From d511f1af00ebc22bc39ef2f31cc8663e8d62d076 Mon Sep 17 00:00:00 2001
From: Max Horn <max@quendi.de>
Date: Tue, 4 Jun 2024 00:28:31 +0200
Subject: [PATCH 1/4] Fix typo

---
 src/HeckeMoreStuff.jl | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/HeckeMoreStuff.jl b/src/HeckeMoreStuff.jl
index 91aa6f528..b20b5712b 100644
--- a/src/HeckeMoreStuff.jl
+++ b/src/HeckeMoreStuff.jl
@@ -698,7 +698,7 @@ function Base.Int128(x::ZZRingElem)
   return Base.Int128(BigInt(x))
 end
 
-## Make zzModRing iteratible
+## Make zzModRing iterable
 
 Base.iterate(R::zzModRing) = (zero(R), zero(UInt))
 
@@ -716,7 +716,7 @@ Base.eltype(::Type{zzModRing}) = zzModRingElem
 Base.IteratorSize(::Type{zzModRing}) = Base.HasLength()
 Base.length(R::zzModRing) = R.n
 
-## Make ZZModRing iteratible
+## Make ZZModRing iterable
 
 Base.iterate(R::ZZModRing) = (zero(R), zero(ZZRingElem))
 
@@ -734,7 +734,7 @@ Base.eltype(::Type{ZZModRing}) = ZZModRingElem
 Base.IteratorSize(::Type{ZZModRing}) = Base.HasLength()
 Base.length(R::ZZModRing) = Integer(R.n)
 
-## Make fpField iteratible
+## Make fpField iterable
 
 Base.iterate(R::fpField) = (zero(R), zero(UInt))
 

From e9293e4b2586f2c08998c28b66252bf012fadcc1 Mon Sep 17 00:00:00 2001
From: Max Horn <max@quendi.de>
Date: Tue, 4 Jun 2024 00:29:21 +0200
Subject: [PATCH 2/4] Move some ZZMatrix code to fmpz_mat.jl

---
 src/HeckeMiscMatrix.jl | 197 -----------------------------------------
 src/flint/fmpz_mat.jl  | 190 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 190 insertions(+), 197 deletions(-)

diff --git a/src/HeckeMiscMatrix.jl b/src/HeckeMiscMatrix.jl
index 4a2dc4f72..f0fe8816e 100644
--- a/src/HeckeMiscMatrix.jl
+++ b/src/HeckeMiscMatrix.jl
@@ -63,16 +63,6 @@ function is_zero_row(M::Matrix{ZZRingElem}, i::Int)
 end
 
 
-function divexact!(a::ZZMatrix, b::ZZMatrix, d::ZZRingElem)
-  ccall((:fmpz_mat_scalar_divexact_fmpz, libflint), Nothing,
-        (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), a, b, d)
-end
-
-function mul!(a::ZZMatrix, b::ZZMatrix, c::ZZRingElem)
-  ccall((:fmpz_mat_scalar_mul_fmpz, libflint), Nothing,
-        (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), a, b, c)
-end
-
 ################################################################################
 #
 ################################################################################
@@ -178,59 +168,6 @@ end
 #
 ################################################################################
 
-@doc raw"""
-    mod!(M::ZZMatrix, p::ZZRingElem)
-
-Reduces every entry modulo $p$ in-place, i.e. applies the mod function to every entry.
-Positive residue system.
-"""
-function mod!(M::ZZMatrix, p::ZZRingElem)
-  GC.@preserve M begin
-    for i = 1:nrows(M)
-      for j = 1:ncols(M)
-        z = ccall((:fmpz_mat_entry, libflint), Ptr{ZZRingElem}, (Ref{ZZMatrix}, Int, Int), M, i - 1, j - 1)
-        ccall((:fmpz_mod, libflint), Nothing, (Ptr{ZZRingElem}, Ptr{ZZRingElem}, Ref{ZZRingElem}), z, z, p)
-      end
-    end
-  end
-  return nothing
-end
-
-@doc raw"""
-    mod(M::ZZMatrix, p::ZZRingElem) -> ZZMatrix
-
-Reduces every entry modulo $p$, i.e. applies the mod function to every entry.
-"""
-function mod(M::ZZMatrix, p::ZZRingElem)
-  N = deepcopy(M)
-  mod!(N, p)
-  return N
-end
-
-@doc raw"""
-    mod_sym!(M::ZZMatrix, p::ZZRingElem)
-
-Reduces every entry modulo $p$ in-place, into the symmetric residue system.
-"""
-function mod_sym!(M::ZZMatrix, B::ZZRingElem)
-  @assert !iszero(B)
-  ccall((:fmpz_mat_scalar_smod, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), M, M, B)
-end
-mod_sym!(M::ZZMatrix, B::Integer) = mod_sym!(M, ZZRingElem(B))
-
-@doc raw"""
-    mod_sym(M::ZZMatrix, p::ZZRingElem) -> ZZMatrix
-
-Reduces every entry modulo $p$ into the symmetric residue system.
-"""
-function mod_sym(M::ZZMatrix, B::ZZRingElem)
-  N = zero_matrix(ZZ, nrows(M), ncols(M))
-  ccall((:fmpz_mat_scalar_smod, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), N, M, B)
-  return N
-end
-mod_sym(M::ZZMatrix, B::Integer) = mod_sym(M, ZZRingElem(B))
-
-
 @doc raw"""
     mod_sym!(A::Generic.Mat{AbsSimpleNumFieldElem}, m::ZZRingElem)
 
@@ -244,140 +181,6 @@ function mod_sym!(A::Generic.Mat{AbsSimpleNumFieldElem}, m::ZZRingElem)
   end
 end
 
-################################################################################
-#
-#  In-place HNF
-#
-################################################################################
-
-function hnf!(x::ZZMatrix)
-  if nrows(x) * ncols(x) > 100
-    z = hnf(x)
-    ccall((:fmpz_mat_set, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}), x, z)
-
-    return x
-  end
-  ccall((:fmpz_mat_hnf, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}), x, x)
-  return x
-end
-
-################################################################################
-#
-#  Smith normal form with trafo
-#
-################################################################################
-
-#=
-g, e,f = gcdx(a, b)
-U = [1 0 ; -divexact(b, g)*f 1]*[1 1; 0 1];
-V = [e -divexact(b, g) ; f divexact(a, g)];
-
-then U*[ a 0; 0 b] * V = [g 0 ; 0 l]
-=#
-@doc raw"""
-    snf_with_transform(A::ZZMatrix, l::Bool = true, r::Bool = true) -> ZZMatrix, ZZMatrix, ZZMatrix
-
-Given some integer matrix $A$, compute the Smith normal form (elementary
-divisor normal form) of $A$. If `l` and/ or `r` are true, then the corresponding
-left and/ or right transformation matrices are computed as well.
-"""
-function snf_with_transform(A::ZZMatrix, l::Bool=true, r::Bool=true)
-  if r
-    R = identity_matrix(ZZ, ncols(A))
-  end
-
-  if l
-    L = identity_matrix(ZZ, nrows(A))
-  end
-  # TODO: if only one trafo is required, start with the HNF that does not
-  #       compute the trafo
-  #       Rationale: most of the work is on the 1st HNF..
-  S = deepcopy(A)
-  while !is_diagonal(S)
-    if l
-      S, T = hnf_with_transform(S)
-      L = T * L
-    else
-      S = hnf!(S)
-    end
-
-    if is_diagonal(S)
-      break
-    end
-    if r
-      S, T = hnf_with_transform(transpose(S))
-      R = T * R
-    else
-      S = hnf!(transpose(S))
-    end
-    S = transpose(S)
-  end
-  #this is probably not really optimal...
-  for i = 1:min(nrows(S), ncols(S))
-    if S[i, i] == 1
-      continue
-    end
-    for j = i+1:min(nrows(S), ncols(S))
-      if S[j, j] == 0
-        continue
-      end
-      if S[i, i] != 0 && S[j, j] % S[i, i] == 0
-        continue
-      end
-      g, e, f = gcdx(S[i, i], S[j, j])
-      a = divexact(S[i, i], g)
-      S[i, i] = g
-      b = divexact(S[j, j], g)
-      S[j, j] *= a
-      if l
-        # U = [1 0; -b*f 1] * [ 1 1; 0 1] = [1 1; -b*f -b*f+1]
-        # so row i and j of L will be transformed. We do it naively
-        # those 2x2 transformations of 2 rows should be a c-primitive
-        # or at least a Nemo/Hecke primitive
-        for k = 1:ncols(L)
-          x = -b * f
-          #          L[i,k], L[j,k] = L[i,k]+L[j,k], x*L[i,k]+(x+1)*L[j,k]
-          L[i, k], L[j, k] = L[i, k] + L[j, k], x * (L[i, k] + L[j, k]) + L[j, k]
-        end
-      end
-      if r
-        # V = [e -b ; f a];
-        # so col i and j of R will be transformed. We do it naively
-        # careful: at this point, R is still transposed
-        for k = 1:nrows(R)
-          R[i, k], R[j, k] = e * R[i, k] + f * R[j, k], -b * R[i, k] + a * R[j, k]
-        end
-      end
-    end
-  end
-
-  # It might be the case that S was diagonal with negative diagonal entries.
-  for i in 1:min(nrows(S), ncols(S))
-    if S[i, i] < 0
-      if l
-        multiply_row!(L, ZZRingElem(-1), i)
-      end
-      S[i, i] = -S[i, i]
-    end
-  end
-
-  if l
-    if r
-      return S, L, transpose(R)
-    else
-      # last is dummy
-      return S, L, L
-    end
-  elseif r
-    # second is dummy
-    return S, R, transpose(R)
-  else
-    # last two are dummy
-    return S, S, S
-  end
-end
-
-
 #Returns a positive integer if A[i, j] > b, negative if A[i, j] < b, 0 otherwise
 function compare_index(A::ZZMatrix, i::Int, j::Int, b::ZZRingElem)
   a = ccall((:fmpz_mat_entry, libflint), Ptr{ZZRingElem}, (Ref{ZZMatrix}, Int, Int), A, i - 1, j - 1)
diff --git a/src/flint/fmpz_mat.jl b/src/flint/fmpz_mat.jl
index 19e48b999..5981f4507 100644
--- a/src/flint/fmpz_mat.jl
+++ b/src/flint/fmpz_mat.jl
@@ -580,6 +580,11 @@ end
 
 divexact(x::ZZMatrix, y::Integer; check::Bool=true) = divexact(x, ZZRingElem(y); check=check)
 
+function divexact!(a::ZZMatrix, b::ZZMatrix, d::ZZRingElem)
+  ccall((:fmpz_mat_scalar_divexact_fmpz, libflint), Nothing,
+        (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), a, b, d)
+end
+
 ###############################################################################
 #
 #   Kronecker product
@@ -618,6 +623,58 @@ Reduce the entries of $x$ modulo $y$ and return the result.
 """
 reduce_mod(x::ZZMatrix, y::Integer) = reduce_mod(x, ZZRingElem(y))
 
+@doc raw"""
+    mod!(M::ZZMatrix, p::ZZRingElem)
+
+Reduces every entry modulo $p$ in-place, i.e. applies the mod function to every entry.
+Positive residue system.
+"""
+function mod!(M::ZZMatrix, p::ZZRingElem)
+  GC.@preserve M begin
+    for i = 1:nrows(M)
+      for j = 1:ncols(M)
+        z = ccall((:fmpz_mat_entry, libflint), Ptr{ZZRingElem}, (Ref{ZZMatrix}, Int, Int), M, i - 1, j - 1)
+        ccall((:fmpz_mod, libflint), Nothing, (Ptr{ZZRingElem}, Ptr{ZZRingElem}, Ref{ZZRingElem}), z, z, p)
+      end
+    end
+  end
+  return nothing
+end
+
+@doc raw"""
+    mod(M::ZZMatrix, p::ZZRingElem) -> ZZMatrix
+
+Reduces every entry modulo $p$, i.e. applies the mod function to every entry.
+"""
+function mod(M::ZZMatrix, p::ZZRingElem)
+  N = deepcopy(M)
+  mod!(N, p)
+  return N
+end
+
+@doc raw"""
+    mod_sym!(M::ZZMatrix, p::ZZRingElem)
+
+Reduces every entry modulo $p$ in-place, into the symmetric residue system.
+"""
+function mod_sym!(M::ZZMatrix, B::ZZRingElem)
+  @assert !iszero(B)
+  ccall((:fmpz_mat_scalar_smod, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), M, M, B)
+end
+mod_sym!(M::ZZMatrix, B::Integer) = mod_sym!(M, ZZRingElem(B))
+
+@doc raw"""
+    mod_sym(M::ZZMatrix, p::ZZRingElem) -> ZZMatrix
+
+Reduces every entry modulo $p$ into the symmetric residue system.
+"""
+function mod_sym(M::ZZMatrix, B::ZZRingElem)
+  N = zero_matrix(ZZ, nrows(M), ncols(M))
+  ccall((:fmpz_mat_scalar_smod, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), N, M, B)
+  return N
+end
+mod_sym(M::ZZMatrix, B::Integer) = mod_sym(M, ZZRingElem(B))
+
 ###############################################################################
 #
 #   Fraction free LU decomposition
@@ -864,6 +921,17 @@ function __hnf(x)
   return z
 end
 
+function hnf!(x::ZZMatrix)
+  if nrows(x) * ncols(x) > 100
+    z = hnf(x)
+    ccall((:fmpz_mat_set, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}), x, z)
+
+    return x
+  end
+  ccall((:fmpz_mat_hnf, libflint), Nothing, (Ref{ZZMatrix}, Ref{ZZMatrix}), x, x)
+  return x
+end
+
 @doc raw"""
     hnf_with_transform(x::ZZMatrix)
 
