-
Notifications
You must be signed in to change notification settings - Fork 1.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fail Check, and if ignore it and install, segfault when slice any object #4141
Comments
same with 2.3.0.alpha0.134 also test 5 fail
|
any notice of this? greetings |
bb62f36#diff-498aa9d7dcf8df8995c4ed3f9b9ba5a1L364 bb62f36#diff-d54379a7db2c0147be63c78752889bd3R48 why is the |
Works with clang (10.0.0) but explode with GCC (10.1.0) also this STL explode with
(sorry, not works with LC_ALL=C) dial_indicator_holder-mod.zip greertings |
Same here on Arch Linux, test 5 fails:
System CFLAGS (seem to be ignored): CFLAGS=-march=haswell -mtune=skylake -O3 --param=ssp-buffer-size=32 -fdevirtualize-speculatively -fno-signed-zeros -fno-trapping-math -fassociative-math -pipe -fno-plt -fdiagnostics-color=always -fstack-protector-strong -D_FORTIFY_SOURCE=2
CXXFLAGS=-march=haswell -mtune=skylake -O3 --param=ssp-buffer-size=32 -fdevirtualize-speculatively -fno-signed-zeros -fno-trapping-math -fassociative-math -pipe -fno-plt -fdiagnostics-color=always -fstack-protector-strong -D_FORTIFY_SOURCE=2
CPPFLAGS=-fstack-protector-strong -D_FORTIFY_SOURCE=2
LDFLAGS=-fstack-protector-strong -D_FORTIFY_SOURCE=2 -march=haswell -mtune=skylake -O3 --param=ssp-buffer-size=32 -fdevirtualize-speculatively -fno-signed-zeros -fno-trapping-math -fassociative-math -Wl,--sort-common,--as-needed,-z,relro -Wl,--check-sections -Wl,--gc-sections
CMake output for the build:
The build warnings generated: |
Update: For me the above mentioned work-around of switching to Clang doesn't work, the result is exactly the same. CMake output:
|
little hack for DBus check diff --git a/cmake/modules/FindDBus.cmake b/cmake/modules/FindDBus.cmake
index 1d0f29dd7..d54d4e516 100644
--- a/cmake/modules/FindDBus.cmake
+++ b/cmake/modules/FindDBus.cmake
@@ -56,4 +56,4 @@ FIND_PATH(DBUS_ARCH_INCLUDE_DIR
SET(DBUS_INCLUDE_DIRS ${DBUS_INCLUDE_DIR} ${DBUS_ARCH_INCLUDE_DIR})
INCLUDE(FindPackageHandleStandardArgs)
-FIND_PACKAGE_HANDLE_STANDARD_ARGS(DBUS REQUIRED_VARS DBUS_INCLUDE_DIRS DBUS_LIBRARIES)
\ No newline at end of file
+FIND_PACKAGE_HANDLE_STANDARD_ARGS(DBus REQUIRED_VARS DBUS_INCLUDE_DIRS DBUS_LIBRARIES)
\ No newline at end of file
builds and check without problem pkgbuild
|
I just did another build, replacing all CXXFLAGS with just "-O2" and it passes the tests now. Without above dbus patch btw. @sl1pkn07 Could you post your CXXFLAGS (and similar)? |
|
And that works now? Or did it fail the test with those? In case it failed it seems the overlap is in extended instruction use (native / haswell) and disabling the PLT. |
works. but only if is builded with clang |
Can you check with gcc? Unsetting LDFLAGS and CPPFLAGS, setting CFLAGS and CXXFLAGS to "-O2"? That worked for me. |
nope. fails with gcc (10.20) for me |
Turns out, last time it built and tested ok, but segfaults on running:
|
Is this issue still valid? Can we close it? |
let me check |
failed in 3 checks
i need more test |
hi, i found the poblem of my latest report. EDITED: my fault. my error above is fixed tested again (commit 1c2d264) with
and now fail with:
greetings |
more test failed 9f4e9da
|
We only support building on Linux using our static dependencies. |
None of the test failed is about static/dynamic dependencies or how is the projet builded So closing the issue about you comment is invalid If you read across the issue, you see one time works and others not. So you make regressions in the code and not fix it Please reopen and fix the problems posted in the issue Greetings |
We do not see your issue, thus we cannot fix it. |
What part exactly you not see? Please copy@paste the code Greetings |
@sl1pkn07 We build statically against dependencies pulled by the script in deps folder. The tests don't fail, the application does not crash. You build against some unknown versions of those libraries, maybe mixing libraries built with gcc and clang, maybe with further incompatibilities in libc++/libstdc++ and who knows what else. That may lead to mysterious crashes and we don't have resources to untangle it and reason what exactly is wrong. Of course there can be a different problem. But again, if we cannot reproduce it, we cannot fix it. |
as you wish, use the instructions on https://github.com/prusa3d/PrusaSlicer/blob/master/doc/How%20to%20build%20-%20Linux%20et%20al.md commit used: 04e4aee
build script:
and now? greetings |
well. i have tested a lot of things, and i think i have found the gulty the systemwide C/CXXFLAGS set up any dev can take a look why fail the 2rd test (and some time get segfault when use the program normally, like posted in #4141 (comment)) and if is fixeable without change that flag? greetings |
Version
version_2.2.0.191.gdb49a4516
Operating system type + version
Archlinux
3D printer brand / version + firmware version (if known)
not aplicable
Behavior
make check fail (ctest -V)
Check OK
segfault
if ignore the test, and install it. when slice any objet (any, no matter what)
Prusa slicer gets Segfault
wxwidgets 3.0.4 over GTK3 3.24.18
cmake 3.17.1
gcc 9.3.0
version_2.2.0.0.gd5bcddeed works ok. seems regression
The text was updated successfully, but these errors were encountered: