Re[3]: Status of MC

[Pavel Tsekov <ptsekov gmx net>]

Hello, Pavel!

PR> Actually, I applied a patch today that resembles my original patch, but it
PR> may be more reliable.

[snip patch]

PR> The difference is that old_mask is "repared" after sigprocmask(), not
PR> before, so if the "glitch" happens in _that_ sigprocmask(), we still can
PR> compensate for it.

Cool! :)

PR> I tried really hard and could not make mc hang with the subshell when
PR> using the new patch.  I'm using the current binary version Cygwin that I
PR> update every day by setup.exe.

Yes! It's not guranteed that anyone running a 'release' dll will suffer
from the SIGSTOP/SIGCONT race. But I usually run the latest snapshot
dll compiled for debugging and it is a bit slower so that is what will
for sure trigger it - anyway there is small number of people who run
a snapshot dll with debugging enabled. I guess heavy load may trigger
it also. But now I really think  this is not a big deal - I will write
something in the Cygwin specific README about that :)

PR> I think this patch should be sufficient, and there is no need to disable
PR> the subshell for older Cygwin versions, unless you can make mc hang even
PR> with this patch, which is now on CVS.

I agree.

[snip]

PR> It didn't work on Linux and caused a crash on Cygwin.  I didn't try to
PR> investigate the reasons for the crash, I simply replaced vfs_die() with
PR> _exit(), which was sufficient.

I'll try to debug it, just out of curiosity, if a I find some free time :)

PR> While fixing this, I also fixed another remote buffer overflow in fish,
PR> hopefully not exloitable (because it can overwrite only one byte after the
PR> buffer), but nasty nevertheless.

Btw I've compiled yesterday MC snapshot (2002-08-18) on an old
sparcstation2 running OpenBSD 3.0. I found that they have MC in their
ports three but it is marked as broken - because of many security
flaws. Maybe all your efforts to fix this overflows will make them
change the current status of things. Btw the MC in their ports tree is 4.5.51 -
pretty old one ;)

Pavel Tsekov