On Mon, Mar 29, 2021 at 12:12:55PM -0700, Henric Jungheim wrote: > > I ran into a bit of an odd problem when building the > sysutils/grub2-bhyve port on 13.0-RC3 (on x64). The bison command > dumps core when the output is going to a console. Redirecting the > build output to a file avoids the problem. I'm not sure if this is > an ncurses issue, a port issue (and which port?), my box, or > something else (clang?). It might be a problem with the application's initialization of ncurses. > This is what the tail end of the port build looks like: > ... > grub-mkfont: No (explicitly disabled) > grub-mount: No (explicitly disabled) > starfield theme: No (No grub-mkfont) > With libzfs support: Yes > ******************************************************* > ===> Building for grub2-bhyve-0.40_8 > gmake[1]: Entering directory '/usr/ports/sysutils/grub2-bhyve/work/grub2-bhyve-0.40' > bison -d -p grub_script_yy -b grub_script ./grub-core/script/parser.y > ./grub-core/script/parser.y:94.1-12: warning: deprecated directive: '%pure-parser', use '%define api.pure' [ > > And here's what lldb has to say: > > > sysutils/grub2-bhyve/work/grub2-bhyve-0.40>lldb -- /usr/local/bin/bison -d -p grub_script_yy -b grub_script ./grub-core/script/parser.y > (lldb) target create "/usr/local/bin/bison" > Current executable set to '/usr/local/bin/bison' (x86_64). > (lldb) settings set -- target.run-args "-d" "-p" "grub_script_yy" "-b" "grub_script" "./grub-core/script/parser.y" > (lldb) run > Process 50108 launching > Process 50108 launched: '/usr/local/bin/bison' (x86_64) > ./grub-core/script/parser.y:94.1-12: warning: deprecated directive: %pure-parser, use %define api.pure [signal SIGSEGV: invalid address (fault address: 0x0) > frame #0: 0x0000000000000000 > error: Bad address > (lldb) bt > * thread #1, name = 'bison', stop reason = signal SIGSEGV: invalid address (fault address: 0x0) > * frame #0: 0x0000000000000000 > frame #1: 0x0000000800c1930a libncursesw.so.9`delay_output_sp + 234 > frame #2: 0x0000000800c19b81 libncursesw.so.9`tputs_sp + 1377 > frame #3: 0x0000000800c19cfb libncursesw.so.9`tputs + 123 > frame #4: 0x0000000800702dbc libtextstyle.so.0`___lldb_unnamed_symbol149$$libtextstyle.so.0 + 92 > frame #5: 0x00000008007037ff libtextstyle.so.0`___lldb_unnamed_symbol152$$libtextstyle.so.0 + 943 > frame #6: 0x00000008007030cf libtextstyle.so.0`___lldb_unnamed_symbol151$$libtextstyle.so.0 + 543 > frame #7: 0x0000000800700d00 libtextstyle.so.0`___lldb_unnamed_symbol125$$libtextstyle.so.0 + 32 Either bison has to tell libtextstyle to initialize, or libtextstyle has to know when to initialize ncurses. For the latter, term_ostream_create calls setupterm, which interestingly enough tries to recover from an error by calling the terminfo functions (which could still fail): if (setupterm (term, fd, &err) || err == 1) { /* Retrieve particular values depending on the terminal type. */ stream->max_colors = tigetnum ("colors"); stream->no_color_video = tigetnum ("ncv"); setupterm could fail for instance if there's no terminal description corresponding to your $TERM variable. If that happens, then the SCREEN pointer passed to tputs_sp might be NULL, or cur_term might be NULL. In ncurses, the corresponding null-pointer checks are done with HasTInfoTerminal: NCURSES_EXPORT(int) NCURSES_SP_NAME(delay_output) (NCURSES_SP_DCLx int ms) { T((T_CALLED("delay_output(%p,%d)"), (void *) SP_PARM, ms)); if (!HasTInfoTerminal(SP_PARM)) returnCode(ERR); if (no_pad_char) { NCURSES_SP_NAME(_nc_flush) (NCURSES_SP_ARG); napms(ms); } else { > frame #8: 0x0000000000407295 bison`___lldb_unnamed_symbol160$$bison + 1669 > frame #9: 0x0000000000407d75 bison`___lldb_unnamed_symbol177$$bison + 149 > frame #10: 0x000000000040805d bison`___lldb_unnamed_symbol181$$bison + 109 > frame #11: 0x000000000042803f bison`___lldb_unnamed_symbol403$$bison + 7679 > frame #12: 0x000000000042e9c8 bison`___lldb_unnamed_symbol459$$bison + 56 > frame #13: 0x0000000000403743 bison`___lldb_unnamed_symbol130$$bison + 291 > frame #14: 0x0000000000403be0 bison`___lldb_unnamed_symbol130$$bison + 1472 > (lldb) > > _______________________________________________ > freebsd-current_at_freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org" -- Thomas E. Dickey <dickey_at_invisible-island.net> https://invisible-island.net ftp://ftp.invisible-island.net
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:27 UTC