[BAILOPAN]

This is an update notice for Natural Selection users.

With mahnsawce's permission I have split NS2AMXx into two portions. The first portion is engine specific commands - I have merged these into the normal ENGINE module.

The second portion is the actual Natural Selection natives and stocks. Most of these remained exactly the same, except all natives/stocks are prepended with "ns_".

This module is not the same as NS2AMX because it a)uses the AMXx API and not the 0.9.6+ API and b) is very stripped to NS specific functions. It will be upkept as an "official" module and mahnsawce can upkeep it too if he wishes too.

I heavily recommend NS plugin writers to use this module - and to give input/suggestions back on how to make it better.
Although no binaries are available (Because this is the new API) you can see the new source code here:

Engine source
NS2AMXx source

New Engine and NS2AMXx includes

Some important info:
I have not tested the stocks. If someone could verify that they compile/work I would be very greatful.

Some things I will not add the NS2AMXx or Engine module:
Registering plugin specific forwards for global things. This is unnecessary and there is no advantage (string comparison is the same if used in C or the AMX Machine - equal() is not written in Small). The 0.20 module API improves forward registering by providing a nice set of functions, so there are no other possible improvements. Global forwards have been moved to Engine.
Using mahnsawce's pev() native. This is because entity_set_* conforms to a nice naming standard and is easy to understand and deal with (except for a longer native name length).

Please give any and all feedback, NS users :]

[AssKicR]

Interesting...

[mICKE]

This sounds great, i hope its not to hard to rewrite the plugins to this module

[CheesyPeteza]

Using mahnsawce's pev() native. This is because entity_set_* conforms to a nice naming standard and is easy to understand and deal with (except for a longer native name length).

I don't understand that bit. First you say you are you using his cool new pev() method, then you give reasons for sticking with the old long style? Anyway I just looked up the include files and see no pev() so I assume you left out the "we are not using...".You already know I disagree with this, I think AmxModX has the opportunity to improve here, and I don't think there would be a problem with 2 methods. But I know you aren't going to change your mind so no point arguing anymore, I'll just swear your name when I code the longest most ugly line of code ever seen using the old vexd style pev editing.

I noticed some things while glancing over the include files.

pvPrivateData, I noticed you are still using a default of +5 for linux, do you intend to add the option to change that? As I have mentioned before its not always +5 for every type of entity.

I can't see the keyvalue creation forward. Is that pfn_spawn(entid)?

For PLAYBACK_EVENT_FULL if you could add Floatrigin[3] = {0.0, 0.0, 0.0} etc. so that it can work like PLAYBACK_EVENT then that would be really handy.

Shouldn't these be engine functinos?

/* Returns the number of specified entities in map if value is 0
   otherwise it returns the index of the #th entity of that class. */
native ns_get_ent(classname[],Value);
/* Mirrors MAKE_STRING() */
native ns_make_string(svValue[]);
/* Mirrors STRING() */
native ns_string(iValue,svReturnedValue[]);

And ns_get_ent appears to be basically just find_ent_by_class() but with the option to find the number of values, perhaps you should just modify find_ent_by_class().

Can all engine forwards be blocked?

[QwertyAccess]

i thought you hated pev bailopan

[BAILOPAN]

We can have a flamewar about pev() some other day

I will remove the engine string functions entirely because they're not needed.

DispatchKey is not forwarded yet, but I will do it for the final release. The only engine forward which can't be blocked is ServerFrame().

For pvPrivateData, I think +5 is a global offset no matter what. I changed it so it only adds 5 if it was on linux (before it subtracted 5 if it was on windows). I can change it to what mahnsawce does (that is a defaulted parameter).

What does playback_event_full look like?

Okay, I'll merge find_ent_by_class and ns_get_ent

[CheesyPeteza]

Originally Posted by BAILOPAN For pvPrivateData, I think +5 is a global offset no matter what. I changed it so it only adds 5 if it was on linux (before it subtracted 5 if it was on windows). I can change it to what mahnsawce does (that is a defaulted parameter). What does playback_event_full look like?

Best to just quote the HL SDK Document about the events system:

C. Playing back the Event ( Server side ) Playing back the Event: The way to invoke an event from the game .dll is to invoke the PLAYBACK_EVENT or PLAYBACK_EVENT_FULL API macro or call. PLAYBACK_EVENT simply invokes PLAYBACK_EVENT_FULL with all default parameters: #define PLAYBACK_EVENT( flags, who, index ) PLAYBACK_EVENT_FULL( flags, who, index, 0, (float *)&g_vecZero, (float *)&g_vecZero, 0.0, 0.0, 0, 0, 0, 0 );

For pvPrivateData in NS the buildings have a linux +3 offset. So please add Mahnsawce's method otherwise it makes life tricky.

Oh and Mahnsawce has written a really cool plugin that uses the DispatchKey forward "Capture The Gorge" so it'd be good if you do get it in for the final release, plus I have a plugin I'd like to do for CS using it.

[QwertyAccess]

whats DispatchKey?

[AssKicR]

Originally Posted by QwertyAccess whats DispatchKey?

Something we need

[EKS]

Originally Posted by QwertyAccess whats DispatchKey?

I personaly use it to crash DoD servers, but im sure it has a more usefull function to