From: Nathaniel Shead <nathanieloshead@gmail.com>
To: gcc-patches@gcc.gnu.org
Cc: Jason Merrill <jason@redhat.com>, Nathan Sidwell <nathan@acm.org>
Subject: [PATCH] c++/modules: Remember that header units have CMIs
Date: Mon, 13 May 2024 12:58:38 +1000 [thread overview]
Message-ID: <664181e2.650a0220.dbfd8.e3d5@mx.google.com> (raw)
Bootstrapped and regtested on x86_64-pc-linux-gnu, OK for trunk?
-- >8 --
This appears to be an oversight in the definition of module_has_cmi_p;
this comes up transitively in other functions used for e.g. determining
whether a name could potentially be accessed in a different translation
unit.
This change will allow us to use the function directly in more places
that need to additional work only if generating a module CMI in the
future.
gcc/cp/ChangeLog:
* cp-tree.h (module_has_cmi_p): Also true for header units.
gcc/testsuite/ChangeLog:
* g++.dg/modules/linkage-3_a.H: New test.
* g++.dg/modules/linkage-3_b.C: New test.
* g++.dg/modules/linkage-3_c.C: New test.
Signed-off-by: Nathaniel Shead <nathanieloshead@gmail.com>
---
gcc/cp/cp-tree.h | 2 +-
gcc/testsuite/g++.dg/modules/linkage-3_a.H | 19 +++++++++++++++++++
gcc/testsuite/g++.dg/modules/linkage-3_b.C | 9 +++++++++
gcc/testsuite/g++.dg/modules/linkage-3_c.C | 10 ++++++++++
4 files changed, 39 insertions(+), 1 deletion(-)
create mode 100644 gcc/testsuite/g++.dg/modules/linkage-3_a.H
create mode 100644 gcc/testsuite/g++.dg/modules/linkage-3_b.C
create mode 100644 gcc/testsuite/g++.dg/modules/linkage-3_c.C
diff --git a/gcc/cp/cp-tree.h b/gcc/cp/cp-tree.h
index db098c32f2d..609904918ba 100644
--- a/gcc/cp/cp-tree.h
+++ b/gcc/cp/cp-tree.h
@@ -7379,7 +7379,7 @@ inline bool module_interface_p ()
inline bool module_partition_p ()
{ return module_kind & MK_PARTITION; }
inline bool module_has_cmi_p ()
-{ return module_kind & (MK_INTERFACE | MK_PARTITION); }
+{ return module_kind & (MK_INTERFACE | MK_PARTITION | MK_HEADER); }
inline bool module_purview_p ()
{ return module_kind & MK_PURVIEW; }
diff --git a/gcc/testsuite/g++.dg/modules/linkage-3_a.H b/gcc/testsuite/g++.dg/modules/linkage-3_a.H
new file mode 100644
index 00000000000..1e0ebd927e2
--- /dev/null
+++ b/gcc/testsuite/g++.dg/modules/linkage-3_a.H
@@ -0,0 +1,19 @@
+// { dg-additional-options "-fmodule-header -Wno-error=c++20-extensions" }
+// { dg-module-cmi {} }
+
+// Like linkage-1, but for header units.
+
+// External linkage definitions must be declared as 'inline' to satisfy
+// [module.import] p6, so we don't need to care about voldemort types in
+// function definitions.
+
+// Strictly speaking this is not required to be supported:
+// [module.import] p5 says that when two different TUs import header-names
+// identifying the same header or source file, it is unspecified whether
+// they import the same header unit, and thus 's' could be a different entity
+// in each TU. But with out current implementation this seems to reasonable to
+// allow (and it does currently work).
+
+struct { int y; } s;
+decltype(s) f(); // { dg-warning "used but not defined" "" { target c++17_down } }
+inline auto x = f();
diff --git a/gcc/testsuite/g++.dg/modules/linkage-3_b.C b/gcc/testsuite/g++.dg/modules/linkage-3_b.C
new file mode 100644
index 00000000000..935ef6150ec
--- /dev/null
+++ b/gcc/testsuite/g++.dg/modules/linkage-3_b.C
@@ -0,0 +1,9 @@
+// { dg-additional-options "-fmodules-ts" }
+
+struct {} unrelated;
+
+import "linkage-3_a.H";
+
+decltype(s) f() {
+ return { 123 };
+}
diff --git a/gcc/testsuite/g++.dg/modules/linkage-3_c.C b/gcc/testsuite/g++.dg/modules/linkage-3_c.C
new file mode 100644
index 00000000000..be527ff552d
--- /dev/null
+++ b/gcc/testsuite/g++.dg/modules/linkage-3_c.C
@@ -0,0 +1,10 @@
+// { dg-module-do run }
+// { dg-additional-options "-fmodules-ts" }
+
+import "linkage-3_a.H";
+
+int main() {
+ auto a = x.y;
+ if (a != 123)
+ __builtin_abort();
+}
--
2.43.2
next reply other threads:[~2024-05-13 2:58 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-13 2:58 Nathaniel Shead [this message]
2024-05-14 22:21 ` Jason Merrill
2024-05-17 6:14 ` [PATCH v2] " Nathaniel Shead
2024-05-17 12:21 ` Nathaniel Shead
2024-05-20 22:00 ` Jason Merrill
2024-05-23 13:27 ` [PATCH 1/2] c++/modules: Fix treatment of unnamed types Nathaniel Shead
2024-05-23 13:29 ` [PATCH 2/2] c++/modules: Remember that header units have CMIs Nathaniel Shead
2024-05-23 19:37 ` Jason Merrill
2024-05-23 19:36 ` [PATCH 1/2] c++/modules: Fix treatment of unnamed types Jason Merrill
2024-05-24 1:27 ` Nathaniel Shead
2024-05-24 13:57 ` Jason Merrill
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=664181e2.650a0220.dbfd8.e3d5@mx.google.com \
--to=nathanieloshead@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jason@redhat.com \
--cc=nathan@acm.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).