Re: [sage-combinat-devel] Re: Ticket #12215

[Anne Schilling]

Hi Simon,

This might not be the best time to switch to the sage-combinat queue,
at least not if you want to rebase against 5457. I am taking this
development off the sage-combinat queue due to conflict with 13109.

In any case, I took the line "- Use #7980 to handle the multiple realizations"
out of sf.py, so you might need to rebase your patch in case it goes
in after 5457.

Viele Gruesse,

Anne


On 7/16/12 2:30 PM, Simon King wrote:
> Hi Anne,
> 
> On 2012-07-16, Anne Schilling  wrote:
>> Dear Simon,
>>
>> I would like to remove the line
>>
>> "- Use #7980 to handle the multiple realizations"
>>
>> in ../combinat/sf/sf.py in our upcoming clean-up of symmetric functions.
>>
>> However, this creates a conflict with your patch
>> trac12215_segfault_fixes.patch
>> in the sage-combinat queue. Would it be possible for you to take out this
>> edit in sf.py from your patch?
> 
> I would not mind to update the patch on trac. But I guess I should finally
> have a look at the Wiki page on the combinat queue and learn how to post
> patches there.
> 
> Best regards,
> Simon

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] please rebase the following patches on sage-5.2.rc0

[Anne Schilling]

Dear Volker Braun,

I just installed sage-5.2.rc0, but your deprecation patch does not seem to be 
part of
this version.

Dear All!

I guarded the following patches with respect to sage-5.2.rc0 as
they did not apply any longer:

trac_11305-rigged_configurations-ts.patch
trac_12882-generalized_cartan_matrix_as_dynkin_diagram_input-cs.patch
trac_11187-finite_reflection_groups-cs.patch
coxeter3-mh_rebase-cs.patch
coxeter3-iterator.patch
12380_move_nfactors_methods-sl.patch
trac_13088_paley_graphs-fc.patch
trac_13072-tuples-of-partitions_am.patch
trac_13074-tuples-of-tableaux-am.patch
extended_affine_weyl_groups_sd40.patch
extended_affine_weyl_groups_sd40-work.patch

Please rebase them with respect to the new version of sage!

Cheers,

Anne

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Keshav Kini]

Dima Pasechnik  writes:
> On Monday, 16 July 2012 20:49:31 UTC+8, Keshav Kini wrote:
> Not really. That's only if you do `hg import --exact`, and your
> patch
> has the correct headers.
> yes, but it should have them if its creator had 
> [diff]
> git = 1
> in her/his .hgrc.

No, this is actually irrelevant. Putting those lines in ~/.hgrc causes
the git diff format to be used, which is able to store the following
information that standard unified diffs cannot (text taken from `hg help
diffs`):

- executable status and other permission bits
- copy or rename information
- changes in binary files
- creation or deletion of empty files

Node ID headers are always created by `hg export`, but they are not
created by `hg qrefresh`, because they are dependent upon properties of
the actual commit object in the history tree, not simply the diff hunks.

The most likely reason for a patch file to not have those headers is
that it was copied directly out of .hg/patches instead of being created
with `hg export qtip > foo.patch`.

-Keshav


Join us in #sagemath on irc.freenode.net !

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: queue broken

[Travis Scrimshaw]

Nicolas,
   I'm apply the same guard to #6538. Let me know if there's a better 
approach to this.

With #12925 applied, #6538 applies for me. Thus the queue should work now 
(at least with respect to my patch).

Thanks,
Travis

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/wY7vZEYNLD4J.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: Ticket #12215

[Simon King]

Hi Anne,

On 2012-07-16, Anne Schilling  wrote:
> Dear Simon,
>
> I would like to remove the line
>
> "- Use #7980 to handle the multiple realizations"
>
> in ../combinat/sf/sf.py in our upcoming clean-up of symmetric functions.
>
> However, this creates a conflict with your patch
> trac12215_segfault_fixes.patch
> in the sage-combinat queue. Would it be possible for you to take out this
> edit in sf.py from your patch?

I would not mind to update the patch on trac. But I guess I should finally
have a look at the Wiki page on the combinat queue and learn how to post
patches there.

