Re: [PATCH] c++: stream PACK_EXPANSION_EXTRA_ARGS [PR106761]
On 9/20/22 10:08, Patrick Palka wrote: On Tue, 20 Sep 2022, Nathan Sidwell wrote: On 9/19/22 09:52, Patrick Palka wrote: It looks like some xtreme-header-* tests are failing after the libstdc++ change r13-2158-g02f6b405f0e9dc ultimately because we're neglecting to stream PACK_EXPANSION_EXTRA_ARGS, which leads to false equivalences of different partial instantiations of _TupleConstraints::__constructible. Tested on x86_64-pc-linux-gnu, does this look OK for trunk? PR c++/106761 gcc/cp/ChangeLog: * module.cc (trees_out::type_node) : Stream PACK_EXPANSION_EXTRA_ARGS. (trees_in::tree_node) : Likewise. Looks good, I wonder why I missed that. (I guess extracting a testcase out of the headers was too tricky?) Many thanks. I managed to produce a small testcase which mirrors the format of the xtreme-header-2* testcase. Does the following look OK? yup, thanks for extracting that! nathan -- >8 -- PR c++/106761 gcc/cp/ChangeLog: * module.cc (trees_out::type_node) : Stream PACK_EXPANSION_EXTRA_ARGS. (trees_in::tree_node) : Likewise. gcc/testsuite/ChangeLog: * g++.dg/modules/pr106761.h: New test. * g++.dg/modules/pr106761_a.H: New test. * g++.dg/modules/pr106761_b.C: New test. --- gcc/cp/module.cc | 3 +++ gcc/testsuite/g++.dg/modules/pr106761.h | 22 ++ gcc/testsuite/g++.dg/modules/pr106761_a.H | 5 + gcc/testsuite/g++.dg/modules/pr106761_b.C | 7 +++ 4 files changed, 37 insertions(+) create mode 100644 gcc/testsuite/g++.dg/modules/pr106761.h create mode 100644 gcc/testsuite/g++.dg/modules/pr106761_a.H create mode 100644 gcc/testsuite/g++.dg/modules/pr106761_b.C diff --git a/gcc/cp/module.cc b/gcc/cp/module.cc index 1a1ff5be574..9a9ef4e3332 100644 --- a/gcc/cp/module.cc +++ b/gcc/cp/module.cc @@ -8922,6 +8922,7 @@ trees_out::type_node (tree type) if (streaming_p ()) u (PACK_EXPANSION_LOCAL_P (type)); tree_node (PACK_EXPANSION_PARAMETER_PACKS (type)); + tree_node (PACK_EXPANSION_EXTRA_ARGS (type)); break; case TYPENAME_TYPE: @@ -9455,12 +9456,14 @@ trees_in::tree_node (bool is_use) { bool local = u (); tree param_packs = tree_node (); + tree extra_args = tree_node (); if (!get_overrun ()) { tree expn = cxx_make_type (TYPE_PACK_EXPANSION); SET_TYPE_STRUCTURAL_EQUALITY (expn); PACK_EXPANSION_PATTERN (expn) = res; PACK_EXPANSION_PARAMETER_PACKS (expn) = param_packs; + PACK_EXPANSION_EXTRA_ARGS (expn) = extra_args; PACK_EXPANSION_LOCAL_P (expn) = local; res = expn; } diff --git a/gcc/testsuite/g++.dg/modules/pr106761.h b/gcc/testsuite/g++.dg/modules/pr106761.h new file mode 100644 index 000..9f22a22a45d --- /dev/null +++ b/gcc/testsuite/g++.dg/modules/pr106761.h @@ -0,0 +1,22 @@ +// PR c++/106761 + +template +struct __and_; + +template +struct is_convertible; + +template +struct _TupleConstraints { + template + using __constructible = __and_...>; +}; + +template +struct tuple { + template + using __constructible += typename _TupleConstraints::template __constructible; +}; + +tuple t; diff --git a/gcc/testsuite/g++.dg/modules/pr106761_a.H b/gcc/testsuite/g++.dg/modules/pr106761_a.H new file mode 100644 index 000..8ad116412af --- /dev/null +++ b/gcc/testsuite/g++.dg/modules/pr106761_a.H @@ -0,0 +1,5 @@ +// PR c++/106761 +// { dg-additional-options -fmodule-header } + +// { dg-module-cmi {} } +#include "pr106761.h" diff --git a/gcc/testsuite/g++.dg/modules/pr106761_b.C b/gcc/testsuite/g++.dg/modules/pr106761_b.C new file mode 100644 index 000..336cb12757e --- /dev/null +++ b/gcc/testsuite/g++.dg/modules/pr106761_b.C @@ -0,0 +1,7 @@ +// PR c++/106761 +// { dg-additional-options -fmodules-ts } + +#include "pr106761.h" +import "pr106761_a.H"; + +tuple u = t; -- Nathan Sidwell
