Re: [AD] namespace conflict with curses

[Peter Wang <tjaden@xxxxxxxxxx>]

On 2005-08-18, Evert Glebbeek <eglebbk@xxxxxxxxxx> wrote:
> > Do you mean you don't want to fix the hline/vline conflict for 4.2?
> > Unlike the case with BITMAP the hline/vline conflict causes problems at
> > the linker level.  The third option I listed is rather simple, too.
> 
> Also true, and if we want to fix it then the time to do so would be right
> now, before 4.2.0 release, and I think that the third option would be the
> way to go.
> The reason I would hesitate to do this has to do with binary compatibility:
> are we already commited to ABI compatibility with either of the beta's or
> RC1 (which isn't public, so shouldn't matter)? I don't think we strictly
> have to be, but I'd still prefer to be if we can.
> That's my only objection in this case. If you can argue that it's no
> problem to break ABI compatibility with the beta's, then feel free to go
> ahead.

We are not commited to ABI compatibility until the actual release.
Besides, RC1 was not publically disclosed.

Patch attached.  Do you think it's worth mentioning in api._tx that
taking the addresses of hline/vline from different C files is going to
result in different addresses?

I've manually changed dllsyms.lst and put _allegro_[vh]line in place of
[vh]line.  I suppose that works?

Peter

Attachment: vhline.diff.gz
Description: application/gunzip