A good test is to grab the old screen binary from some not yet updated installation and try the reattach. Comment 2 says that succeeds. This is a strong indicator that we have a changed struct msg.
Please make sure that we always increment the message version counter when we change struct msg! Cheers, JW Am 08.09.2015 16:26 schrieb "Amadeusz Sławiński" <am...@asmblr.net>: > On Tue, 8 Sep 2015 12:08:54 +0200 > Petr Hracek <phra...@redhat.com> wrote: > > > On 08/31/2015 07:41 PM, Amadeusz Sławiński wrote: > > >> I have received a bug > > >> https://bugzilla.redhat.com/show_bug.cgi?id=1253697 > > >> which causes that screen indefinite hangs. > > >> > > >> Strace is a part of the bug. > > >> Did you ever discover this problem? > > >> > > >> This bug was discovered in screen-4.1.0. > > > Hi, > > > > > > I recall that someone on irc had similar problem, it seemed like > > > his /dev/pts was being mounted with wrong options > > > it needed something like those mount options to work: > > > "gid=5,mode=620" (group 5 being tty on my system). > > > > > > Amadeusz > > > > > Hi, > > > > they provided a options: > > > > /dev/pts mount: > > > > $ mount | grep pts > > devpts on /dev/pts type devpts > > (rw,nosuid,noexec,relatime,seclabel,gid=5,mode=620,ptmxmode=000) > > > > But it seems that it is not working still. > > > > What about ptmxmode? > > Could it be relevant? > > Hi, > > I don't think so, mine is also 000. > As far as I can see from bugreport it happens between updates. > What was version before and after update? > Also do you have somewhere diff between those versions? > > It can be that there is some difference in "struct msg" between > versions which causes this problem. > > Does it only fails to reattach after update? > By which I mean, does new sessions after update attach properly? > > Amadeusz > >