Best regards,
Simon

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] The deprecation change

[Anne Schilling]

Thanks, but this was really unnecessary since a more updated version of the 
patch is
on the sage-combinat queue (where it is under development) and that one needs 
to be
updated (not the one on trac).

Best regards,

Anne

On 7/16/12 10:33 AM, Volker Braun wrote:
> I've rebased http://trac.sagemath.org/5457 on top of the deprecation change 
> and it was a very easy change.
> 
> 
> 
> On Monday, July 16, 2012 12:53:52 PM UTC-4, Anne Schilling wrote:
> 
> Dear Volker,
> 
> The ticket http://trac.sagemath.org/5457 uses many deprecation warnings.
> I tried to make your change last week, but then it breaks the entire 
> sage-combinat
> queue.
> 
> Dear Sage-combinat developers,
> 
> Would it be ok to move to sage-5.2.rc0 and have the queue apply there to 
> avoid the
> conflicts?
> 
> Best,
> 
> Anne
> 
> On 7/16/12 9:49 AM, Volker Braun wrote:
> > The change in the deprecation framework (http://trac.sagemath.org/10508 
> ) has been merged into sage-5.2.rc0. This is 
> likely to break any patches that are using
> deprecations. I can assure you that all
> > jokes about deprecating the deprecations have already been made.
> >
> > What needs to be done is switch this:
> >
> >   from sage.misc.misc import deprecation
> >   deprecation('description', 'Sage version X')
> >
> > to
> >
> >   from sage.misc.superseded import deprecation
> >   deprecation(ticket_number, 'description')
> >
> > where ticket_number is an integer, the trac ticket number. If you have 
> any questions, I'd be happy to help.
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "sage-combinat-devel" group.
> To view this discussion on the web visit 
> https://groups.google.com/d/msg/sage-combinat-devel/-/w0b1kcpIw4oJ.
> To post to this group, send email to sage-combinat-devel@googlegroups.com.
> To unsubscribe from this group, send email to 
> sage-combinat-devel+unsubscr...@googlegroups.com.
> For more options, visit this group at 
> http://groups.google.com/group/sage-combinat-devel?hl=en.

-- 
Anne Schilling  E-mail: a...@math.ucdavis.edu
Professor   Web:http://www.math.ucdavis.edu/~anne
Department of Mathematics   Office: MSB 3222
University of CaliforniaPhone:  (530) 554 2326
One Shields Ave Fax:(530) 752 6635
Davis, CA 95616

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: queue broken

[Nicolas M. Thiery]

Hi Travis,

On Sun, Jul 15, 2012 at 07:20:39AM -0700, Travis Scrimshaw wrote:
>   I have a hunch as to what is going on with my patch. There's a
>dependency on Nicolas's 12925 tutorial patch which has been merged into
>5.2, thus it no longer appears in the patch queue. In particular it
>created the combinat/tutorial.py file which my little change depends on.
>My hunch is that this should work fine in 5.2.
>Can someone with 5.2 test this please?

It's in the queue: trac_12925-combinatorics_thematic_tutorial-nt.patch 
#+5_2_rc0

Note: I just updated the guard; this may change something for you if
you were running a beta version of 5.2. Otherwise it should not change
anything (e.g. it should be applied on Sage <= 5.1).

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] The deprecation change

[Volker Braun]

I've rebased http://trac.sagemath.org/5457 on top of the deprecation change 
and it was a very easy change.



On Monday, July 16, 2012 12:53:52 PM UTC-4, Anne Schilling wrote:
>
> Dear Volker, 
>
> The ticket http://trac.sagemath.org/5457 uses many deprecation warnings. 
> I tried to make your change last week, but then it breaks the entire 
> sage-combinat 
> queue. 
>
> Dear Sage-combinat developers, 
>
> Would it be ok to move to sage-5.2.rc0 and have the queue apply there to 
> avoid the 
> conflicts? 
>
> Best, 
>
> Anne 
>
> On 7/16/12 9:49 AM, Volker Braun wrote: 
> > The change in the deprecation framework (http://trac.sagemath.org/10508) 
> has been merged into sage-5.2.rc0. This is likely to break any patches that 
> are using deprecations. I can assure you that all 
> > jokes about deprecating the deprecations have already been made. 
> > 
> > What needs to be done is switch this: 
> > 
> >   from sage.misc.misc import deprecation 
> >   deprecation('description', 'Sage version X') 
> > 
> > to 
> > 
> >   from sage.misc.superseded import deprecation 
> >   deprecation(ticket_number, 'description') 
> > 
> > where ticket_number is an integer, the trac ticket number. If you have 
> any questions, I'd be happy to help. 
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/w0b1kcpIw4oJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Sage Days 40: Anne, Nicolas, and Daniel

