Opened 14 years ago

Closed 14 years ago

#1994 closed Bugs (worksforme)

headers out of order somewhere in function/functional

Reported by: Steven Scott <Steven.Scott@…> Owned by: John Maddock
Milestone: Boost 1.36.0 Component: TR1
Version: Boost 1.35.0 Severity: Problem
Keywords: Cc:

Description

Here's a test case that shows it is not my fault :)

#include <boost/program_options.hpp> int main(){}

compiled with g++ -o po.o -c -g -I/home/sls/include -I/home/sls/include/boost/tr1/tr1 po.cpp

copious compiler errors supplied in the attached file.

Attachments (1)

err2 (240.3 KB ) - added by Steven.Scott@… 14 years ago.
first 2000 lines of compiler output

Download all attachments as: .zip

Change History (6)

by Steven.Scott@…, 14 years ago

Attachment: err2 added

first 2000 lines of compiler output

comment:1 by Marshall Clow, 14 years ago

Component: Noneprogram_options
Owner: set to Vladimir Prus

comment:2 by Vladimir Prus, 14 years ago

Component: program_optionsfunction
Owner: changed from Vladimir Prus to Marshall Clow
Summary: headers out of order somewhere in program_options/function/functionalheaders out of order somewhere in function/functional

Given that the error in inside boost.function headers, and that it's triggered by adding tr1 to includes, I don't think it's my bug at any way. I don't know who's the right person to look at it

comment:3 by Marshall Clow, 14 years ago

Owner: changed from Marshall Clow to Douglas Gregor

Just another data point: I tried g++ -o po.o -c -g -I ~/Sources/boost/release -I~/Sources/boost/release/boost/tr1/tr1 po.cpp on my Mac OS X box (gcc 4.01), and got no errors.

comment:4 by Douglas Gregor, 14 years ago

Component: functionTR1
Owner: changed from Douglas Gregor to John Maddock

Not that I like passing the buck, but if it's the introduction of the Boost.TR1 headers that causes the problem, then it's a Boost.TR1 issue (even if it's headers from another library that need tweaking).

comment:5 by John Maddock, 14 years ago

Resolution: worksforme
Status: newclosed

I believe this has already been fixed in SVN Trunk: either way, I've added a test case for this to the TR1 tests and they pass OK for me here.

John.

Note: See TracTickets for help on using tickets.