Update ExaModels and MadNLP

[pulsipher]

This tests out the ExaModels changes proposed in exanauts/ExaModels.jl#256.

This now updates the versioning of ExaModels and MadNLP with the new versions.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 61.97183% with 27 lines in your changes missing coverage. Please review.
✅ Project coverage is 49.92%. Comparing base (64c85bb) to head (239e584).

Files with missing lines	Patch %	Lines
src/transform.jl	61.97%	27 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (64c85bb) and HEAD (239e584). Click for more details.

HEAD has 5 uploads less than BASE

Flag	BASE (64c85bb)	HEAD (239e584)
	6	1

Additional details and impacted files

@@             Coverage Diff             @@
##             main      #31       +/-   ##
===========================================
- Coverage   85.09%   49.92%   -35.17%     
===========================================
  Files           6        5        -1     
  Lines         711      701       -10     
===========================================
- Hits          605      350      -255     
- Misses        106      351      +245     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[pulsipher]

This appears to make very little difference for performance and allocations/memory. Using the quadrotor example from our benchmarks with 1000 time points, I get the following with the current release:

using NLPModelsIpopt, InfiniteOpt, InfiniteExaModels

m = quad(num_supports = 1000, backend = ExaTranscriptionBackend(IpoptSolver))
set_silent(m)
@time build_transcription_backend!(m) # builds the ExaModel
@time optimize!(m)

0.168895 seconds (450.06 k allocations: 24.030 MiB)
0.865922 seconds (310 allocations: 2.626 MiB)

With this PR, I get:

0.174419 seconds (449.24 k allocations: 24.468 MiB)
0.856932 seconds (310 allocations: 2.629 MiB)

[pulsipher]

This is broken because of exanauts/ExaModels.jl#269. Previously, we could add constraints without a generator by directly giving an ExaModels expression and an iterator, but that method was removed.

[pulsipher]

Requires exanauts/ExaModels.jl#286.