[Bug gdb/27176] “demangler-warning: unable to demangle” while loading debugging telegram-desktop under Debian 10 GNU/Linux

[simark at simark dot ca]

https://sourceware.org/bugzilla/show_bug.cgi?id=27176

--- Comment #3 from Simon Marchi <simark at simark dot ca> ---
(In reply to Garreau, Alexandre from comment #2)
> Oh thank you!  But I’m curious: why did *gdb* tell me it had a bug? if the
> bug is not due to gdb but wrong input, shouldn’t gdb not say that? isn’t
> that mere fact a sort of bug?

Perhaps.  gdb uses the demangler from gcc.  It's code shared between the two
repos, but it's maintained by gcc and we just import it.  The demangler is
never supposed to crash, but it happens some times.  I haven't written that
code, but I suppose that it was decided that when this happens, GDB can't
reasonably continue, so we might as well consider an error internal to the
demangler internal to GDB too.

> 
> Also, since I didn’t used directly gcc but only gdb, should I, now, rather
> report the bug to my distribution (Debian) rather than gcc? or should I
> investigate the gcc version that has been used to produce these debugging
> symbols (likely gcc (Debian 8.3.0-6) 8.3.0), and report to gcc anyway
> (pointing them to the wrongly mangled debugging symbols)?

Well, you raise a good point about the versions.  You should ideally try the
latest version of GDB (which might include a newer version of the demangler),
see if the problem still exists.

When passing your symbol in c++filt on Debian 10, I get a segfault, likely the
same root cause as your problem:

root@36f2fe0af0bb:/# c++filt 
_ZNSt11_Tuple_implILm0EJN3rpl8producerIN4base5flagsIN11MTPDchannel4FlagEEENS0_8no_errorEZNS0_7details10map_helperIZN4Data18FlagsValueWithMaskINSA_5FlagsIS6_Lj2147486693EE6ChangeES7_ZNS8_11then_helperISE_S7_ZNKS0_12event_streamISE_E6eventsEvEUlRKT_E_EclISE_S7_ZNS0_6singleISE_EEDaOSI_EUlSK_E_SE_S7_EEDaONS1_ISI_T0_T1_EEEUlSK_E_EEDaSU_NSI_4TypeEEUlRKSE_E0_EclISE_S7_ZNS8_13filter_helperIZNSB_ISE_S7_SV_EEDaSU_SW_EUlSY_E_EclISE_S7_SV_vEEDaSU_EUlSK_E_S6_EEDaSU_EUlSK_E_EENS1_IbS7_ZNS9_IZNSA_15SingleFlagValueINSC_INS3_IN19MTPDchatAdminRights4FlagEEELj4294967295EE6ChangeES7_ZNSF_IS1E_S7_ZNKSG_IS1E_E6eventsEvEUlSK_E_EclIS1E_S7_ZNSO_IS1E_EEDaSP_EUlSK_E_S1E_S7_EEDaSU_EUlSK_E_EEDaSU_NSI_4EnumEEUlS1C_E_EclIS1C_S7_ZNS9_IZNSB_IS1E_S7_S1K_EEDaSU_SW_EUlRKS1E_E0_EclIS1E_S7_ZNS12_IZNSB_IS1E_S7_S1K_EEDaSU_SW_EUlS1Q_E_EclIS1E_S7_S1K_vEEDaSU_EUlSK_E_S1C_EEDaSU_EUlSK_E_bEEDaSU_EUlSK_E_EENS1_IbS7_ZNS9_IZNS19_INSC_INS3_IN20MTPDchatBannedRights4FlagEEELj4294967295EE6ChangeES7_ZNSF_IS25_S7_ZNKSG_IS25_E6eventsEvEUlSK_E_EclIS25_S7_ZNSO_IS25_EEDaSP_EUlSK_E_S25_S7_EEDaSU_EUlSK_E_EEDaSU_S1L_EUlS23_E_EclIS23_S7_ZNS9_IZNSB_IS25_S7_S2B_EEDaSU_SW_EUlRKS25_E0_EclIS25_S7_ZNS12_IZNSB_IS25_S7_S2B_EEDaSU_SW_EUlS2G_E_EclIS25_S7_S2B_vEEDaSU_EUlSK_E_S23_EEDaSU_EUlSK_E_bEEDaSU_EUlSK_E_EES2Q_EED1Ev
Segmentation fault (core dumped)

When I run it through c++file in my binutils-gdb development tree, it doesn't
seem to demangle, but at least it doesn't segfault.

