Re: [AD] set_config_file jumping around on OS X bundles |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
- To: alleg-developers@xxxxxxxxxx
- Subject: Re: [AD] set_config_file jumping around on OS X bundles
- From: Chris <chris.kcat@xxxxxxxxxx>
- Date: Mon, 22 Aug 2005 14:15:18 -0700
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:from:to:subject:date:user-agent:references:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:message-id; b=fwYKm/KQzDCqskws5xNlZ3RCCPf+grseICbrYdkkx+5E2gYrNCFtw4W2zDv7Co05Cs0q6XpL6n1tyL+32IE/5fXmAMrd52dvBLqlGASk9ngQEO4ogdE6H3+Bwd8pwt6sKswXdSl02mv47uOXvKtQoG4Rb27O8u8PoSdyc3rUmJg=
On Monday 22 August 2005 12:26 pm, Evert Glebbeek wrote:
> I'd rather not. I don't think there's much difference in adding a null
> vtable entry and adding one with functions filled in.
I'm not sure what you mean by this. AFIAK, if we add an extra vtable entry now
that's empty on all platforms, we can fill it in later as long as the DLL
doesn't get any extra global functions or variables. Since the vtable is/can
be filled with static/non-global functions, simply filling it in later with a
static function wouldn't break ABI compatiblity in either direction (as long
as you take proper care that it can be NULL). Same with any global functions
we'd need.. make the functions, but leave them as null-ops, then fix them up
in 4.2.1.
But it's something we'd have to do now. If we add the extra global symbols
now, the DLL should be fine when the functions themselves are filled in
later. This won't break backwards ABI compatibilty since a 4.2.0 program
would still work with a 4.2.1 DLL (all the same symbols are still present,
and the struct sizes are the same). Unless the program tried to use the
null-op functions, but as long as we don't announce it in the public API yet,
it would be their fault for using it.