WIP Support for Celltotex

[mmatera]

In this branch, I am trying to include all the fixes needed to make that CellsToTeX package

Import@"https://raw.githubusercontent.com/jkuczm/MathematicaCellsToTeX/master/NoInstall.m"
This involves improvements in WL native Patterns support, Protect/Unprotect mechanisms, Names[], etc. The idea would be to merge this branch when the Import stops showing errors, including adding some tests.

[rocky]

I love this idea! I can find a number of other packages including some of the ones mentioned at sagemath that basically use the same kind of install mechanism.

I bet https://github.com/jkuczm/MathematicaSyntaxAnnotations#no-installation would be another one that would work because it outside of the install mechanism what's going on is pretty straightforward.

Maybe after this is done we can come up with a list of packages that can be installed this way and mention that in the repository somewhere.

[mmatera]

Status update: several bugs and implementation issues on the Package and context subsystems were fixed.
The first part of CellsToTeX package loads without reporting problems:
BeginPackage@"CellsToTeX"
Import@"https://raw.githubusercontent.com/jkuczm/MathematicaSyntaxAnnotations/v0.2.2/SyntaxAnnotations/SyntaxAnnotations.m";
EndPackage[]
The second part Import@"https://raw.githubusercontent.com/jkuczm/MathematicaCellsToTeX/master/CellsToTeX/CellsToTeX.m"`
loads properly until the line 820, where starts raising messages connected with pattern matching functions.

[mmatera]

A test that actually goes out to the internet, can be fragile: an internet connection might not be available and the URL can change, to name a couple of fragilities.

Marking the test with @pytest.skipif is one way to allow running the test conditionally. The flaky module is another way to indicate that a test might fail.

I added a line looking if the Import was successful. If it is, it performs the test. An improvement would be to raise a warning when this happens.

The problem with flaky and pytest.skipif is that I would need to know (ahead) if there is an internet connection, and the package is available.

@rocky , if you can give me a hand with the part of testing, I can continue with the fixings for the interpreter.

[rocky]

@rocky , if you can give me a hand with the part of testing, I can continue with the fixings for the interpreter.

Sure. Over the weekend I can add some commits in a branch off of this that you can look at.

[mmatera]

Update: I found and fixed some issues with Systemp`Protect. Also, I included some support for DeleteCases with the levelspect parameter, and n. Still does not work, but now the number of errors is smaller...

[rocky]

Thanks for undertaking this.

Being able to do some sort of import of simple packages is very important. I am guessing that after this is done there will be many many more packages that we will be able to load which will add both power to Mathics and help draw in a community using and working on Mathics.

[rocky]

Just curious, How close is this to working? I also wonder what our chances are for importing other things of this kind are now?

[mmatera]

Well, I am not sure when this package is going to work, but in the way I think we have improved a lot the support for the existing symbols. At this point, it seems that there are not fails in the import process, but for some reason, it still fails when the main routines are called. I hope to have more time during the next weeks to continue with this.

[rocky]

It might be good to break this PR up into little pieces that work and get that merged into master.

[mmatera]

It might be good to break this PR up into little pieces that work and get that merged into master.

Actually, I did it and a couple of them were already merged. The remaining, working changes are in #1088

[rocky]

@mmatera It would be good to merge this into master.

[rocky]

@mmatera It would also be good to get this into master

[rocky]

I think I've run into this problem a bit too, so it might be good to get this into master.

[mmatera]

I open a new branch (PR #1102) with all these improvements. The only thing is that for some reason I can not do a rebase in order to keep just the changes. Maybe @rocky can help me to tide it up.

[rocky]

I open a new branch (PR #1102) with all these improvements. The only thing is that for some reason I can not do a rebase in order to keep just the changes. Maybe @rocky can help me to tide it up.

Sure. Will do. Oh - and by the way - thanks for all of the great work here!