a huge burden to login to tigris
to file a bug.
-Bill

On Tue, Apr 18, 2017 at 10:33 PM, Russel Winder 
wrote:

> On Tue, 2017-04-18 at 09:23 -0700, Bill Deegan wrote:
> > I think reporting bugs for one part of SCons in a different
> > repository is
> > not a good idea.
>
> I do not disagree. But having the codebase on BitBucket and the issue
> handler on Tigris is a dreadful set up in my view. If the issue handler
> was on BitBucket life would be a lot easier. However, every time this
> topic comes up we end up at "someone insists that the project cannot
> survive without the entire history, and no-one can be bothered to dothe
> transfer" so nothing happens.
>
> Of course Tigris looks like it is going away by the look of the sign
> saying "no more projects". If that happens and nothing is done, history
> gets lost anyway.
>
> I'd suggest opening up the issue tracker on the SCons repository on
> BitBucket sooner rather than later.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-18 Thread Russel Winder
On Tue, 2017-04-18 at 09:23 -0700, Bill Deegan wrote:
> I think reporting bugs for one part of SCons in a different
> repository is
> not a good idea.

I do not disagree. But having the codebase on BitBucket and the issue
handler on Tigris is a dreadful set up in my view. If the issue handler
was on BitBucket life would be a lot easier. However, every time this
topic comes up we end up at "someone insists that the project cannot
survive without the entire history, and no-one can be bothered to dothe 
transfer" so nothing happens. 

Of course Tigris looks like it is going away by the look of the sign
saying "no more projects". If that happens and nothing is done, history
gets lost anyway.

I'd suggest opening up the issue tracker on the SCons repository on
BitBucket sooner rather than later.

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev


Re: [Scons-dev] D tool development workflow

2017-04-18 Thread Bill Deegan
I think reporting bugs for one part of SCons in a different repository is
not a good idea.
If the intent is to get it into the scons core.
If the intent is to provide as a standalone, then bugtrack on whichever
system you'd like.

-Bill

On Tue, Apr 18, 2017 at 1:07 AM, Russel Winder  wrote:

> I believe that development of the D tools is going to be improved by
> assuming that other D tool users want to work with a standard SCons and
> updates just to the D tools; that they are not interested in working
> with SCons from a Mercurial clone that has to be managed. I am
> therefore going to make https://github.com/russel/SCons_D_Experiment
> the focus of all D tool evolution and then providing pull requests to
> the SCons BitBucket mainline. A little more work for me in that I
> develop in one repository and then update another, but the D world is a
> Git world and so using a Git repository is more natural for handling
> experiments.
>
> I am also going to break ranks and have D tool issues come in to this
> repository. D folk are GitHub folk, I am not going to push them to
> Tigris for bug reporting – especially as I never go there myself any
> more.
>
> This is about pragmatics, I believe more will be achieved this way.
>
> Hopefully, what has started is a race for SCons and Meson to beat Dub
> in building D code.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev