develooper Front page | perl.perl5.porters | Postings from January 2008

dtrace and Perl (again)

Thread Next
From:
Andy Armstrong
Date:
January 5, 2008 11:10
Subject:
dtrace and Perl (again)
Message ID:
4033D95A-F418-4CAC-BB0C-34D20AF68A55@hexten.net
(this is Yak Shaving; the background is that I'm using dtrace to  
analyse which temporary files, directories and other global resources  
are used by core's tests)

I've just been reading about Alan Burlison's experiments with dtrace  
and Perl:

http://blogs.sun.com/alanbur/entry/dtrace_and_perl

I took Richard Dawe's patch[1] (mentioned in the above post) and  
applied it to blead on Mac OS. It works fine. Actually dtrace seems to  
have changed slightly since Richard produced his patch with the result  
that it's now slightly less involved: you don't have to create an  
extra object file to link against.

The resulting perl binary is, indeed, dtraceable - which is a rather  
nice tool to have.

There are two test failures:

t/op/ 
caller...................................................Assertion  
failed: (SvTYPE(shplep) == SVt_PVGV || SvTYPE(shplep) == SVt_PVLV),  
function Perl_pp_entersub, file pp_hot.c, line 2764.
FAILED--no leader found

t/op/ 
glob.....................................................Assertion  
failed: (SvTYPE(shplep) == SVt_PVGV || SvTYPE(shplep) == SVt_PVLV),  
function Perl_pp_entersub, file pp_hot.c, line 2764.
FAILED--expected 15 tests, saw 0

I don't /think/ they're going to be too difficult to track down.

Is there any interest in having the usedtrace option in core? Given  
that dtrace seems to be spreading beyond Solaris I can see it being a  
requested feature at some point. If it's likely to be accepted I'll do  
some more work so that it all just works.

[1] http://rich.phekda.org/perl-dtrace/perl-5.8.8-dtrace-20070720.patch

-- 
Andy Armstrong, Hexten





Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About