From @khwilliamson

After having fixed several memory leaks that I knew how to, or Tony Cook gave me advice on fixing, there remain a bunch. Tony indicated that we have lived with some for a long time, so those are not a priority to fix now, but obviously it would be nice to. But we do want to not introduce memory leaks. Some of these are coming from, for example, multi deref, which is fairly new. So they should be looked at anyway before 5.24, so I will add this ticket to the blockers. Attached is a .gz of the output of clang on a non-threaded build, with the leak sanitizer enabled.
--
Karl Williamson

From @khwilliamson

On Fri Mar 04 10​:51​:00 2016, khw wrote​:

Oops, attached is a 'make test' output that has the lines with embedded \n expanded out, so it is more readable.

--
Karl Williamson

From @khwilliamson

test.out.gz

From @iabyn

On Fri, Mar 04, 2016 at 10​:51​:00AM -0800, Karl Williamson wrote​:

# New Ticket Created by Karl Williamson
# Please include the string​: [perl #127657]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl5/Ticket/Display.html?id=127657 >

After having fixed several memory leaks that I knew how to, or Tony Cook
gave me advice on fixing, there remain a bunch. Tony indicated that we
have lived with some for a long time, so those are not a priority to fix
now, but obviously it would be nice to. But we do want to not introduce
memory leaks. Some of these are coming from, for example, multi deref,
which is fairly new. So they should be looked at anyway before 5.24, so
I will add this ticket to the blockers. Attached is a .gz of the output
of clang on a non-threaded build, with the leak sanitizer enabled. --

I disabled LeakSanitizer locally (with ASAN_OPTIONS='detect_leaks=0')
because on my system, even the most basic of executions, e.g.
  PERL_DESTRUCT_LEVEL=2 ./perl -e 1
gave loads of (what I assumed were) false positives.

What version of clang are you using, and/or what (if any) report
suppression are you using?

--
Overhead, without any fuss, the stars were going out.
  -- Arthur C Clarke

The RT System itself - Status changed from 'new' to 'open'

From @khwilliamson

On 03/04/2016 04​:44 PM, Dave Mitchell wrote​:

On Fri, Mar 04, 2016 at 10​:51​:00AM -0800, Karl Williamson wrote​:

# New Ticket Created by Karl Williamson
# Please include the string​: [perl #127657]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl5/Ticket/Display.html?id=127657 >

After having fixed several memory leaks that I knew how to, or Tony Cook
gave me advice on fixing, there remain a bunch. Tony indicated that we
have lived with some for a long time, so those are not a priority to fix
now, but obviously it would be nice to. But we do want to not introduce
memory leaks. Some of these are coming from, for example, multi deref,
which is fairly new. So they should be looked at anyway before 5.24, so
I will add this ticket to the blockers. Attached is a .gz of the output
of clang on a non-threaded build, with the leak sanitizer enabled. --

I disabled LeakSanitizer locally (with ASAN_OPTIONS='detect_leaks=0')
because on my system, even the most basic of executions, e.g.
PERL_DESTRUCT_LEVEL=2 ./perl -e 1
gave loads of (what I assumed were) false positives.

What version of clang are you using, and/or what (if any) report
suppression are you using?

This happened to me too, but not in current blead; I assumed I had fixed
enough leaks that things would build. But a way to get it to work,
leaks and all, is to say

LSAN_OPTIONS="exitcode=0;
external_symbolizer_path=/usr/bin/llvm-symbolizer-3.5"
PERL_DESTRUCT_LEVEL=2 make test

On dromedary, I don't have to use the symbolizer.

$ clang -v
Ubuntu clang version 3.6.0-2ubuntu1 (tags/RELEASE_360/final) (based on
LLVM 3.6.0)
Target​: x86_64-pc-linux-gnu
Thread model​: posix

Hmm. maybe I should be using symbolizer-3.6. But it seemed to work fine

From @jkeenan

On Fri Mar 04 10​:58​:34 2016, khw wrote​:

On Fri Mar 04 10​:51​:00 2016, khw wrote​:

Oops, attached is a 'make test' output that has the lines with
embedded \n expanded out, so it is more readable.

Karl, what were the tools or commands you used to generate this output?

Thank you very much.

--
James E Keenan (jkeenan@​cpan.org)

From @khwilliamson

On 03/18/2016 05​:55 AM, James E Keenan via RT wrote​:

On Fri Mar 04 10​:58​:34 2016, khw wrote​:

On Fri Mar 04 10​:51​:00 2016, khw wrote​:

Oops, attached is a 'make test' output that has the lines with
embedded \n expanded out, so it is more readable.

Karl, what were the tools or commands you used to generate this output?

Thank you very much.

./Configure -des -Uversiononly -Dprefix=/home/khw/devel -Dusedevel
-D'optimize =-ggdb3' -A'optimize=-ggdb3' -A'optimize=-O0'
-Accflags='-DPERL_BOOL_AS_CHAR' -Accflags='-DPERL_EXTERNAL_GLOB'
-Dman1dir=none -Dman3dir=none -DDEBUGGING -Dusemorebits -Dcc=clang
-Accflags='-fsanitize=address' -Aldflags='-fsanitize=address'
-Alddlflags='-shared'

$ clang -v
Ubuntu clang version 3.6.0-2ubuntu1 (tags/RELEASE_360/final) (based on
LLVM 3.6.0)
Target​: x86_64-pc-linux-gnu
Thread model​: posix