[Bug c++/77522] [6/7/8 Regression] ICE on invalid code C++14 code: in tsubst_decl, at cp/pt.c:12447
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77522 --- Comment #5 from paolo at gcc dot gnu.org --- Author: paolo Date: Fri Feb 9 11:37:14 2018 New Revision: 257521 URL: https://gcc.gnu.org/viewcvs?rev=257521&root=gcc&view=rev Log: 2018-02-09 Paolo Carlini PR c++/77522 * g++.dg/cpp1y/lambda-init16.C: New. Added: trunk/gcc/testsuite/g++.dg/cpp1y/lambda-init16.C Modified: trunk/gcc/testsuite/ChangeLog
[Bug c++/77522] [6/7/8 Regression] ICE on invalid code C++14 code: in tsubst_decl, at cp/pt.c:12447
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77522 --- Comment #4 from Jakub Jelinek --- The ICE got fixed with r251433. Since then the error is: pr77522.C: In instantiation of ‘void f(T) [with T = int]’: pr77522.C:8:7: required from here pr77522.C:3:8: error: use of ‘f(T) [with T = int]__a’ before deduction of ‘auto’ auto g = [&a = f] () {}; ^ pr77522.C:3:12: error: invalid initialization of non-const reference of type ‘auto&’ from an rvalue of type ‘’ auto g = [&a = f] () {}; ^ Paolo, will you check the testcase in and remove 8 Regression?
[Bug c++/77522] [6/7/8 Regression] ICE on invalid code C++14 code: in tsubst_decl, at cp/pt.c:12447
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77522 Paolo Carlini changed: What|Removed |Added CC||dmalcolm at gcc dot gnu.org, ||paolo.carlini at oracle dot com --- Comment #3 from Paolo Carlini --- In trunk we don't ICE anymore on this, I think we can safely remove the 8 Regression marker and commit a testcase. I'm wondering however why when initialize_reference calls error_at the input_location is at the open square bracket. Well, just a curiosity, at some point expr itself will have its own usable location.
[Bug c++/77522] [6/7/8 Regression] ICE on invalid code C++14 code: in tsubst_decl, at cp/pt.c:12447
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77522 Richard Biener changed: What|Removed |Added Target Milestone|5.5 |6.5