Re: [PATCH] c++: stream PACK_EXPANSION_EXTRA_ARGS [PR106761]
On Tue, 20 Sep 2022, Nathan Sidwell wrote: > On 9/19/22 09:52, Patrick Palka wrote: > > It looks like some xtreme-header-* tests are failing after the libstdc++ > > change r13-2158-g02f6b405f0e9dc ultimately because we're neglecting > > to stream PACK_EXPANSION_EXTRA_ARGS, which leads to false equivalences > > of different partial instantiations of _TupleConstraints::__constructible. > > > > Tested on x86_64-pc-linux-gnu, does this look OK for trunk? > > > > PR c++/106761 > > > > gcc/cp/ChangeLog: > > > > * module.cc (trees_out::type_node) : > > Stream PACK_EXPANSION_EXTRA_ARGS. > > (trees_in::tree_node) : Likewise. > > > Looks good, I wonder why I missed that. (I guess extracting a testcase out of > the headers was too tricky?) Many thanks. I managed to produce a small testcase which mirrors the format of the xtreme-header-2* testcase. Does the following look OK? -- >8 -- PR c++/106761 gcc/cp/ChangeLog: * module.cc (trees_out::type_node) : Stream PACK_EXPANSION_EXTRA_ARGS. (trees_in::tree_node) : Likewise. gcc/testsuite/ChangeLog: * g++.dg/modules/pr106761.h: New test. * g++.dg/modules/pr106761_a.H: New test. * g++.dg/modules/pr106761_b.C: New test. --- gcc/cp/module.cc | 3 +++ gcc/testsuite/g++.dg/modules/pr106761.h | 22 ++ gcc/testsuite/g++.dg/modules/pr106761_a.H | 5 + gcc/testsuite/g++.dg/modules/pr106761_b.C | 7 +++ 4 files changed, 37 insertions(+) create mode 100644 gcc/testsuite/g++.dg/modules/pr106761.h create mode 100644 gcc/testsuite/g++.dg/modules/pr106761_a.H create mode 100644 gcc/testsuite/g++.dg/modules/pr106761_b.C diff --git a/gcc/cp/module.cc b/gcc/cp/module.cc index 1a1ff5be574..9a9ef4e3332 100644 --- a/gcc/cp/module.cc +++ b/gcc/cp/module.cc @@ -8922,6 +8922,7 @@ trees_out::type_node (tree type) if (streaming_p ()) u (PACK_EXPANSION_LOCAL_P (type)); tree_node (PACK_EXPANSION_PARAMETER_PACKS (type)); + tree_node (PACK_EXPANSION_EXTRA_ARGS (type)); break; case TYPENAME_TYPE: @@ -9455,12 +9456,14 @@ trees_in::tree_node (bool is_use) { bool local = u (); tree param_packs = tree_node (); + tree extra_args = tree_node (); if (!get_overrun ()) { tree expn = cxx_make_type (TYPE_PACK_EXPANSION); SET_TYPE_STRUCTURAL_EQUALITY (expn); PACK_EXPANSION_PATTERN (expn) = res; PACK_EXPANSION_PARAMETER_PACKS (expn) = param_packs; + PACK_EXPANSION_EXTRA_ARGS (expn) = extra_args; PACK_EXPANSION_LOCAL_P (expn) = local; res = expn; } diff --git a/gcc/testsuite/g++.dg/modules/pr106761.h b/gcc/testsuite/g++.dg/modules/pr106761.h new file mode 100644 index 000..9f22a22a45d --- /dev/null +++ b/gcc/testsuite/g++.dg/modules/pr106761.h @@ -0,0 +1,22 @@ +// PR c++/106761 + +template +struct __and_; + +template +struct is_convertible; + +template +struct _TupleConstraints { + template + using __constructible = __and_...>; +}; + +template +struct tuple { + template + using __constructible += typename _TupleConstraints::template __constructible; +}; + +tuple t; diff --git a/gcc/testsuite/g++.dg/modules/pr106761_a.H b/gcc/testsuite/g++.dg/modules/pr106761_a.H new file mode 100644 index 000..8ad116412af --- /dev/null +++ b/gcc/testsuite/g++.dg/modules/pr106761_a.H @@ -0,0 +1,5 @@ +// PR c++/106761 +// { dg-additional-options -fmodule-header } + +// { dg-module-cmi {} } +#include "pr106761.h" diff --git a/gcc/testsuite/g++.dg/modules/pr106761_b.C b/gcc/testsuite/g++.dg/modules/pr106761_b.C new file mode 100644 index 000..336cb12757e --- /dev/null +++ b/gcc/testsuite/g++.dg/modules/pr106761_b.C @@ -0,0 +1,7 @@ +// PR c++/106761 +// { dg-additional-options -fmodules-ts } + +#include "pr106761.h" +import "pr106761_a.H"; + +tuple u = t; -- 2.38.0.rc0.52.gdda7228a83