[Nicolas M. Thiery]

On Sun, Jul 15, 2012 at 07:03:13PM -0400, Mark Shimozono wrote:
> P.P.S Nicolas T.,  thanks much for the recent functorial construction
> of the group algebra of a root-lattice-realization, 
> which was duly shown to me by Nicolas B.  From this it will be easy for
> me to gracefully do GKM theory for torus-equivariant K-theory of Kac-Moody 
> flags
> (well, modulo that annoying division, which I can manage).
> In order to do cohomology I need something similar for the symmetric algebra
> of the root-lattice-realization (polynomial ring with degree one part given 
> by 
> the root-lattice-realization).  Did you implement that?

Nope. One would just need to add a new functorial construction
"SymmetricAlgebra", similar to that in
sage.categories.algebra_functor.

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] The deprecation change

[Nicolas M. Thiery]

Hi Anne!

On Mon, Jul 16, 2012 at 09:53:52AM -0700, Anne Schilling wrote:
> Would it be ok to move to sage-5.2.rc0 and have the queue apply
> there to avoid the conflicts?

That's fine for me, but this would mean that all the Sage Days 40
newcomers from last week will need to instantly update their 5.0 to
5.2 to pursue development; not the nicest welcome message ...

Hopefully we can "just" insert the deprecation patches in the queue,
and then rebase all the patches there on top of it. I guess it's best
if a single person handles it.

Any volunteer? (worst case I guess I can handle it, but officially I
am on vacation).

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] The deprecation change

[Nicolas M. Thiery]

On Mon, Jul 16, 2012 at 09:49:47AM -0700, Volker Braun wrote:
>The change in the deprecation framework (http://trac.sagemath.org/10508)
>has been merged into sage-5.2.rc0. This is likely to break any patches
>that are using deprecations.

Sight ...

Oh well, let us all count and report how many hours this is going to
waste us, so that we can at least learn from it and see if and how
much we should weight in next time.

Best,
Nicolas
--
Nicolas M. Thiéry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] The deprecation change

[Anne Schilling]

Dear Volker,

The ticket http://trac.sagemath.org/5457 uses many deprecation warnings.
I tried to make your change last week, but then it breaks the entire 
sage-combinat
queue.

Dear Sage-combinat developers,

Would it be ok to move to sage-5.2.rc0 and have the queue apply there to avoid 
the
conflicts?

Best,

Anne

On 7/16/12 9:49 AM, Volker Braun wrote:
> The change in the deprecation framework (http://trac.sagemath.org/10508) has 
> been merged into sage-5.2.rc0. This is likely to break any patches that are 
> using deprecations. I can assure you that all
> jokes about deprecating the deprecations have already been made.
> 
> What needs to be done is switch this:
> 
>   from sage.misc.misc import deprecation
>   deprecation('description', 'Sage version X')
> 
> to 
> 
>   from sage.misc.superseded import deprecation
>   deprecation(ticket_number, 'description')
> 
> where ticket_number is an integer, the trac ticket number. If you have any 
> questions, I'd be happy to help.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] The deprecation change

[Volker Braun]

The change in the deprecation framework (http://trac.sagemath.org/10508) 
has been merged into sage-5.2.rc0. This is likely to break any patches that 
are using deprecations. I can assure you that all jokes about deprecating 
the deprecations have already been made.

What needs to be done is switch this:

  from sage.misc.misc import deprecation
  deprecation('description', 'Sage version X')

to 

  from sage.misc.superseded import deprecation
  deprecation(ticket_number, 'description')

where ticket_number is an integer, the trac ticket number. If you have any 
questions, I'd be happy to help.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/0J5WIZPKz3MJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: Sage Days 40

[bump]

Mark Shimozono wrote:

| P.P.P.S Daniel Bump, let me know when I can have a go at the extended 
affine Weyl groups, 
| Hecke algebras and their polynomial modules, etc. 

I'm keeping my revisions to the extended affine patch in a separate patch 
to be folded. I
have some debugging and further coding to do before I'm ready to push it to 
the queue. 
So I'd like to keep possession of the extended patch for the time being. My 
plan is to 
make it work with IwahoriHeckeAlgebrasT. 

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/0tQh6J_GiuQJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: queue broken

[Nicolas M. Thiery]

On Sat, Jul 14, 2012 at 10:25:16PM -0700, Anne Schilling wrote:
> I do not think there is support for symmetric functions in several alphabets 
> yet,
> but I could be wrong.

To emulate symmetric functions in two alphabets, you can take tensor products:

sage: Sym = SymmetricFunctions(QQ)
sage: Sym.inject_shorthands()
sage: ss = tensor([s, s])
sage: ss.category()
Category of tensor products of hopf algebras with basis over Rational Field

Basic arithmetic is implemented:

sage: x = ss.an_element()
s[] # s[] + 2*s[] # s[1] + 3*s[] # s[2]
sage: x * x
s[] # s[] + 4*s[] # s[1] + 4*s[] # s[1, 1] + 10*s[] # s[2] + 12*s[] # s[2, 
1] + 9*s[] # s[2, 2] + 12*s[] # s[3] + 9*s[] # s[3, 1] + 9*s[] # s[4]

Now many desirable features are missing:

sage: x.coproduct()
... NotImplemented ...
sage: pp = tensor([p,p])
sage: pp(x)
... NotImplementedError ...

Those should be implemented generically for tensor products. The first
one should be just a couple lines in:

Coalgebras.TensorProducts.ParentMethods.coproduct

The other is probably best implemented by adding an appropriate
_coerce_map_from in tensor products (I would need to search for the
best spot for this.

Plethysm (in a stupid way) should not be much more work. To be put in:

 sage.combinat.sf.sfa.SymmetricFunctionsBases.TensorProducts.ElementMethods

or something similar.

Please open a ticket and list whatever features you might want!

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Ticket #12215

[Anne Schilling]

Dear Simon,

I would like to remove the line

"- Use #7980 to handle the multiple realizations"

in ../combinat/sf/sf.py in our upcoming clean-up of symmetric functions.

However, this creates a conflict with your patch
trac12215_segfault_fixes.patch
in the sage-combinat queue. Would it be possible for you to take out this
edit in sf.py from your patch?

Thanks,

Anne

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Sage Days 40

[Nicolas M. Thiery]

On Sat, Jul 14, 2012 at 12:29:21AM -0700, Anne Schilling wrote:
> Thanks again to everyone who participated in the Sage Days 40 this past
> week in Minneapolis. It was great to meet you all and work with you.
> I hope everyone got something out of the workshop.
>
> Many thanks to Gregg Musiker and IMA for all the local organization,
> and the NSF for funding!
>
> Let's keep in touch on this mailing list about everyone's progress.

+1! And thanks Anne for your talks and leadership!

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" 
http://Nicolas.Thiery.name/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: Patches for PartitionTuples and TableauTuples

[Anne Schilling]

Dear Andrew,

> Finally, I would like to move these patches up to the "needs review" section 
> of the series file. Do I just do this - and check that everything still 
> applies - or should I defer to some one who knows
> what they are doing:) ? I am happy to do either.

You can just move your patch up in the series file as long as you check that
the whole queue still applies when you do so.

Could you for the moment please avoid putting it before or close to

trac_5457-symmetric_functions-mz.patch
anne-sf-as.patch

since Mike and I are trying to get this into sage in the next couple of days
and it would be good to avoid a conflict in the series file (Mike is going to
add another patch right after mine).

As long as you add your patch one or two blocks down, it should be fine.

Best,

Anne

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Sage Days 40: Anne, Nicolas, and Daniel

[Anne Schilling]

Hi Mark,

> Where can I get the crystal demo worksheet?
> (Thanks Nicolas for the cool interacts).

You can find it here

http://www.ima.umn.edu/2011-2012/SW7.9-13.12/abstracts.html

or on the wiki for Sage Days 40

http://wiki.sagemath.org/days40_tutorials

> P.S. Can you/someone please review my massive coxeter patch,
> which has been languishing for months?

Christian Stump offered to review the patch. If you still need someone
after that, I can do it.

> P.P.P.S Daniel Bump, let me know when I can have a go at the extended affine 
> Weyl groups,
> Hecke algebras and their polynomial modules, etc.

Please also keep me updated and let me know when/if I can help.

Anne

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Dima Pasechnik]

On Monday, 16 July 2012 20:49:31 UTC+8, Keshav Kini wrote:
>
> Dima Pasechnik  writes: 
> > This pain is caused by mq (sorry, mq, you are evil. live with it 
> > :-)). 
> > Just pretending that the patch is a regular patch, importing it with 
> > hg import 
> > would land you in a workflow very much like the git's workflow for 
> > the 
> > problem Ralf described. 
>
> Not really. That's only if you do `hg import --exact`, and your patch 
> has the correct headers. 
>

yes, but it should have them if its creator had 
[diff]
git = 1
in her/his .hgrc. 
Anyways, if you have to rebase your patch created from a tree you have,
against something based in a not so distant node on your tree, you are all 
set.

 

>
> Working with patches is always going to be painful, whether you use mq 
> or not. 
>
> > (git doesn't really have an advantage here, rather it's us shooting 
> > ourselves 
> > in the foot by sticking to the mq workflow (workf#$%, if you ask me 
> > :-))) 
>
> It has some advantages, such as the fact that its rebase command is 
> built in, and so is its configuration for using various 3-way diff 
> viewers like kdiff3. 
>
> -Keshav 
>
>  
> Join us in #sagemath on irc.freenode.net ! 
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/rGY1ttl4jtMJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Keshav Kini]

Dima Pasechnik  writes:
> This pain is caused by mq (sorry, mq, you are evil. live with it
> :-)). 
> Just pretending that the patch is a regular patch, importing it with
> hg import
> would land you in a workflow very much like the git's workflow for
> the
> problem Ralf described.

Not really. That's only if you do `hg import --exact`, and your patch
has the correct headers.

Working with patches is always going to be painful, whether you use mq
or not.

> (git doesn't really have an advantage here, rather it's us shooting
> ourselves
> in the foot by sticking to the mq workflow (workf#$%, if you ask me
> :-)))

It has some advantages, such as the fact that its rebase command is
built in, and so is its configuration for using various 3-way diff
viewers like kdiff3.

-Keshav


Join us in #sagemath on irc.freenode.net !

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Dima Pasechnik]

This pain is caused by mq (sorry, mq, you are evil. live with it :-)). 
Just pretending that the patch is a regular patch, importing it with hg 
import
would land you in a workflow very much like the git's workflow for the
problem Ralf described.
(git doesn't really have an advantage here, rather it's us shooting 
ourselves
in the foot by sticking to the mq workflow (workf#$%, if you ask me :-)))


On Monday, 16 July 2012 20:33:02 UTC+8, Keshav Kini wrote:
>
> Ralf Hemmecke  writes: 
> > Or simpler still, since no additional setup would be necessary (except 
> > for "git config"). 
> > 
> > git config --global merge.too kdiff3 
>
> s/too/tool/ 
>
> > git checkout somebranch 
> > git rebase otherbranch 
> > # in case of conflict 
> > git mergetool 
> > git rebase --continue 
> > 
> > But unfortunately you are not using git. 
>
> Yet. :) 
>
> -Keshav 
>
>  
> Join us in #sagemath on irc.freenode.net ! 
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/Wymomhh0_JwJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Keshav Kini]

Ralf Hemmecke  writes:
> Or simpler still, since no additional setup would be necessary (except
> for "git config").
>
> git config --global merge.too kdiff3

s/too/tool/

> git checkout somebranch
> git rebase otherbranch
> # in case of conflict
> git mergetool
> git rebase --continue
>
> But unfortunately you are not using git.

Yet. :)

-Keshav


Join us in #sagemath on irc.freenode.net !

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: how to rebase a patch ?

