Modularization of sagelib: Break out separate package sagemath-symbolics
Opened this issue · 9 comments
mkoeppe commented
After merging the pynac sources as src/sage/symbolic/ginac (#32386), we are able to create a distribution package sagemath-symbolics, containing the integrated pynac from that ticket and most of
sage.symbolic,sage.calculus,sage.functions,sage.interfaces(parts),sage.rings.asymptotic,sage.geometry.hyperbolic_spacesage.manifolds
There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, sage.graphs has 0 imports from sage.symbolic, sage.combinat just a handful, etc. (Removing unnecessary uses of SR and symbolic functions is done in #32411, #32413, #32415, #32416, ...; refactoring to avoid dependencies via is_... functions and similar is addressed in #32414.)
As of this ticket, sagemath-symbolics will be a subset distribution that is a superset of sagemath-categories (#29865), similar to sagemath-linbox (#31444) and sagemath-polyhedra (#32432). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
Like sagemath-linbox, also sagemath-symbolics has non-pip-installable dependencies: ecl, maxima, giac, singular (for pynac's use of libfactory)
See also:
- #32601 Modularization of sagelib: Break out a separate package sagemath-standard-no-symbolics
Depends on #29865
Depends on #32386
Depends on #32637
CC: @dimpase @tscrim @kliem @egourgoulhon
Component: refactoring
Issue created by migration from https://trac.sagemath.org/ticket/31695
mkoeppe commented
Description changed:
---
+++
@@ -1,5 +1,8 @@
-We prepare a prototype package based on `sagemath-categories`, similar to `sagemath-linbox` (#31444), which packages everything related to the `SR`, with minimal dependencies.
+After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package `sagemath-symbolics`, containing the integrated pynac from the present ticket and most of `sage.symbolic`, `sage.calculus`, `sage.functions`, and parts of `sage.interfaces`.
-We use it for the Continuous Integration workflow of `pynac`.
+There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (#32411 tracks the task of removing unnecessary uses of `SR` and symbolic functions.)
+As of this ticket, `sagemath-symbolics` will be a subset distribution that is a superset of `sagemath-categories` (#29865), similar to `sagemath-linbox` (#31444). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
+Like `sagemath-linbox`, also `sagemath-symbolics` has non-pip-installable dependencies: `ecl`, `maxima`, `giac`, `singular` (for pynac's use of libfactory)
+mkoeppe commented
Description changed:
---
+++
@@ -1,6 +1,12 @@
-After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package `sagemath-symbolics`, containing the integrated pynac from the present ticket and most of `sage.symbolic`, `sage.calculus`, `sage.functions`, and parts of `sage.interfaces`.
+After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package `sagemath-symbolics`, containing the integrated pynac from the present ticket and most of
+- `sage.symbolic`,
+- `sage.calculus`,
+- `sage.functions`,
+- `sage.interfaces` (parts),
+- `sage.rings.asymptotic`,
+- `sage.manifolds`
-There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (#32411 tracks the task of removing unnecessary uses of `SR` and symbolic functions.)
+There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (removing unnecessary uses of `SR` and symbolic functions is done in #32411, #32413, ...)
As of this ticket, `sagemath-symbolics` will be a subset distribution that is a superset of `sagemath-categories` (#29865), similar to `sagemath-linbox` (#31444). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
mkoeppe commented
Description changed:
---
+++
@@ -4,6 +4,7 @@
- `sage.functions`,
- `sage.interfaces` (parts),
- `sage.rings.asymptotic`,
+- `sage.geometry.hyperbolic_space`
- `sage.manifolds`
There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (removing unnecessary uses of `SR` and symbolic functions is done in #32411, #32413, ...)mkoeppe commented
Description changed:
---
+++
@@ -7,7 +7,7 @@
- `sage.geometry.hyperbolic_space`
- `sage.manifolds`
-There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (removing unnecessary uses of `SR` and symbolic functions is done in #32411, #32413, ...)
+There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (Removing unnecessary uses of `SR` and symbolic functions is done in #32411, #32413, #32415, #32416, ...; refactoring to avoid dependencies via `is_...` functions and similar is addressed in #32414.)
As of this ticket, `sagemath-symbolics` will be a subset distribution that is a superset of `sagemath-categories` (#29865), similar to `sagemath-linbox` (#31444). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
mkoeppe commented
Description changed:
---
+++
@@ -1,4 +1,4 @@
-After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package `sagemath-symbolics`, containing the integrated pynac from the present ticket and most of
+After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package `sagemath-symbolics`, containing the integrated pynac from that ticket and most of
- `sage.symbolic`,
- `sage.calculus`,
- `sage.functions`, mkoeppe commented
Description changed:
---
+++
@@ -9,7 +9,7 @@
There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (Removing unnecessary uses of `SR` and symbolic functions is done in #32411, #32413, #32415, #32416, ...; refactoring to avoid dependencies via `is_...` functions and similar is addressed in #32414.)
-As of this ticket, `sagemath-symbolics` will be a subset distribution that is a superset of `sagemath-categories` (#29865), similar to `sagemath-linbox` (#31444). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
+As of this ticket, `sagemath-symbolics` will be a subset distribution that is a superset of `sagemath-categories` (#29865), similar to `sagemath-linbox` (#31444) and `sagemath-polyhedra` (#32432). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
Like `sagemath-linbox`, also `sagemath-symbolics` has non-pip-installable dependencies: `ecl`, `maxima`, `giac`, `singular` (for pynac's use of libfactory)
mkoeppe commented
Description changed:
---
+++
@@ -1,4 +1,4 @@
-After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package `sagemath-symbolics`, containing the integrated pynac from that ticket and most of
+After merging the pynac sources as `src/sage/symbolic/ginac` (#32386), we are able to create a distribution package **sagemath-symbolics**, containing the integrated pynac from that ticket and most of
- `sage.symbolic`,
- `sage.calculus`,
- `sage.functions`,
@@ -9,7 +9,9 @@
There are many parts of Sage that (fortunately!) do not depend at all on Sage symbolics, or only depend on it for some smaller features. For example, `sage.graphs` has 0 imports from `sage.symbolic`, `sage.combinat` just a handful, etc. (Removing unnecessary uses of `SR` and symbolic functions is done in #32411, #32413, #32415, #32416, ...; refactoring to avoid dependencies via `is_...` functions and similar is addressed in #32414.)
-As of this ticket, `sagemath-symbolics` will be a subset distribution that is a superset of `sagemath-categories` (#29865), similar to `sagemath-linbox` (#31444) and `sagemath-polyhedra` (#32432). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
+As of this ticket, **sagemath-symbolics** will be a subset distribution that is a superset of **sagemath-categories** (#29865), similar to **sagemath-linbox** (#31444) and **sagemath-polyhedra** (#32432). (Later, when the namespace package mechanism is in place, it will become a native namespace package.)
-Like `sagemath-linbox`, also `sagemath-symbolics` has non-pip-installable dependencies: `ecl`, `maxima`, `giac`, `singular` (for pynac's use of libfactory)
+Like **sagemath-linbox**, also **sagemath-symbolics** has non-pip-installable dependencies: `ecl`, `maxima`, `giac`, `singular` (for pynac's use of libfactory)
+See also:
+- #32601 Modularization of sagelib: Break out a separate **package sagemath-standard-no-symbolics**