$ ../binutils/cxxfilt
_ZNSt11_Tuple_implILm0EJN3rpl8producerIN4base5flagsIN11MTPDchannel4FlagEEENS0_8no_errorEZNS0_7details10map_helperIZN4Data18FlagsValueWithMaskINSA_5FlagsIS6_Lj2147486693EE6ChangeES7_ZNS8_11then_helperISE_S7_ZNKS0_12event_streamISE_E6eventsEvEUlRKT_E_EclISE_S7_ZNS0_6singleISE_EEDaOSI_EUlSK_E_SE_S7_EEDaONS1_ISI_T0_T1_EEEUlSK_E_EEDaSU_NSI_4TypeEEUlRKSE_E0_EclISE_S7_ZNS8_13filter_helperIZNSB_ISE_S7_SV_EEDaSU_SW_EUlSY_E_EclISE_S7_SV_vEEDaSU_EUlSK_E_S6_EEDaSU_EUlSK_E_EENS1_IbS7_ZNS9_IZNSA_15SingleFlagValueINSC_INS3_IN19MTPDchatAdminRights4FlagEEELj4294967295EE6ChangeES7_ZNSF_IS1E_S7_ZNKSG_IS1E_E6eventsEvEUlSK_E_EclIS1E_S7_ZNSO_IS1E_EEDaSP_EUlSK_E_S1E_S7_EEDaSU_EUlSK_E_EEDaSU_NSI_4EnumEEUlS1C_E_EclIS1C_S7_ZNS9_IZNSB_IS1E_S7_S1K_EEDaSU_SW_EUlRKS1E_E0_EclIS1E_S7_ZNS12_IZNSB_IS1E_S7_S1K_EEDaSU_SW_EUlS1Q_E_EclIS1E_S7_S1K_vEEDaSU_EUlSK_E_S1C_EEDaSU_EUlSK_E_bEEDaSU_EUlSK_E_EENS1_IbS7_ZNS9_IZNS19_INSC_INS3_IN20MTPDchatBannedRights4FlagEEELj4294967295EE6ChangeES7_ZNSF_IS25_S7_ZNKSG_IS25_E6eventsEvEUlSK_E_EclIS25_S7_ZNSO_IS25_EEDaSP_EUlSK_E_S25_S7_EEDaSU_EUlSK_E_EEDaSU_S1L_EUlS23_E_EclIS23_S7_ZNS9_IZNSB_IS25_S7_S2B_EEDaSU_SW_EUlRKS25_E0_EclIS25_S7_ZNS12_IZNSB_IS25_S7_S2B_EEDaSU_SW_EUlS2G_E_EclIS25_S7_S2B_vEEDaSU_EUlSK_E_S23_EEDaSU_EUlSK_E_bEEDaSU_EUlSK_E_EES2Q_EED1Ev
_ZNSt11_Tuple_implILm0EJN3rpl8producerIN4base5flagsIN11MTPDchannel4FlagEEENS0_8no_errorEZNS0_7details10map_helperIZN4Data18FlagsValueWithMaskINSA_5FlagsIS6_Lj2147486693EE6ChangeES7_ZNS8_11then_helperISE_S7_ZNKS0_12event_streamISE_E6eventsEvEUlRKT_E_EclISE_S7_ZNS0_6singleISE_EEDaOSI_EUlSK_E_SE_S7_EEDaONS1_ISI_T0_T1_EEEUlSK_E_EEDaSU_NSI_4TypeEEUlRKSE_E0_EclISE_S7_ZNS8_13filter_helperIZNSB_ISE_S7_SV_EEDaSU_SW_EUlSY_E_EclISE_S7_SV_vEEDaSU_EUlSK_E_S6_EEDaSU_EUlSK_E_EENS1_IbS7_ZNS9_IZNSA_15SingleFlagValueINSC_INS3_IN19MTPDchatAdminRights4FlagEEELj4294967295EE6ChangeES7_ZNSF_IS1E_S7_ZNKSG_IS1E_E6eventsEvEUlSK_E_EclIS1E_S7_ZNSO_IS1E_EEDaSP_EUlSK_E_S1E_S7_EEDaSU_EUlSK_E_EEDaSU_NSI_4EnumEEUlS1C_E_EclIS1C_S7_ZNS9_IZNSB_IS1E_S7_S1K_EEDaSU_SW_EUlRKS1E_E0_EclIS1E_S7_ZNS12_IZNSB_IS1E_S7_S1K_EEDaSU_SW_EUlS1Q_E_EclIS1E_S7_S1K_vEEDaSU_EUlSK_E_S1C_EEDaSU_EUlSK_E_bEEDaSU_EUlSK_E_EENS1_IbS7_ZNS9_IZNS19_INSC_INS3_IN20MTPDchatBannedRights4FlagEEELj4294967295EE6ChangeES7_ZNSF_IS25_S7_ZNKSG_IS25_E6eventsEvEUlSK_E_EclIS25_S7_ZNSO_IS25_EEDaSP_EUlSK_E_S25_S7_EEDaSU_EUlSK_E_EEDaSU_S1L_EUlS23_E_EclIS23_S7_ZNS9_IZNSB_IS25_S7_S2B_EEDaSU_SW_EUlRKS25_E0_EclIS25_S7_ZNS12_IZNSB_IS25_S7_S2B_EEDaSU_SW_EUlS2G_E_EclIS25_S7_S2B_vEEDaSU_EUlSK_E_S23_EEDaSU_EUlSK_E_bEEDaSU_EUlSK_E_EES2Q_EED1Ev

Again, I don't know if this is a valid mangling or not.  But the odds are that
with a newer GDB, you'll be able to go further.

-- 
You are receiving this mail because:
You are on the CC list for the bug.