In latest CI runs, see: https://github.com/boostorg/math/actions/runs/9427754774/job/25973291514?pr=... Any ideas? Thanks! John.
On 6/8/24 14:01, John Maddock via Boost wrote:
In latest CI runs, see: https://github.com/boostorg/math/actions/runs/9427754774/job/25973291514?pr=...
Any ideas?
The error code 0xC0000005 suggests it's an access violation. Perhaps, some recent changes to the Boost tree started causing b2 crash? I don't see any recent commits in tools/build develop.
Andrey Semashev wrote:
On 6/8/24 14:01, John Maddock via Boost wrote:
In latest CI runs, see:
https://github.com/boostorg/math/actions/runs/9427754774/job/25973291 5
14?pr=1144
Any ideas?
The error code 0xC0000005 suggests it's an access violation. Perhaps, some recent changes to the Boost tree started causing b2 crash? I don't see any recent commits in tools/build develop.
It's an access violation, and it only happens with the new VS 2022 17.10, and only on GHA. I'm not able to reproduce the crash locally.
Any ideas?
The error code 0xC0000005 suggests it's an access violation. Perhaps, some recent changes to the Boost tree started causing b2 crash? I don't see any recent commits in tools/build develop.
It's an access violation, and it only happens with the new VS 2022 17.10, and only on GHA. My current understanding of the issue isfoggy, but I believe it's something like, ... I think GHA has a new version/patch of VC143.b2 searches for the version/patch via direct(hard coded) lookup and it can't be foundon GHA. On Sunday, June 9, 2024 at 01:25:52 AM GMT+2, Peter Dimov via Boost
wrote:
Andrey Semashev wrote:
On 6/8/24 14:01, John Maddock via Boost wrote:
In latest CI runs, see:
https://github.com/boostorg/math/actions/runs/9427754774/job/25973291 5
14?pr=1144
Any ideas?
The error code 0xC0000005 suggests it's an access violation. Perhaps, some recent changes to the Boost tree started causing b2 crash? I don't see any recent commits in tools/build develop.
It's an access violation, and it only happens with the new VS 2022 17.10, and only on GHA. I'm not able to reproduce the crash locally. _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
сб, 8 июн. 2024 г. в 23:19, Andrey Semashev via Boost
The error code 0xC0000005 suggests it's an access violation. Perhaps, some recent changes to the Boost tree started causing b2 crash? I don't see any recent commits in tools/build develop.
This is a breakage in the GHA image. This issue lists several solutions to the problem that people have discovered : https://github.com/actions/runner-images/issues/10004. I suggest a long-term solution: to NEVER use GHA without a container, as the maintainers keep breaking it.
> This is a breakage in the GHA image. Yes we see that now. If it helps, I ended up solving this problemin 3 non-boost-related repo projectson GHA. * I explicitly boot-strapped and built b2with vc142. OK, you sidestep the brokenDLL. * Then I simply temporarily walkedthe vc143 jobs back to vc142, hopingthat an image fix is forthcoming. Also along the way, since vc144 isout there, someone sooner or laterneeds to change bootstrap.bat.Otherwise we will get zapped inthe butt by that too down the road. Christopher On Sunday, June 9, 2024 at 09:45:36 AM GMT+2, Дмитрий Архипов via Boostwrote: сб, 8 июн. 2024 г. в 23:19, Andrey Semashev via Boost : > The error code 0xC0000005 suggests it's an access violation. Perhaps, > some recent changes to the Boost tree started causing b2 crash? I don't > see any recent commits in tools/build develop. This is a breakage in the GHA image. This issue lists several solutions to the problem that people have discovered : https://github.com/actions/runner-images/issues/10004. I suggest a long-term solution: to NEVER use GHA without a container, as the maintainers keep breaking it. _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Christopher Kormanyos wrote:
Also along the way, since vc144 isout there, someone sooner or laterneeds to change bootstrap.bat.Otherwise we will get zapped inthe butt by that too down the road.
There is no vc144. Toolset version 14.40 is still vc143; it's just the next version after 14.39. They ran out of 14.3x numbers.
There is no vc144. Toolset version 14.40> is still vc143 Understood. That had confused me. Thank you for clarifying this, Peter. Christopher On Sunday, June 9, 2024 at 06:29:20 PM GMT+2, Peter Dimov via Boost
wrote:
Christopher Kormanyos wrote:
Also along the way, since vc144 isout there, someone sooner or laterneeds to change bootstrap.bat.Otherwise we will get zapped inthe butt by that too down the road.
There is no vc144. Toolset version 14.40 is still vc143; it's just the next version after 14.39. They ran out of 14.3x numbers. _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
On Sun, Jun 9, 2024 at 2:45 AM Дмитрий Архипов via Boost
сб, 8 июн. 2024 г. в 23:19, Andrey Semashev via Boost
: The error code 0xC0000005 suggests it's an access violation. Perhaps, some recent changes to the Boost tree started causing b2 crash? I don't see any recent commits in tools/build develop.
This is a breakage in the GHA image. This issue lists several solutions to the problem that people have discovered : https://github.com/actions/runner-images/issues/10004.
I suggest a long-term solution: to NEVER use GHA without a container, as the maintainers keep breaking it.
Would you happen to know how to do that without violating the Windows software licensing terms? -- -- René Ferdinand Rivera Morell -- Don't Assume Anything -- No Supone Nada -- Robot Dreams - http://robot-dreams.net
participants (6)
-
Andrey Semashev
-
Christopher Kormanyos
-
John Maddock
-
Peter Dimov
-
René Ferdinand Rivera Morell
-
Дмитрий Архипов