Stuart Henderson wrote: > On 2026/04/30 21:02, Chris Billington wrote: >> >> I attempted to run with gdb. tdesktop seems to already contain debugging >> symbols, but gdb exits before I am able to 'run' it: > > you should have a core file I think, so you can try > > egdb Telegram Telegram.core > >> Maybe the coming release of 7.9 with newer versions and debugging symbols >> stripped will resolve it? > > What's currently available in snapshots (including packages for aarch64 > amd64 i386 sparc64) are either extremely close to, or possibly the same > as, 7.9 release. > > (tdesktop works for me in -current). > There's no .core file for Telegram, oddly. But there is one for egdb: Reading symbols from egdb... (No debugging symbols found in egdb) [New process 205927] [New process 152837] [New process 286037] [New process 607633] [New process 547542] Core was generated by `egdb'. Program terminated with signal SIGABRT, Aborted. #0 thrkill () at /tmp/-:2 warning: Source file is more recent than executable. [Current thread is 1 (process 205927)] (gdb) bt #0 thrkill () at /tmp/-:2 #1 0xd6299ff2435ba494 in ?? () #2 0x0000037c98109e76 in handle_fatal_signal(int) () #3 <signal handler called> #4 thrkill () at /tmp/-:2 #5 0xd6299ff1c2f5b3a9 in ?? () #6 0x0000037f19be894b in _libc_abort () at /usr/src/lib/libc/stdlib/abort.c:51 #7 0x0000037c984f7b1e in abort_with_message(char const*) () #8 0x0000037c984f2d0d in internal_vproblem(internal_problem*, char const*, int, char const*, __va_list_tag*) () #9 0x0000037c984f27ff in internal_verror(char const*, int, char const*, __va_list_tag*) () #10 0x0000037c98a9e3b8 in internal_error_loc(char const*, int, char const*, ...) () #11 0x0000037c984f303a in malloc_failure(long) () #12 0x0000037c97f0a728 in xmalloc () #13 0x0000037c98aa0bf1 in format_pieces::format_pieces(char const**, bool, bool) () #14 0x0000037c984eda3f in ui_out::vmessage(ui_file_style const&, char const*, __va_list_tag*) () #15 0x0000037c984eb086 in ui_file::vprintf(char const*, __va_list_tag*) () #16 0x0000037c984f33fe in defaulted_query(char const*, char, __va_list_tag*) () #17 0x0000037c984f37b5 in query(char const*, ...) () #18 0x0000037c984f2a4c in internal_vproblem(internal_problem*, char const*, int, char const*, __va_list_tag*) () #19 0x0000037c984f27ff in internal_verror(char const*, int, char const*, __va_list_tag*) () #20 0x0000037c98a9e3b8 in internal_error_loc(char const*, int, char const*, ...) () #21 0x0000037c984f303a in malloc_failure(long) () #22 0x0000037c98aa31a8 in operator new(unsigned long) () #23 0x0000037c97ed6ed0 in ada_decode(char const*, bool, bool, bool) () #24 0x0000037c97ef821e in ada_language::sniff_from_mangled_name(char const*, std::__1::unique_ptr<char, gdb::xfree_deleter<char> >*) const () #25 0x0000037c98448403 in symbol_find_demangled_name(general_symbol_info*, char const*) () #26 0x0000037c9825061a in minimal_symbol_reader::install()::$_0::operator()(minimal_symbol*, minimal_symbol*) const () #27 0x0000037c9824fad6 in minimal_symbol_reader::install() () #28 0x0000037c980fd710 in elf_symfile_read(objfile*, enum_flags<symfile_add_flag>) () #29 0x0000037c9843d606 in read_symbols(objfile*, enum_flags<symfile_add_flag>) () #30 0x0000037c9843a8b0 in symbol_file_add_with_addrs(gdb::ref_ptr<bfd, gdb_bfd_ref_policy> const&, char const*, enum_flags<symfile_add_flag>, std::__1::vector<other_sections, std::__1::allocator<other_sections> >*, enum_flags<objfile_flag>, objfile*) () #31 0x0000037c9843ae97 in symbol_file_add_main_1(char const*, enum_flags<symfile_add_flag>, enum_flags<objfile_flag>, unsigned long long) () #32 0x0000037c98215c2e in captured_main_1(captured_main_args*) () #33 0x0000037c982143cf in gdb_main(captured_main_args*) () #34 0x0000037c97eae740 in main () I retried egdb Telegram a few times, and once out of 10 perhaps I got: Reading symbols from Telegram... /usr/obj/ports/gdb-16.3/gdb-16.3/gdb/utils.c:667: internal-error: virtual memory exhausted: can't allocate 576 bytes. A problem internal to GDB has been detected, further debugging may prove unreliable. Quit this debugging session? (y or n) Recursive internal problem. Abort trap (core dumped) The memory usage for 'egdb Telegram' climbs to about 4.5GB before falling back as egdb aborts. tdesktop-6.7.6 does run OK on for me on a not-very recent -current (GENERIC.MP#304). But 'egdb Telegram' aborts before all the symbols are loaded like on 7.8. Telegram executable contains too many symbols for egdb to handle with default login.conf settings? Chris
No comments:
Post a Comment