@@ -1216,6 +1284,122 @@ function is_snf(x::ZZMatrix)
                (Ref{ZZMatrix},), x)
 end
 
+################################################################################
+#
+#  Smith normal form with trafo
+#
+################################################################################
+
+#=
+g, e,f = gcdx(a, b)
+U = [1 0 ; -divexact(b, g)*f 1]*[1 1; 0 1];
+V = [e -divexact(b, g) ; f divexact(a, g)];
+
+then U*[ a 0; 0 b] * V = [g 0 ; 0 l]
+=#
+@doc raw"""
+    snf_with_transform(A::ZZMatrix, l::Bool = true, r::Bool = true) -> ZZMatrix, ZZMatrix, ZZMatrix
+
+Given some integer matrix $A$, compute the Smith normal form (elementary
+divisor normal form) of $A$. If `l` and/ or `r` are true, then the corresponding
+left and/ or right transformation matrices are computed as well.
+"""
+function snf_with_transform(A::ZZMatrix, l::Bool=true, r::Bool=true)
+  if r
+    R = identity_matrix(ZZ, ncols(A))
+  end
+
+  if l
+    L = identity_matrix(ZZ, nrows(A))
+  end
+  # TODO: if only one trafo is required, start with the HNF that does not
+  #       compute the trafo
+  #       Rationale: most of the work is on the 1st HNF..
+  S = deepcopy(A)
+  while !is_diagonal(S)
+    if l
+      S, T = hnf_with_transform(S)
+      L = T * L
+    else
+      S = hnf!(S)
+    end
+
+    if is_diagonal(S)
+      break
+    end
+    if r
+      S, T = hnf_with_transform(transpose(S))
+      R = T * R
+    else
+      S = hnf!(transpose(S))
+    end
+    S = transpose(S)
+  end
+  #this is probably not really optimal...
+  for i = 1:min(nrows(S), ncols(S))
+    if S[i, i] == 1
+      continue
+    end
+    for j = i+1:min(nrows(S), ncols(S))
+      if S[j, j] == 0
+        continue
+      end
+      if S[i, i] != 0 && S[j, j] % S[i, i] == 0
+        continue
+      end
+      g, e, f = gcdx(S[i, i], S[j, j])
+      a = divexact(S[i, i], g)
+      S[i, i] = g
+      b = divexact(S[j, j], g)
+      S[j, j] *= a
+      if l
+        # U = [1 0; -b*f 1] * [ 1 1; 0 1] = [1 1; -b*f -b*f+1]
+        # so row i and j of L will be transformed. We do it naively
+        # those 2x2 transformations of 2 rows should be a c-primitive
+        # or at least a Nemo/Hecke primitive
+        for k = 1:ncols(L)
+          x = -b * f
+          #          L[i,k], L[j,k] = L[i,k]+L[j,k], x*L[i,k]+(x+1)*L[j,k]
+          L[i, k], L[j, k] = L[i, k] + L[j, k], x * (L[i, k] + L[j, k]) + L[j, k]
+        end
+      end
+      if r
+        # V = [e -b ; f a];
+        # so col i and j of R will be transformed. We do it naively
+        # careful: at this point, R is still transposed
+        for k = 1:nrows(R)
+          R[i, k], R[j, k] = e * R[i, k] + f * R[j, k], -b * R[i, k] + a * R[j, k]
+        end
+      end
+    end
+  end
+
+  # It might be the case that S was diagonal with negative diagonal entries.
+  for i in 1:min(nrows(S), ncols(S))
+    if S[i, i] < 0
+      if l
+        multiply_row!(L, ZZRingElem(-1), i)
+      end
+      S[i, i] = -S[i, i]
+    end
+  end
+
+  if l
+    if r
+      return S, L, transpose(R)
+    else
+      # last is dummy
+      return S, L, L
+    end
+  elseif r
+    # second is dummy
+    return S, R, transpose(R)
+  else
+    # last two are dummy
+    return S, S, S
+  end
+end
+
 ###############################################################################
 #
 #   manual linear algebra: row and col operations
@@ -1676,6 +1860,12 @@ function mul!(y::ZZMatrix, x::ZZRingElem)
   return y
 end
 
