Front page | perl.perl5.porters |
Postings from October 2014
On Thu Jun 05 14:03:56 2014, greg.lindahl wrote: > Thanks. I agree that your example does not reproduce my actual > problem... it's more subtle. > > I started wading into the test. It's a single test of our text > tokenization code that sees the problem, on a long string which > generates many short tokens. The newSVpvf call usually takes 1,200 cpu > cycles for short strings, but at the start of this test it takes > 300,000 cpu cycles (yes, on the same short strings) and then slowly > falls back to 1,200 cpu cycles by the end of this single test. I need > to walk up the call stack and see what kind of loop is involved. > > (Thank you for the attempt at reproducer code, I don't know enough XS > to have set that up!) Are you able to provide more detailed instructions for reproducing this problem? Or should I close the ticket? Tony --- via perlbug: queue: perl5 status: open https://rt.perl.org/Ticket/Display.html?id=121984Thread Previous