On Thu, Dec 23, 2010 at 7:31 PM, Michel MORIN <mimomorin@gmail.com> wrote:
Hi Neil,
Neil Groves wrote:
I will submit a fix to boost-trunk today. After a couple of days of passes on boost-trunk I shall merge to boost-release.
The ticket (#5014) just has been closed, but the test case in the ticket still prints out wrong results.
You are correct. The change I made yesterday on the trunk for the strided adaptor is fundamentally flawed in a number of ways. I have committed a new fix to the trunk and added your test cases, plus additional test cases to cover all of the flaws I noticed from yesterdays changes.
I don't know whether we should implement decrement() function for Bidirectional Range.
My new implementation works safely for all traversal types, and correctly and safely handles decrement.
Regards, Michel
Please accept my apologies for the previous 'intellectually challenged' commit. Regards, Neil Groves