Client crashes upon login

[Luc1]

Since the last update to the client, Spotify kicks me back to login instead of remembering that I am logged in (it remembered previously). Then when I try to log in, it crashes (segfault). I fill in my username, fill in any password (correct or incorrect) and click "Log in" or hit enter. As soon as I click that button, it segfaults. I tried toggling the Remember Me option to no avail.

Console output:

$ spotify
[0104/104352:ERROR:main_delegate.cc(752)] Could not load cef_extensions.pak
[0104/104352:ERROR:main_delegate.cc(752)] Could not load cef_extensions.pak
[0104/104354:ERROR:PlatformKeyboardEvent.cpp(94)] Not implemented reached in static bool blink::PlatformKeyboardEvent::currentCapsLockState()
Segmentation fault

Strace:

...
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(3, {msg_namelen=0}, 0)          = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=20, events=POLLIN}, {fd=21, events=POLLIN}], 5, 300) = 1 ([{fd=21, revents=POLLIN}])
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
read(21, "!", 2)                        = 1
recvmsg(3, {msg_namelen=0}, 0)          = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=20, events=POLLIN}, {fd=21, events=POLLIN}], 5, 0) = 0 (Timeout)
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(3, {msg_namelen=0}, 0)          = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=20, events=POLLIN}, {fd=21, events=POLLIN}], 5, 0) = 1 ([{fd=21, revents=POLLIN}])
recvmsg(20, {msg_namelen=0}, 0)         = -1 EAGAIN (Resource temporarily unavailable)
read(21, "!", 2)                        = 1
gettid()                                = 26304
gettid()                                = 26304
gettid()                                = 26304
gettid()                                = 26304
gettid()                                = 26304
gettid()                                = 26304
gettid()                                = 26304
gettid()                                = 26304
epoll_ctl(64, EPOLL_CTL_MOD, 63, {EPOLLIN|EPOLLERR|EPOLLET, {u32=1342405048, u64=140030161156536}}) = 0
futex(0x2d89430, FUTEX_WAKE_PRIVATE, 1) = 1
gettid()                                = 26304
gettid()                                = 26304
+++ killed by SIGSEGV +++
Segmentation fault

gdb output:

...
Thread 19 "Core Thread" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fff937fd700 (LWP 26389)]
0x0000000000606d31 in ?? ()
(gdb) bt
#0  0x0000000000606d31 in ?? ()
#1  0x0000000000605727 in ?? ()
#2  0x000000000062538a in ?? ()
#3  0x000000000062532a in ?? ()
#4  0x000000000062525d in ?? ()
#5  0x0000000000625239 in ?? ()
#6  0x00000000006251e1 in ?? ()
#7  0x0000000000625107 in ?? ()
#8  0x00000000006111a0 in boost::asio::detail::task_io_service_operation::complete(boost::asio::detail::task_io_service&, boost::system::error_code const&, unsigned long) ()
#9  0x0000000000610d6a in boost::asio::detail::task_io_service::do_run_one(boost::asio::detail::scoped_lock<boost::asio::detail::posix_mutex>&, boost::asio::detail::task_io_service_thread_info&, boost::system::error_code const&) ()
#10 0x0000000000610a85 in boost::asio::detail::task_io_service::run(boost::system::error_code&) ()
#11 0x00000000006109d6 in boost::asio::io_service::run() ()
#12 0x00000000006109a0 in ?? ()
#13 0x0000000000610951 in ?? ()
#14 0x0000000000610930 in ?? ()
#15 0x00000000006108a9 in ?? ()
#16 0x000000000158f25f in ?? ()
#17 0x00007ffff1518464 in start_thread (arg=0x7fff937fd700) at pthread_create.c:333
#18 0x00007fffedfc19df in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:105

I'm on Debian Stretch with Spotify version 1.0.45.186.g3b5036d6 (latest from repository).

I tried removing .config/spotify and .cache/spotify, again to no avail. There is no --verbose option or anything either.

[jooon · Accepted solution]

It is a known problem

https://community.spotify.com/t5/Desktop-Linux-Windows-Web-Player/Spotify-1-0-45-186-crashes-on-logi...

and a fixed version (1.0.47) will be released in a few days to the testing repository. Meanwhile, you can downgrade to version 1.0.44 from the stable repository.

There is no --verbose, but there is --show-console that will print more things, rarely anything relevant for the crash though. The flag is called --show-console because it opens a console window in Windows.

[Luc1]

Thanks. Weird that such a breaking bug is not just fixed with a hotfix instead of letting it go through the testing repository (and that "in a few days") before landing in stable.

[jooon]

Yeah, I agree. We don't have the same thorough release procedure for the Linux client as we do for Win/Mac.

Since 1.0.44 is still in stable, this only affects users who are using 1.0.45 from testing who login with username and password. You are the first user that has explicitly stated that you were logged out, but I am guessing others who reported the bug also were. Most testing users don't see this bug, because they were already logged in and upgraded to 1.0.45 and would not need to have login again, because they log in with a stored token (similar to how cookies on the web work) which was not affected, or they use the Facebook login button which was also not affected.

If it had happened for the stable release, I might have forced a quick update somehow.

[Luc1]

I was actually logged in but it kicked me out. I think I remember being able to log in (and being annoyed at having to log in ever time, despite having set "remember me") before this segfault started appearing, but I'm not sure. Anyway, we'll see if .47 fixes that too. Thanks! 🙂

[jooon]

It should be fixed in 1.0.47 in testing now.