[doc] update install instructions use triqs/unstable adapol/triqs_xca

Author: HugoStrand

doc/install.rst

@@ -8,19 +8,17 @@ Install xca
 Compiling xca from source
 ===============================
 
-.. note:: To guarantee reproducibility in scientific calculations we strongly recommend the use of a stable `release <https://github.com/TRIQS/triqs/releases>`_ of both TRIQS and its applications.
-
 Prerequisites
 -------------
 
 #. The :ref:`TRIQS <triqslibs:welcome>` library, see :ref:`TRIQS installation instruction <triqslibs:triqs_install>`.
    In the following, we assume that TRIQS is installed in the directory ``path_to_triqs``.
 
-   We are currently depending on the `unstable` development version of TRIQS (that will be released as version 4.0 in the end of 2025). However, for a smoother installation experience we provide tested branches with the naming pattern `triqs/DEV_PPSC_*` (as of writing `triqs/DEV_PPSC_20250912`) that trail the `triqs/unstable` branch with fixed versions of all subpackages (e.g. `itertools`, `mpi`, `h5`, `nda`, etc.). Long story short. **Please install TRIQS on the latest `triqs/DEV_PPSC_*` branch.**
-
-#. The `adapol <https://github.com/flatironinstitute/adapol>`_ package (Adaptive Pole Fitting for Quantum Many-Body Physics) has to be installed on the `main` branch. This can be achieved using a local `pip` install, e.g.::
+   We are currently depending on the `unstable` development branch of TRIQS that has to be compiled from source. (Note that the packaged versions of TRIQS are not supported.)
+   
+#. The `adapol <https://github.com/flatironinstitute/adapol>`_ package (Adaptive Pole Fitting for Quantum Many-Body Physics) has to be installed on the `triqs_xca` branch. This can be achieved using a local `pip` install, e.g.::
 
-    $ git clone git@github.com:flatironinstitute/adapol.git --branch main
+    $ git clone git@github.com:flatironinstitute/adapol.git --branch triqs_xca
     $ python -m pip install -e ./adapol
 
 #. The `pyed <https://github.com/HugoStrand/pyed>`_ package (PYED: Exact diagonalization for finite quantum systems) has to be available in the python environment. This can be tested by running::