samth
2017-8-16 21:10:50

@mflatt running the handin server for our course causes racket to segfault consistently; this didn’t happen previously (ie a year ago)


samth
2017-8-16 21:11:00

this happens both with 6.6 and 6.10


samth
2017-8-16 21:11:05

stacktrace looks like:


samth
2017-8-16 21:11:09
#2  0x00000000006c2019 in fault_handler ()                                                                                [0/1442]
#3  0x00007ffff76c85e0 in <signal handler called> () at /lib64/libpthread.so.0
#4  0x00007ffff738601f in __rawmemchr_sse2 () at /lib64/libc.so.6
#5  0x00007ffff736ffa2 in _IO_str_init_static_internal () at /lib64/libc.so.6
#6  0x00007ffff735f1b7 in __isoc99_vsscanf () at /lib64/libc.so.6
#7  0x00007ffff735f157 in __isoc99_sscanf () at /lib64/libc.so.6
#8  0x00007fffe62a7862 in epoxy_glx_version () at /lib64/libepoxy.so.0
#9  0x00007fffe7754757 in gdk_x11_screen_init_gl () at /lib64/libgdk-3.so.0
#10 0x00007fffe7754cd8 in _gdk_x11_screen_update_visuals_for_gl () at /lib64/libgdk-3.so.0
#11 0x00007fffe775da95 in _gdk_x11_screen_init_visuals () at /lib64/libgdk-3.so.0
#12 0x00007fffe775a910 in _gdk_x11_screen_new () at /lib64/libgdk-3.so.0
#13 0x00007fffe774a278 in _gdk_x11_display_open () at /lib64/libgdk-3.so.0
#14 0x00007fffe771e389 in gdk_display_manager_open_display () at /lib64/libgdk-3.so.0
#15 0x00007fffe54af0da in gtk_init_check () at /lib64/libgtk-3.so.0
#16 0x00000000006be858 in ffi_call_unix64 ()
#17 0x00000000006be21f in ffi_call ()

mflatt
2017-8-16 21:12:05

Does running just gracket crash in the same way?


samth
2017-8-16 21:12:12

it goes away with PLT_GTK2=1, which I just tested


samth
2017-8-16 21:12:36

and yes, gracket crashes the same way


samth
2017-8-16 21:12:59

I don’t think PLT_GTK2 was needed a year ago (although that may have been an older Racket)


samth
2017-8-16 21:13:06

(older than 6.6, that is)


mflatt
2017-8-16 21:13:56

And other Gtk+ 3 apps work? I ask because the crash appears to be in gtk_init_check.


samth
2017-8-16 21:16:12

no, they don’t — same segfault with xvfb


samth
2017-8-16 21:16:21

so probably not racket’s fault


samth
2017-8-16 21:16:32

I will report this to other people, then — thanks