Details
-
Bug
-
Resolution: Done
-
Not Evaluated
-
Qt Creator 8.0.0
-
None
-
Windows host, Qt Creator 8.0 (~beta1), NDK 23.1.7779620 with LLDB 12.0.8, x86_64 API29 Android10 image in Emulator, Android Qt 6.3.1 from Qt installer
-
Android,
Windows
-
83c380665d (qt-creator/qt-creator/8.0) 83c380665d (qt-creator/qt-creator/master) 83c380665d (qt-creator/tqtc-qt-creator/8.0)
Description
With the specified environment and with a cleanly built C++ test application, I get differences in the debugger logs between the first and second debugger starts.
- Debugger log of first start: DebuggerLog_FirstStart.txt
- Debugger log of second start: DebuggerLog_SecondStart.txt
Quite early in the log, in the ELF SECTIONS message differ:
dELF SECTIONS: .interp .note.android.ident .note.gnu.build-id .dynsym .gnu.version .gnu.version_r .gnu.hash .dynstr .rel.dyn .rel.plt .rodata .eh_frame_hdr .eh_frame .text .plt .preinit_array .init_array .fini_array .data.rel.ro .dynamic .got .got.plt .bss .shstrtab .gnu_debugdata
dELF SECTIONS: .interp .note.android.ident .note.gnu.build-id .dynsym .gnu.version .gnu.version_r .gnu.hash .dynstr .rela.dyn .rela.plt .rodata .eh_frame_hdr .eh_frame .text .plt .preinit_array .init_array .fini_array .data.rel.ro .dynamic .got .got.plt .bss .shstrtab .gnu_debugdata
The first start talks about .rel.dyn which is 32-bit and the second start talks about .rela.dyn (64-bit).
But in both cases, the ABI is detected correctly:
dABI: x86-linux-android-elf-64bit
In fact, on the first start, the build directory is populated with these files: app_process_FirstStart, linker_FirstStart, libc_FirstStart.so.
Before the second start it is different files: app_process_SecondStart, linker64_SecondStart, libc_SecondStart.so.
I suspect that this can cause follow-up misbehavior.
Attachments
Issue Links
- relates to
-
-
- Closed
-