[Nicolas Borie]

Le 16/07/2012 12:45, Simon King a écrit :
While we are at it: One thing that prevents me from working with the 
combinat queue is that I never understood what I need to set up in 
order to push my local changes to the "official" combinat queue. Best 
regards, Simon 

Hi,

Pushing - Pulling on the official combinat repository can be set in your 
~/.hgrc file : look the documentation at :

http://wiki.sagemath.org/combinat/MercurialStepByStep#Mercurial_configuration

I currently have something like ten different mercurial local 
repositories on my computer and just being located in 
/opt/sage/devel/sage-combinat/.hg/patches/ is enought for my hg command 
to select the right place to push/pull.


Best regards,
Nicolas B.

--
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Simon King]

Hi Cristian,

On 2012-07-16, Christian Stump  wrote:
>> what I usually do is:
>
> this is a very nice and detailed description -- thanks! maybe you
> wanna add it to wiki?

I thought there already is a Wiki page or a page in the developers guide?
I didn't try to find it though.

>>  * hg qrefresh followed by hg export qtip > foo.patch
>
> is there, in this situation, a difference between hg export qtip >
> foo.patch and hg qrefresh (that's what I usually do here) ?

hg qrefresh means that the patch in your queue is updated (you can add a
commit message when you do "hg qrefresh -e", by the way - it will open
a text editor). If it concerns a patch foo.patch, then this updating
happens in .hg/patches/foo.patch.

In contrast, hg export qtip means that you write the patch into a "regular"
file (i.e., one that is not in a hidden folder), and I think the file
format is slightly different (namely, the patch in .hg/patches also
contains information on your local queue, that is not relevant for other
people's queues. Hence, IIRC, the stuff in .hg/patches/ is *not* what
you post on trac. 

If you attempt hg export qtip when there are local changes (i.e.,
*before* you do hg qrefresh), I think mercurial will refuse.

While we are at it: One thing that prevents me from working with the
combinat queue is that I never understood what I need to set up in order
to push my local changes to the "official" combinat queue.

Best regards,
Simon

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: how to rebase a patch ?

[Ralf Hemmecke]

* Turn on the rebase extension in your ~/.hgrc file
* Set up a merge tool by means of some voodoo which I've now
  forgotten (this may even be handled automatically in newer versions
  of Mercurial)
* Do `hg update -C foo` where foo is the version of Sage on which the
  patch is known to apply
* Apply the patch
* Do `hg heads` and note the ID of the commit which is the latest
  version of Sage (as opposed to the commit which is the patch you
  just applied)
* Do `hg rebase -d ID` where ID is the ID previously noted
* If a merge window pops up, resolve merge conflicts with the handy
  dandy merge tool you configured
* Voila, the patch is now rebased

This is more complicated to set up but also allows you to avoid retyping
code and fixing up files manually.


Or simpler still, since no additional setup would be necessary (except 
for "git config").


git config --global merge.too kdiff3
git checkout somebranch
git rebase otherbranch
# in case of conflict
git mergetool
git rebase --continue

But unfortunately you are not using git.

"git gui" is also quite a powerful tool.

Sorry for the noise.

Ralf

--
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Keshav Kini]

Christian Stump
 writes:
> is there, in this situation, a difference between hg export qtip >
> foo.patch and hg qrefresh (that's what I usually do here) ?

`hg qrefresh` adds uncommitted changes into the topmost patch; `hg
export qtip > foo.patch` simply copies the existing patch file into
whatever path you specify ("foo.patch" in this case), while also adding
a couple of headers specifying the parent commit ID and the patch's own
commit ID, which can change even when the content of the patch is the
same.

-Keshav


Join us in #sagemath on irc.freenode.net !

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Keshav Kini]

