Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: werner at beroux dot com
Target Milestone: ---
gcc (Debian 4.9.2-10) 4.9.2
Steps:
1. Install Docker
2. Retrieve `Dockerfile`: `$ wget
https://raw.githubusercontent.com/wernight/docker
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #2 from werner at beroux dot com ---
It worked on gcc (Debian 4.9.2-2).
Also I tried on two machines with pretty different spec. I'm not excluding the
memory issue, just saying it used to work on those two machines and now it
doesn't.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #3 from werner at beroux dot com ---
I should clarify that the code being compiled might have changed as well
between the two:
https://github.com/ariya/phantomjs/commits/master
from 2015-06-25T20:16:02.750261Z (working)
to 2015-09
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #5 from werner at beroux dot com ---
I'm trying to build with newer GCC and older code to confirm that (build takes
a while).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #6 from werner at beroux dot com ---
Humm no sorry I'm building from the same git tag (which is of Jan 24). So it
looks like something caused GCC to use more memory than before.
I did notice a GCC5 compilation fix, probably unre
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #8 from werner at beroux dot com ---
Tried to build the exact same just on newer gcc and nothing else should have
changed, and it failed.
I'll try to build on older gcc as well.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #9 from werner at beroux dot com ---
(In reply to Markus Trippelsdorf from comment #7)
> Can you please attach the preprocessed JSBindingsAllInOne.ii file (by
> adding -save-temps to the compiler invocation)?
I added that fl
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67571
--- Comment #11 from werner at beroux dot com ---
(In reply to Markus Trippelsdorf from comment #10)
> (In reply to werner from comment #9)
> > (In reply to Markus Trippelsdorf from comment #7)
> > > Can you please attac
