Re: [PATCH] avoid use after free

Hello Waldek/All, Sorry Waldek, I should have specified it better Actually, I used the clang static analiser to uncover this use after free (the report i have uploaded): http://exdev.sourceforge.net/osv/clang-20180122/report-a87eab.html#EndPath This is a piece of the code, apparently, we use

Re: [PATCH] reduce hostname length in order to follow procfs hostname implementation

I am a little bit reluctant to apply this patch. I do agree that we should be consistent (on other hand core/dhcp.cc uses 256 char long buffer - another place where it is different than 65). But I am afraid that some times especially in cloud environment what gets passed through cloud init may

Re: [PATCH] avoid use after free

I am not really familiar with this code but I do not see where ii is used after original call to if_freenameindex. Do you have specific scenario when something would fail before your patch and now passes after it? Concluding I am not convinced this patch is correct. @nyh what do you think?

Re: mysql: failed looking up symbol numa_all_nodes_ptr

On Sat, Feb 10, 2018 at 7:31 AM, Geraldo Netto wrote: > Dear Friends, > > I was trying to upgrade mysql but in order to upgrade to version 5.6.39 > It seems mysql requires the libraries > - libnuma.so.1 > libnuma is a separate shared library, it is not part of glibc, and