develooper Front page | perl.perl5.porters | Postings from February 2020

Re: How to bisect for first commit at which a test passes?

Thread Previous | Thread Next
James E Keenan
February 14, 2020 22:49
Re: How to bisect for first commit at which a test passes?
Message ID:
On 2/13/20 3:56 PM, wrote:
> James E Keenan <> wrote:
> :$ sh ./Configure -des -Dusedevel -Duseithreads \
> :                      -Dcc="g++7" \
> :-Accflags="-Wl,-rpath=/usr/local/lib/gcc7" \
> :-Aldflags="-Wl,-rpath=/usr/local/lib/gcc7" \
> After rather more back and forth, it turned out after all to be rather
> simple - it goes wrong only if you build perl with -Duseithreads, so
> you have to pass that flag to as well. Without that the
> script does not die, so the bisector is correctly reporting that it
> is not failing as expected at the start revision.
> As such, this invocation:
>    Porting/ --start=a3c77565 --end=a06a4d45 \
>      -Dcc="g++7" -Accflags="-Wl,-rpath=/usr/local/lib/gcc7" \
>      -Aldflags="-Wl,-rpath=/usr/local/lib/gcc7" -Duseithreads \
>      --expect-fail --target t/op/sprintf2.t
> .. successfully finds a06a4d45 as the first commit succeeding, and I'm
> sure would also find it for a wider range such as v5.31.2 to v5.31.3.
> Hugo

Hugo, many thanks for your investigation.  I can confirm that what was 
missing in the invocation with which I began this thread was 

I know this is not the first time my attempts to bisect some problem 
have been befuddled by "Runner returned 256, not 0 for start revision". 
So I'm going to take notes when I/we encounter this again so that we can 
add to the documentation in Porting/

Thank you very much.
Jim Keenan

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About