The WIP Thread
- DoomRater
- Posts: 8270
- Joined: Wed Jul 28, 2004 8:21 am
- Preferred Pronouns: He/Him
- Location: WATR HQ
- Contact:
Like Graf's gonna want to do that.
Anyway, here's a shot of a gun I'm taking pics of to attempt to Doomify since I thought it looked so badass (may have to retake these with a better camera): http://img213.imageshack.us/img213/9967 ... adymi8.jpg
Anyway, here's a shot of a gun I'm taking pics of to attempt to Doomify since I thought it looked so badass (may have to retake these with a better camera): http://img213.imageshack.us/img213/9967 ... adymi8.jpg
It's possible, but that's for the SkullTag coders' to worry about.
I suppose it depends. I doubt the client side flag can't be backported to ZDoom unless it changes networking model, so it doesn't have a value in that regard. But when it comes to Graf I suspect the decision would rest on just how many ZDoom maps use the flag, either for intended SkullTag support or for the hell of it, and thus how many maps would be broken by a new ZDoom flag. If there's not a whole lot to lose, then I can't see him really caring.DoomRater wrote:Like Graf's gonna want to do that.
I don't know. Sometimes the situation calls for it. It's really also could happen for any of a number of unpredicted reasons. It also just may never come to ZDoom needing an extra A_SpawnItemEx flag. Compatibility with SkullTag is an overall good thing to be sure, but it's cracks like this that threaten that. Personally I think the burden of compatibility is on SkullTag's side. I don't know how far the developers intend to go with maintaining friendly compatibility relations, but to me SkullTag is a specialized offshoot port, as opposed to ZDoom which is more general and base, so SkullTag needs to be maintaining compatibility with ZDoom if they want it.DoomRater wrote:Why break a good thing?
Sorry, Zippy, but that post is a steaming pile of shit. (and I mean that in as nice a way as possible.)
There is absolutely no harm in ST implementing a useful flag that happens to only make sense for their port.
On the contrary there would be very much harm in Graf (or another coder) blatantly ignoring ST's flag and intentionally breaking compatibility.
The flag is known to be used by a ZDoom port. The smart thing to do would be to put a note in the source code marking flag 128 as "reserved by SkullTag" so it doesn't get accidentally used.
And what if SpawnItemEx does have need for a new flag? Oh no, what to do then!?
Um...
256
512
1024
2048
4096
8192
16384
32768
65536
.
.
.
Duh.
There is absolutely no harm in ST implementing a useful flag that happens to only make sense for their port.
On the contrary there would be very much harm in Graf (or another coder) blatantly ignoring ST's flag and intentionally breaking compatibility.
The flag is known to be used by a ZDoom port. The smart thing to do would be to put a note in the source code marking flag 128 as "reserved by SkullTag" so it doesn't get accidentally used.
And what if SpawnItemEx does have need for a new flag? Oh no, what to do then!?
Um...
256
512
1024
2048
4096
8192
16384
32768
65536
.
.
.
Duh.
While that's certainly a possibility, I'd shy away from it. You never know what may happen down the road. If ST at some point does implement support, then they'll be stuck having to find a new flag number to use to do the same thing ZDoom's flag 128 would do... which then ZDoom would have to mark as reserved. In short, it would be a mess. The same would hold true if at some future point randy decides to implement a C/S model into ZDoom.
As unlikely as either possibility is, there's still a slight chance of either happening and it would just be better coding practice to avoid such conflicts.
There's just no reason not to leave 128 as an unavailable flag permanently (or until the same facility becomes useful to ZDoom). Even if you run out of numbers, you can just move on to a second (or third, or fourth) flag argument.
As unlikely as either possibility is, there's still a slight chance of either happening and it would just be better coding practice to avoid such conflicts.
There's just no reason not to leave 128 as an unavailable flag permanently (or until the same facility becomes useful to ZDoom). Even if you run out of numbers, you can just move on to a second (or third, or fourth) flag argument.
Oh, don't worry about that. I think we've dealt with each other long enough for me to know that that's just how you express your endearment.HotWax wrote:Sorry, Zippy, but that post is a steaming pile of shit. (and I mean that in as nice a way as possible.)
Yeah, it's not a serious issue because there are ways around it. I still think that it is SkullTag's concern, though. I think they could have taken a more compatibility friendly approach (maybe attaching client-side-ness to specific actors, etc.). I just don't see it as ZDoom's concern if it were to get in the way as a problem, in part because it doesn't appear to me that development between the two is tied closely enough together to merit it. There's also the matter that the directions between the two ports are different, and how far things will diverge/converge as the developers of their respective ports decide in which direction they want to continue or what direction they are willing to part from.
Also, what got me giggling the most about this:
My, my, how the tables have turned! This seems to be somewhat the opposite of the HUD discussion.HotWax wrote:there's still a slight chance of either happening and it would just be better coding practice to avoid such conflicts.
There's just no reason not to leave 128 as an unavailable flag permanently (or until the same facility becomes useful to ZDoom).