Re: [PATCH] c++: stream PACK_EXPANSION_EXTRA_ARGS [PR106761]
On 9/19/22 09:52, Patrick Palka wrote: It looks like some xtreme-header-* tests are failing after the libstdc++ change r13-2158-g02f6b405f0e9dc ultimately because we're neglecting to stream PACK_EXPANSION_EXTRA_ARGS, which leads to false equivalences of different partial instantiations of _TupleConstraints::__constructible. Tested on x86_64-pc-linux-gnu, does this look OK for trunk? PR c++/106761 gcc/cp/ChangeLog: * module.cc (trees_out::type_node) : Stream PACK_EXPANSION_EXTRA_ARGS. (trees_in::tree_node) : Likewise. Looks good, I wonder why I missed that. (I guess extracting a testcase out of the headers was too tricky?) nathan --- gcc/cp/module.cc | 3 +++ 1 file changed, 3 insertions(+) diff --git a/gcc/cp/module.cc b/gcc/cp/module.cc index 1a1ff5be574..9a9ef4e3332 100644 --- a/gcc/cp/module.cc +++ b/gcc/cp/module.cc @@ -8922,6 +8922,7 @@ trees_out::type_node (tree type) if (streaming_p ()) u (PACK_EXPANSION_LOCAL_P (type)); tree_node (PACK_EXPANSION_PARAMETER_PACKS (type)); + tree_node (PACK_EXPANSION_EXTRA_ARGS (type)); break; case TYPENAME_TYPE: @@ -9455,12 +9456,14 @@ trees_in::tree_node (bool is_use) { bool local = u (); tree param_packs = tree_node (); + tree extra_args = tree_node (); if (!get_overrun ()) { tree expn = cxx_make_type (TYPE_PACK_EXPANSION); SET_TYPE_STRUCTURAL_EQUALITY (expn); PACK_EXPANSION_PATTERN (expn) = res; PACK_EXPANSION_PARAMETER_PACKS (expn) = param_packs; + PACK_EXPANSION_EXTRA_ARGS (expn) = extra_args; PACK_EXPANSION_LOCAL_P (expn) = local; res = expn; } -- Nathan Sidwell
[PATCH] c++: stream PACK_EXPANSION_EXTRA_ARGS [PR106761]
It looks like some xtreme-header-* tests are failing after the libstdc++ change r13-2158-g02f6b405f0e9dc ultimately because we're neglecting to stream PACK_EXPANSION_EXTRA_ARGS, which leads to false equivalences of different partial instantiations of _TupleConstraints::__constructible. Tested on x86_64-pc-linux-gnu, does this look OK for trunk? PR c++/106761 gcc/cp/ChangeLog: * module.cc (trees_out::type_node) : Stream PACK_EXPANSION_EXTRA_ARGS. (trees_in::tree_node) : Likewise. --- gcc/cp/module.cc | 3 +++ 1 file changed, 3 insertions(+) diff --git a/gcc/cp/module.cc b/gcc/cp/module.cc index 1a1ff5be574..9a9ef4e3332 100644 --- a/gcc/cp/module.cc +++ b/gcc/cp/module.cc @@ -8922,6 +8922,7 @@ trees_out::type_node (tree type) if (streaming_p ()) u (PACK_EXPANSION_LOCAL_P (type)); tree_node (PACK_EXPANSION_PARAMETER_PACKS (type)); + tree_node (PACK_EXPANSION_EXTRA_ARGS (type)); break; case TYPENAME_TYPE: @@ -9455,12 +9456,14 @@ trees_in::tree_node (bool is_use) { bool local = u (); tree param_packs = tree_node (); + tree extra_args = tree_node (); if (!get_overrun ()) { tree expn = cxx_make_type (TYPE_PACK_EXPANSION); SET_TYPE_STRUCTURAL_EQUALITY (expn); PACK_EXPANSION_PATTERN (expn) = res; PACK_EXPANSION_PARAMETER_PACKS (expn) = param_packs; + PACK_EXPANSION_EXTRA_ARGS (expn) = extra_args; PACK_EXPANSION_LOCAL_P (expn) = local; res = expn; } -- 2.38.0.rc0