+function mul!(z::ZZMatrix, y::ZZMatrix, x::ZZRingElem)
+  ccall((:fmpz_mat_scalar_mul_fmpz, libflint), Nothing,
+        (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), z, y, x)
+  return z
+end
+
 function addmul!(z::ZZMatrix, y::ZZMatrix, x::ZZRingElem)
   ccall((:fmpz_mat_scalar_addmul_fmpz, libflint), Nothing,
         (Ref{ZZMatrix}, Ref{ZZMatrix}, Ref{ZZRingElem}), z, y, x)

From 54517955fa282cbdcae3a3aed0c6729507ecd3cc Mon Sep 17 00:00:00 2001
From: Max Horn <max@quendi.de>
Date: Tue, 4 Jun 2024 00:30:29 +0200
Subject: [PATCH 3/4] Enable 'f(x)' support where f is an arb/acb/real/complex
 poly

---
 src/HeckeMiscPoly.jl |  4 ----
 src/polysubst.jl     | 16 +++++++++++++++-
 2 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/src/HeckeMiscPoly.jl b/src/HeckeMiscPoly.jl
index ace4136d0..c1270f3d9 100644
--- a/src/HeckeMiscPoly.jl
+++ b/src/HeckeMiscPoly.jl
@@ -407,10 +407,6 @@ function mulhigh(a::PolyRingElem{T}, b::PolyRingElem{T}, n::Int) where {T}
   return mulhigh_n(a, b, degree(a) + degree(b) - n)
 end
 
-function (f::AcbPolyRingElem)(x::AcbFieldElem)
-  return evaluate(f, x)
-end
-
 function mod(f::AbstractAlgebra.PolyRingElem{T}, g::AbstractAlgebra.PolyRingElem{T}) where {T<:RingElem}
   check_parent(f, g)
   if length(g) == 0
diff --git a/src/polysubst.jl b/src/polysubst.jl
index 6b71138da..a45cfdde9 100644
--- a/src/polysubst.jl
+++ b/src/polysubst.jl
@@ -1,4 +1,18 @@
-for T in [zzModPolyRingElem, fpPolyRingElem, ZZModPolyRingElem, FpPolyRingElem, QQPolyRingElem, ZZPolyRingElem, FqPolyRepPolyRingElem, fqPolyRepPolyRingElem, FqPolyRingElem]
+for T in [
+  zzModPolyRingElem,
+  fpPolyRingElem,
+  ZZModPolyRingElem,
+  FpPolyRingElem,
+  QQPolyRingElem,
+  ZZPolyRingElem,
+  FqPolyRepPolyRingElem,
+  fqPolyRepPolyRingElem,
+  FqPolyRingElem,
+  AcbPolyRingElem,
+  ArbPolyRingElem,
+  ComplexPoly,
+  RealPoly,
+  ]
   (f::T)(a) = subst(f, a)
 
   function (f::T)(a::T)

From 5887a5a1e74a8cb17529b393ea0541f4f502681a Mon Sep 17 00:00:00 2001
From: Max Horn <max@quendi.de>
Date: Tue, 4 Jun 2024 00:31:16 +0200
Subject: [PATCH 4/4] Remove redundant is_zero_row methods

The generic ones in AbstractAlgebra already takes care of these,
and are equally efficient thanks to is_zero_entry
---
 src/HeckeMiscMatrix.jl | 23 -----------------------
 1 file changed, 23 deletions(-)

diff --git a/src/HeckeMiscMatrix.jl b/src/HeckeMiscMatrix.jl
index f0fe8816e..33982c682 100644
--- a/src/HeckeMiscMatrix.jl
+++ b/src/HeckeMiscMatrix.jl
@@ -29,19 +29,6 @@ function Array(a::ZZMatrix; S::Type{T}=ZZRingElem) where {T}
   return A
 end
 
-function is_zero_row(M::ZZMatrix, i::Int)
-  GC.@preserve M begin
-    for j = 1:ncols(M)
-      m = ccall((:fmpz_mat_entry, libflint), Ptr{ZZRingElem}, (Ref{ZZMatrix}, Int, Int), M, i - 1, j - 1)
-      fl = ccall((:fmpz_is_zero, libflint), Bool, (Ptr{ZZRingElem},), m)
-      if !fl
-        return false
-      end
-    end
-  end
-  return true
-end
-
 function is_zero_row(M::zzModMatrix, i::Int)
   zero = UInt(0)
   for j in 1:ncols(M)
@@ -53,16 +40,6 @@ function is_zero_row(M::zzModMatrix, i::Int)
   return true
 end
 
-function is_zero_row(M::Matrix{ZZRingElem}, i::Int)
-  for j = 1:Base.size(M, 2)
-    if !iszero(M[i, j])
-      return false
-    end
-  end
-  return true
-end
-
-
 ################################################################################
 #
 ################################################################################