Opened 13 years ago

Closed 13 years ago

#3593 closed Patches (fixed)

[range] regresion test failures in VC10

Reported by: Richard Webb <richard.webb@…> Owned by: Neil Groves
Milestone: Boost 1.43.0 Component: range
Version: Boost Development Trunk Severity: Problem
Keywords: Cc:

Description

The range regresion tests have a number of failures on VC10 Beta 2 due to ambiguities between boost::begin/end and std::begin/end. e.g

http://tinyurl.com/yeg59bs

http://tinyurl.com/y8dx87f

These failures can be avoid by explicitly qualifying calls to begin/end in the tests with boost:: .

Attachments (2)

range.diff (4.4 KB ) - added by Richard Webb <richard.webb@…> 13 years ago.
range2.diff (946 bytes ) - added by rwebb <richard.webb@…> 13 years ago.

Download all attachments as: .zip

Change History (7)

by Richard Webb <richard.webb@…>, 13 years ago

Attachment: range.diff added

comment:1 by rwebb <richard.webb@…>, 13 years ago

Milestone: Boost 1.42.0Boost 1.43.0
Summary: [range] regresion test failures in VC10 Beta 2[range] regresion test failures in VC10
Type: BugsPatches

Theres also a couple of failures due to conflicts between boost::next and std::next, which effect both VC10 and GCC 4.4 in c++0x mode. Additional patch attached.

by rwebb <richard.webb@…>, 13 years ago

Attachment: range2.diff added

comment:2 by Neil Groves, 13 years ago

Owner: changed from Thorsten Ottosen to Neil Groves
Status: newassigned

I've started fixing compiler compatibility issues with Boost.Range

comment:3 by Neil Groves, 13 years ago

Resolution: fixed
Status: assignedclosed

These issues were fixed as part of debugging other compiler compatibility issues with the Boost.RangeEx merge into Boost.Range.

comment:4 by Richard Webb <richard.webb@…>, 13 years ago

Resolution: fixed
Status: closedreopened

The iterator_range test is still failing (as of yesterday): http://tinyurl.com/yblef9g

comment:5 by Neil Groves, 13 years ago

Resolution: fixed
Status: reopenedclosed

Heck. I reproduced the problem on GCC with C++0x enabled and cleared the defects. I assumed this would also clear VC10. It does not due to a missing C++0x library function in the current GCC.

This time it really is fixed for VC10. Sorry for the previous confusion.

Note: See TracTickets for help on using tickets.