Doing a release

[Aaron Meurer]

It has come to my attention that the IPython notebook LaTeX rendering is broken in the latest versions of SymPy and IPython (now known as Jupyter), but works in SymPy master.

Because of this, we should get a release out. The plans were to make the next release 1.0. Are there still major blockers to getting 1.0 out soon? If so, we should do a patch release with the fix to the notebook.  

Regardless, I would like to get 1.0 out relatively soon (definitely by the end of the year).

Aaron Meurer

[AMiT Kumar]

Hi,

Can we delay 1.0 with some other version like 0.8 or so?

It would be nice if have atleast the solve mess being fixed in 1.0

Also we have yet not settled with a single assumption system.

AMiT Kumar

[Francesco Bonazzi]

On Tuesday, 1 September 2015 06:30:20 UTC+2, AMiT Kumar wrote:

Also we have yet not settled with a single assumption system.

Maybe the two assumption systems should be merged before the release of version 1.0. It would be a bit weird to release version 1.0, and later deprecate the old-style assumptions system.

[Aaron Meurer]

I think there are already some breaking changes in master, no? Maybe we should just do a bugfix 0.7.7 release to fix the Jupyter stuff. 

Aaron Meurer

--
You received this message because you are subscribed to the Google Groups "sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
To post to this group, send email to sy...@googlegroups.com.
Visit this group at http://groups.google.com/group/sympy.
To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/471ddadb-0eb4-4139-98ec-a7b193b70a15%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Aaron Meurer]

OK, I have created a release branch for 0.7.6.1, which is just 0.7.6 plus the IPython fixes https://github.com/sympy/sympy/pull/9883. Are there any other fixes needed for IPython? I tested the notebook in the release branch and the MathJax printing works. 

Otherwise, I will probably push this out later tonight or tomorrow. 

Aaron Meurer

[Aaron Meurer]

The current plan is not to deprecate the old assumptions syntax, but to allow both. In fact, in master, ask(Q.real(x)) will call x.is_real (it also calls some other stuff too, so it should be at least as smart), so you can start migrating to the new assumptions syntax. The reverse doesn't work yet (even if you put Q.real(x) in the global assumptions).  

It would be nice to finish completing the breaking changes that are planned. I believe the only one remaining to be fixed after Sudhanshu's GSoC project relates to oo.is_real. Is there anything else Sudhanshu? Perhaps we should try to work out some of the performance issues.

How long will it be until the solvers are in a 1.0 ready state, especially now that GSoC is over?

Aaron Meurer

On Tue, Sep 1, 2015 at 2:45 AM, Francesco Bonazzi <franz....@gmail.com> wrote:

--

You received this message because you are subscribed to the Google Groups "sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
To post to this group, send email to sy...@googlegroups.com.
Visit this group at http://groups.google.com/group/sympy.

[AMiT Kumar]

> How long will it be until the solvers are in a 1.0 ready state, especially now that GSoC is over?

This can take some time. Since the solveset is not complete with respect to old `solve`, 

one of the things that can be done for faster release of 1.0 is rename `solveset` as

`solve` and the previous `solve` as `old_solve` (old_solve not in public API) so that the people

using old solve in there code can quickly fix breaking of their code, and we can continue

the development of `solve` (formerly `solveset`). This transition may require a couple of

versions. Though, we also need to decide the behavior deprecation policy for `solve`.

AMiT Kumar