There is a "/proc/self/exe" support under GNU/kFreeBSD, but it is
limited, namely inside combination of bind mounts and chroots.
What's the problem with /proc/self/exe anyway? It works fine here,
even inside chroots.
Did you nullfs-mount /proc?
No, I expect that the problem is when the execu
2011/10/19 Petr Salinger :
>>> There is a "/proc/self/exe" support under GNU/kFreeBSD, but it is
>>> limited, namely inside combination of bind mounts and chroots.
What's the problem with /proc/self/exe anyway? It works fine here,
even inside chroots.
Did you nullfs-mount /proc?
--
Robert Mill
There is a "/proc/self/exe" support under GNU/kFreeBSD, but it is
limited, namely inside combination of bind mounts and chroots.
It is possible that this is a linuxism, but if this should not be used on
bsd, wouldn't it be better if the proc filesystem is NOT available on the
porter boxes? Or
2011/10/19 Christian T. Steigies :
> It is possible that this is a linuxism, but if this should not be used on
> bsd, wouldn't it be better if the proc filesystem is NOT available on the
> porter boxes?
/proc is much more than /proc/self/exe. In particular it is needed by
ps, killall, etc.
--
R
On Tue, Oct 18, 2011 at 08:51:06AM +0200, Petr Salinger wrote:
>
> The problem seems be in src/gle/file_io.cpp:
>
> #if defined(__UNIX__) && !defined(__MACOS__) && !defined(__FREEBSD__)
> /* try to read location from the /proc/self/exe file */
> char path[PATH_MAX];
> stru
So to summarize, both you and I can build the package on kfreebsd
machines (debian porter boxes and your box),
but the buildds fail to build since 4.2.2-4+b2,
whereas 4.2.2-4+b1 and 4.2.2-4 have built successfully. Note
that in a binNMU the source code has not been changed. From this I conclude
t
6 matches
Mail list logo
