[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#954681: [gnat-9] crash caused by SOURCE_DATE_EPOCH patch



On Sat, 4 Apr 2020 13:37:23 +0200 Nicolas Boulenguez wrote:
 > Commit 2e5e37249f09e54bb62e3ab87508e43ff709a4fb should improve the
 > situation.
 > It reverts the part affecting s-os_lib.adb. A GCC build produces the
 > same checksum for it than 9.3.0-8 in libgnat-9, hopefully making gnat
 > usable again by existing packaged libraries (I don't know how to test
 > this on a porterbox).
 > The remaining part, slightly adapted, passes the regression tests for
 > this specific bug.
 >
 > If all goes well, please lower the severity of this bug, but keep it
 > open as a reminder.
 > * Unless an objection arises, the gcc-10 branch should eventually
 > cherry-pick c1d276b0e2cb60140389aa4c23ac3addb7a77af3.
 > * Ludovic has noticed that the initial error message
 > CONSTRAINT_ERROR : uintp.adb:1959 overflow check failed
 > probably reveals a bug that should be investigated by itself.

 > No time computation is supposed to trigger a range overflow in an
 > unrelated package.

I believe I'm running into this bug building PHCpack [1] in sid. The
build fails with:

     gnatbind phctop use_c2phc pieri_solver
     raised CONSTRAINT_ERROR : uintp.adb:1959 overflow check failed

An equivalent error happens when I use gnat-10:

     gnatbind phctop use_c2phc pieri_solver
     raised CONSTRAINT_ERROR : uintp.adb:2002 overflow check failed

Is there a known workaround?

Thanks!
Doug

[1] https://github.com/janverschelde/PHCpack


Reply to: