[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 Jonathan Wakely changed: What|Removed |Added Status|ASSIGNED|RESOLVED Known to work||6.4.0, 7.3.0, 8.2.0 Resolution|--- |FIXED Known to fail||6.4.1, 7.3.1, 8.2.1, 9.0 --- Comment #14 from Jonathan Wakely --- .
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822
--- Comment #13 from Jonathan Wakely ---
Created attachment 44936
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=44936=edit
Patch for GCC 6.5
Fixed in the gcc-7 and gcc-8 branches.
This will not be fixed on gcc-6-branch, but for anybody who wants to apply it
locally, the patch for GCC 6.5 is attached, and reproduced here:
--- a/libstdc++-v3/include/bits/stl_pair.h
+++ b/libstdc++-v3/include/bits/stl_pair.h
@@ -187,7 +187,7 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
};
#endif // C++11
- class __pair_base
+ template class __pair_base
{
#if __cplusplus >= 201103L
template friend struct pair;
@@ -206,7 +206,7 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
*/
template
struct pair
-: private __pair_base
+: private __pair_base<_T1, _T2>
{
typedef _T1 first_type;/// @c first_type is the first bound type
typedef _T2 second_type; /// @c second_type is the second bound type
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #12 from Jonathan Wakely --- Author: redi Date: Wed Oct 31 13:16:10 2018 New Revision: 265683 URL: https://gcc.gnu.org/viewcvs?rev=265683=gcc=rev Log: PR libstdc++/87822 fix layout change for nested std::pair The introduction of the empty __pair_base base class for PR 86751 changed the layout of std::pair, ...>. The outer pair and its first member both have a base class of the same type, which cannot exist at the same address. This causes the first member to be at a non-zero offset. The solution is to make the base class depend on the template parameters, so that each pair type has a different base class type, which allows the base classes of the outer pair and its first member to have the same address. PR libstdc++/87822 * include/bits/stl_pair.h (__pair_base): Change to class template. (pair): Make base class type depend on template parameters. * testsuite/20_util/pair/87822.cc: New test. Added: branches/gcc-7-branch/libstdc++-v3/testsuite/20_util/pair/87822.cc Modified: branches/gcc-7-branch/libstdc++-v3/ChangeLog branches/gcc-7-branch/libstdc++-v3/include/bits/stl_pair.h
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #11 from Jonathan Wakely --- Author: redi Date: Wed Oct 31 13:03:25 2018 New Revision: 265681 URL: https://gcc.gnu.org/viewcvs?rev=265681=gcc=rev Log: PR libstdc++/87822 fix layout change for nested std::pair The introduction of the empty __pair_base base class for PR 86751 changed the layout of std::pair, ...>. The outer pair and its first member both have a base class of the same type, which cannot exist at the same address. This causes the first member to be at a non-zero offset. The solution is to make the base class depend on the template parameters, so that each pair type has a different base class type, which allows the base classes of the outer pair and its first member to have the same address. PR libstdc++/87822 * include/bits/stl_pair.h (__pair_base): Change to class template. (pair): Make base class type depend on template parameters. * testsuite/20_util/pair/87822.cc: New test. Added: branches/gcc-8-branch/libstdc++-v3/testsuite/20_util/pair/87822.cc Modified: branches/gcc-8-branch/libstdc++-v3/ChangeLog branches/gcc-8-branch/libstdc++-v3/include/bits/stl_pair.h
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #10 from Romain Geissler --- Thanks for the quick patch ! If no commit is planned in the branch 6, I am going to apply this patch on top myself. I hope people do read the release notes to figure out about this potential ABI breaking.
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #9 from Jonathan Wakely --- Author: redi Date: Wed Oct 31 12:29:02 2018 New Revision: 265678 URL: https://gcc.gnu.org/viewcvs?rev=265678=gcc=rev Log: PR libstdc++/87822 fix layout change for nested std::pair The introduction of the empty __pair_base base class for PR 86751 changed the layout of std::pair, ...>. The outer pair and its first member both have a base class of the same type, which cannot exist at the same address. This causes the first member to be at a non-zero offset. The solution is to make the base class depend on the template parameters, so that each pair type has a different base class type, which allows the base classes of the outer pair and its first member to have the same address. PR libstdc++/87822 * include/bits/stl_pair.h (__pair_base): Change to class template. (pair): Make base class type depend on template parameters. * testsuite/20_util/pair/87822.cc: New test. Added: trunk/libstdc++-v3/testsuite/20_util/pair/87822.cc Modified: trunk/libstdc++-v3/ChangeLog trunk/libstdc++-v3/include/bits/stl_pair.h
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #8 from rguenther at suse dot de --- On Wed, 31 Oct 2018, redi at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 > > --- Comment #7 from Jonathan Wakely --- > (In reply to Richard Biener from comment #5) > > Unfortunate :/ Can you add a 6.5 specific note to 6.5/changes.html? > > Will do. > > I think it would also be good to commit the fix to the gcc-6-branch, even if > it's closed, so it can be picked up from there if needed. Hmm. I think it may be better to provide the fix as patch, referenced from the changes.html note? > Do we even want to consider a 6.6 release, or just officially bless a 6.5.1 > snapshot post-fix? Neither of that (there are no further snapshots from the branch anyways). Since 6.5 isn't supported anymore I'd rather point people to 7.x.
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #7 from Jonathan Wakely --- (In reply to Richard Biener from comment #5) > Unfortunate :/ Can you add a 6.5 specific note to 6.5/changes.html? Will do. I think it would also be good to commit the fix to the gcc-6-branch, even if it's closed, so it can be picked up from there if needed. Do we even want to consider a 6.6 release, or just officially bless a 6.5.1 snapshot post-fix?
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 Jonathan Wakely changed: What|Removed |Added Status|NEW |ASSIGNED Assignee|unassigned at gcc dot gnu.org |redi at gcc dot gnu.org --- Comment #6 from Jonathan Wakely --- It only affects std::pair, z> (for any x, y, z) i.e. when the first member of the pair is also a pair.
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 Richard Biener changed: What|Removed |Added CC||jakub at gcc dot gnu.org --- Comment #5 from Richard Biener --- (In reply to Jonathan Wakely from comment #4) > The fix isn't in any other releases yet, only 6.5 Unfortunate :/ Can you add a 6.5 specific note to 6.5/changes.html? How pervasive is this issue?
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #4 from Jonathan Wakely --- The fix isn't in any other releases yet, only 6.5
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 --- Comment #3 from Jonathan Wakely --- (In reply to Romain Geissler from comment #0) > (introduce a new > tagged std::pair type and provide dual abi ?). No, no, no! Anything but that.
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 Jonathan Wakely changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2018-10-31 Ever confirmed|0 |1 --- Comment #2 from Jonathan Wakely --- Argh, dammit. Yes, it should have been a template.
[Bug libstdc++/87822] [6/7/8/9 Regression] Binary incompatibility in std::pair introduced by PR 86751
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87822 Richard Biener changed: What|Removed |Added Keywords||ABI CC||rguenth at gcc dot gnu.org Blocks||86751 Target Milestone|--- |7.4 Summary|[regression 6/7/8/9] Binary |[6/7/8/9 Regression] Binary |incompatibility in |incompatibility in |std::pair introduced by PR |std::pair introduced by PR |86751 |86751 Referenced Bugs: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86751 [Bug 86751] [6/7/8 Regression] Ambiguous operator= overload for std::pair
