On Wed, Apr 18, 2018 at 07:09:38PM +0100, Chris Lamb wrote:
>
>>>I suspect the underlying problem is that we are not detecting
>>>profiling information on armhf correctly. The relevant code is:
>>>
>>>
>>> https://salsa.debian.org/lintian/lintian/blob/master/checks/binaries.pm#L192-207
>>>
>>>I have attached the "readelf -WltdVs" output of "basic.c" compiled
>>>on the harris.debian.org porterbox.
>>>
>>>Whilst I see a "GLIBC_" section, I do see an mcount:
>>>
>>> 117: 00000000 0 FUNC GLOBAL DEFAULT UND
>>> __gnu_mcount_nc@@GLIBC_2.8
>>>
>>>Can someone with some ELF knowledge chime in here? :)
I think I've sussed it after a lot of digging - the regex you're using
is simply too tight and won't match that:
if ($sec =~ /^GLIBC_.*/ and $sym =~ m{\A _?+ _?+ mcount \Z}xsm){
That's not going to match "__gnu_mcount_nc" at all. Simple fix:
if ($sec =~ /^GLIBC_.*/ and $sym =~ m{\A _?+ _?+ (gnu_)?+mcount(_nc)?+
\Z}xsm) {
works both on armhf and amd64. Wow, debugging lintian tests is a lot of effort!
--
Steve McIntyre, Cambridge, UK. st...@einval.com
"Managing a volunteer open source project is a lot like herding
kittens, except the kittens randomly appear and disappear because they
have day jobs." -- Matt Mackall
