pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = ty `+` calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well Default is 0755 or 0644, de- pending on the blob. 913 914 CHAPTER 79. LIBGIT2 • file_open_flags: bi�lags used to open any files during the checkout. • notify_flags: Flags for what sort of conflicts

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = ty `+` calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well Default is 0755 or 0644, de- pending on the blob. 885 886 CHAPTER 77. LIBGIT2 • file_open_flags: bi�lags used to open any files during the checkout. • notify_flags: Flags for what sort of conflicts

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = ty `+` calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well Default is 0755 or 0644, de- pending on the blob. 883 884 CHAPTER 77. LIBGIT2 • file_open_flags: bi�lags used to open any files during the checkout. • notify_flags: Flags for what sort of conflicts

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = ty `+` calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well Default is 0755 or 0644, de- pending on the blob. 915 916 CHAPTER 80. LIBGIT2 • file_open_flags: bi�lags used to open any files during the checkout. • notify_flags: Flags for what sort of conflicts

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = ty `+` calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well Default is 0755 or 0644, de- pending on the blob. 923 924 CHAPTER 82. LIBGIT2 • file_open_flags: bi�lags used to open any files during the checkout. • notify_flags: Flags for what sort of conflicts

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = ty `+` calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well Default is 0755 or 0644, de- pending on the blob. 921 922 CHAPTER 81. LIBGIT2 • file_open_flags: bi�lags used to open any files during the checkout. • notify_flags: Flags for what sort of conflicts

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = t calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well similar(::SpecialSparseMatrix, ::Type, ::Dims) returns a dense zero matrix. As a consequence, products of bi-, tri- and symmetric tridiagonal matrices with each other result in dense output. Moreover, constructing

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = t calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well similar(::SpecialSparseMatrix, ::Type, ::Dims) returns a dense zero matrix. As a consequence, products of bi-, tri- and symmetric tridiagonal matrices with each other result in dense output. Moreover, constructing

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = t calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well similar(::SpecialSparseMatrix, ::Type, ::Dims) returns a dense zero matrix. As a consequence, products of bi-, tri- and symmetric tridiagonal matrices with each other result in dense output. Moreover, constructing

pseudo-code might look like: function matmul(a::AbstractMatrix, b::AbstractMatrix) op = (ai, bi) -> ai * bi + ai * bi ## this is insufficient because it assumes `one(eltype(a))` is constructable: # R = t calls `promote_type` ## but this is not true for some types, such as Bool: # R = promote_type(ai, bi) # this is wrong, since depending on the return value # of type-inference is very brittle (as well similar(::SpecialSparseMatrix, ::Type, ::Dims) returns a dense zero matrix. As a consequence, products of bi-, tri- and symmetric tridiagonal matrices with each other result in dense output. Moreover, constructing