Re: [Xen-devel] [xen-4.7-testing bisection] complete build-amd64

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 12.06.17 at 12:45, <ian.jackson@xxxxxxxxxxxxx> wrote:
> Jan Beulich writes ("Re: [Xen-devel] [xen-4.7-testing bisection] complete 
> build-amd64"):
>> On 11.06.17 at 17:55, <osstest-admin@xxxxxxxxxxxxxx> wrote:
>> >   commit 164c34dd23bc3ea8d5285752d9270627a93c91f5
>> >   Author: Jan Beulich <jbeulich@xxxxxxxx>
>> >   Date:   Fri Jun 9 13:51:34 2017 +0200
>> >   
>> >       hvmloader: avoid tests when they would clobber used memory
>> 
>> I appreciate the test pointing out an oversight of mine, but I
>> wonder why debug builds are being done/run in osstest (as me
>> having done only release builds was the reason for not spotting
>> the issue earlier on) - this way osstest tests what likely only
>> very few people are going to actually use in production.
> 
> Normally, the debug build of some software would (i) generate better
> diagnostics (at the cost of possible information leaks log volume
> problems) and (ii) add additional checks to the code which will detect
> more bugs (at the a possible performance cost).
> 
> osstest doesn't do much perf testing and the other downsides aren't
> relevant.

Are you sure? Higher optimization levels may expose bugs (and
not only such in the compiler's optimizer).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel