From: Andreas Scherer (as_at_[hidden])
Date: 2000-03-06 02:04:48
beman dawes <bema-_at_[hidden]> wrote:
> As far as general "cleaning" of boost file line terminators, what do
> you consider "clean"? CR, CR/LF, or LF?
Dunno, maybe the majority of EOFs in <smart_ptr.hpp>?
> I'm sure you see the
> problem; if boost standardized on a particular platform's terminator,
> people using other platforms take offense.
As I posted in another thread, sticking to exactly one kind of EOFs
would greatly improve the portability of the source files. If EOFs are
used consistently, files can be adapted to the local system
automatically, either by setting certain options of the "unzip" program
or by applying "recode". (Version 3.5 of the latter hickups on
"mixed-mode" files: "Ambigous output in data..CR-LF")
The funny thing is, that on this WinNT machine (with VC6) I did not
encounter problems after applying some patches. "patch 2.5" not only
has introduced VC-specific modifications but also has clean
<smart_ptr.hpp> for Win32 format.
> to tell you the truth I am surprised there are still
> compilers out there that can't handle any combination of CR, CR/LF,
> or LF.
The "incorrect" EOF in line 213 of <smart_ptr.hpp> and its effect on
the preprocessor can easily be seen when you open the file in VIM 5.
This editor shows the "incorrect" EOFs as "^M", hence the subject line.
Thanks for fixing this minor problem.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk