Lars Gullik Bjønnes wrote:
> | Hmmmm, I see it's gone in, but it looks to me like this patch has the
> | potential to break Cygwin and Windows builds.
>
> Sure... unfortunately it seems that the only way to get (cyg)Win
> people to comment on the patch is to check it in and wait for breakage
> reports.
Understood.
> | You construct fs::path objects on the fly from the filename strings
> | that we (LyX) store. That's fine on posix machines because we store all
> | filenames as posix-like paths.
>
> There was no special handling in FileInfo for differences like this,
> we just used what we got... should be the same now.
Maybe. I was just flagging up a possible reason for breakage. Presumably
in the future we'll store fs::path variables rather than as std::strings.
Will enable us to use fs::native format perhaps.
There will still be 'interesting things to do with files that are to be
found using {BIB,TEX}INPUT environment variables though.
> | Will things continue to work on Cygwin and
> | on Windows?
>
> No. functions is fs_extras are not implemented for win.
> Testing is not quite enought... we need some doer as well.
Ahhh well. Breakage leads to progress. Haven't I read, however, that
Boost.Filesystem doesn't behave well on Cygwin ATM? Maybe I just dreamt
that.
--
Angus
