Regression: tests failing on Solaris 10 for >= 2.0.23

[Laurent Blume]

Le 2014/08/14 14:37 +0200, Werner Koch a écrit:
> There is
>
> 571bcd46 * common: Fix build problem with Sun Studio compiler.
>
> thus I wonder how it was possible to build before.

I've been using only GCC4 for a while. I only gave a try to Studio this 
time as it was suggested it could be a compiler optimization issue.

> The agent died, right?  The best way to debug this is to put
>
> --8<---------------cut here---------------start------------->8---
> log-file socket:///home/foo/.gnupg/S.log
> debug 1024
> verbose
> --8<---------------cut here---------------end--------------->8---
>
> into gpg-agent.conf and watch the output using watchgnupg.  You may of
> course also use a plain file.  If you notice that the gpg-agent died,
> you should start it again and attach a debugger to gpg-agent.

Here is what I get in the log:

gpg-agent[10736]: chan_6 -> OK Pleased to meet you, process 10734
gpg-agent[10736]: chan_6 <- RESET
gpg-agent[10736]: chan_6 -> OK
gpg-agent[10736]: chan_6 <- OPTION ttytype=ansi
gpg-agent[10736]: chan_6 -> OK
gpg-agent[10736]: chan_6 <- OPTION allow-pinentry-notify
gpg-agent[10736]: chan_6 -> OK
gpg-agent[10736]: chan_6 <- AGENT_ID

And of course, it's only now that I notice there is a core file 
generated each time in /var/core. Sorry about that, there was no hint of 
it in the output so far:

# pstack core.gpg-agent.10736
core 'core.gpg-agent.10736' of 10736:   gpg-agent --server
  fffffd7fff2679f0 strlen () + 30
  fffffd7fff2c7fd3 vsprintf () + 33
  fffffd7ffef60273 int_vasprintf () + 2c3
  fffffd7ffef58a8d _assuan_debug () + bd
  fffffd7ffef57e41 assuan_set_error () + 51
  fffffd7ffef5bfcf process_request () + 8f
  fffffd7ffef5c048 assuan_process () + 28
  0000000000417db5 start_command_handler () + 145
  000000000043dadf main () + c5f
  00000000004125fc _start () + 6c

Does it help? An strlen() error, that sounds rather trivial?

Laurent