Simon King  writes:
> Hi Frederic,
>
> what I usually do is: 
>
>  * When I have a patch foo.patch that needs rebasing, then I take a
>Sage installations with all required prerequisites (i.e.,
>the dependencies of foo.patch).
>  * Then, I try to apply foo.patch (i.e., hg qimport foo.patch followed
>by hg qpush).
>  * If applying the patch does work, then I complain on the ticket that
>the patch does not need rebasing, from my perspective, and that the
>release manager or the reviewer please state which other tickets need
>to be applied to see the merge conflict.
>  * If applying the patch does *not* work (as expected), then mercurial
>will automatically create some files that show the conflicts. Say, if
>foo.patch shall modify groups.py, but one hunk of the patch does not
>fit, then groups.py.rej is created.
>  * Now, I'd read groups.py.rej, to see what part of the patch did not
>apply, and why. Note that every part of foo.patch that is not
>mentioned in one of the rej files is fine, so, no need to worry
>about them.
>  * I'd edit groups.py, so that the parts of groups.py mentioned in
>groups.py.rej are as they are meant to be with foo.patch.
>  * hg qrefresh followed by hg export qtip > foo.patch
>  * Post the new version of foo.patch to trac, and don't forget to change
>the state of the ticket back to "needs review" or post a comment (if
>you just update the attachment, trac will not notify the reviewer)

Here's another way:

   * Turn on the rebase extension in your ~/.hgrc file
   * Set up a merge tool by means of some voodoo which I've now
 forgotten (this may even be handled automatically in newer versions
 of Mercurial)
   * Do `hg update -C foo` where foo is the version of Sage on which the
 patch is known to apply
   * Apply the patch
   * Do `hg heads` and note the ID of the commit which is the latest
 version of Sage (as opposed to the commit which is the patch you
 just applied)
   * Do `hg rebase -d ID` where ID is the ID previously noted
   * If a merge window pops up, resolve merge conflicts with the handy
 dandy merge tool you configured
   * Voila, the patch is now rebased

This is more complicated to set up but also allows you to avoid retyping
code and fixing up files manually. Often, a conflict that would have
caused .rej files to appear if you just tried to apply the patch will be
automatically handled by the logic of the `hg rebase` command, and even
if it isn't handled, you get to manipulate the hunks directly, hopefully
making sure you don't lose anything.

-Keshav


Join us in #sagemath on irc.freenode.net !

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

Re: [sage-combinat-devel] Re: how to rebase a patch ?

[Christian Stump]

> what I usually do is:

this is a very nice and detailed description -- thanks! maybe you
wanna add it to wiki?

>  * hg qrefresh followed by hg export qtip > foo.patch

is there, in this situation, a difference between hg export qtip >
foo.patch and hg qrefresh (that's what I usually do here) ?

Thanks, Christian

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] Re: how to rebase a patch ?

[Simon King]

Hi Frederic,

what I usually do is: 

 * When I have a patch foo.patch that needs rebasing, then I take a
   Sage installations with all required prerequisites (i.e.,
   the dependencies of foo.patch).
 * Then, I try to apply foo.patch (i.e., hg qimport foo.patch followed
   by hg qpush).
 * If applying the patch does work, then I complain on the ticket that
   the patch does not need rebasing, from my perspective, and that the
   release manager or the reviewer please state which other tickets need
   to be applied to see the merge conflict.
 * If applying the patch does *not* work (as expected), then mercurial
   will automatically create some files that show the conflicts. Say, if
   foo.patch shall modify groups.py, but one hunk of the patch does not
   fit, then groups.py.rej is created.
 * Now, I'd read groups.py.rej, to see what part of the patch did not
   apply, and why. Note that every part of foo.patch that is not
   mentioned in one of the rej files is fine, so, no need to worry
   about them.
 * I'd edit groups.py, so that the parts of groups.py mentioned in
   groups.py.rej are as they are meant to be with foo.patch.
 * hg qrefresh followed by hg export qtip > foo.patch
 * Post the new version of foo.patch to trac, and don't forget to change
   the state of the ticket back to "needs review" or post a comment (if
   you just update the attachment, trac will not notify the reviewer)

Best regards,
Simon

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.

[sage-combinat-devel] how to rebase a patch ?

[Frédéric Chapoton]

Hello,

I was told to rebase a patch (namely 
http://trac.sagemath.org/sage_trac/ticket/13067) by the release manager.

What is the procedure to do that ? I have not found it on the mercurial 
guide (http://wiki.sagemath.org/combinat/MercurialStepByStep).

I have only a rather basic understanding of the intricacies of mercurial, 
and usually just follow the instructions blindly..

Thanks for your help.

Frederic

-- 
You received this message because you are subscribed to the Google Groups 
"sage-combinat-devel" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/sage-combinat-devel/-/8iXr7Dk3c6wJ.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to 
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sage-combinat-devel?hl=en.