Attached Particle Spawners #4409

raymoo · 2016-08-04T20:17:23Z

Implements #567 (attached particle spawners).

Here is the code for a test mod implementing rain: http://lpaste.net/174254

The particles in the demo are shown to all players, so you can also use it to test if the spawners get deleted when an object is removed, by entering with multiple clients and logging out of one.

juhdanad · 2016-08-04T20:44:06Z

src/particles.cpp

+		if (attached != NULL)
+		{
+			// CAO coordinates in irrlicht units, but particles are
+			// not.


You could use BS there.

juhdanad · 2016-08-04T20:58:53Z

If ActiveObjects contained references to the particle spawners, the spawner would be deleted when the object is deleted/unloaded. The class ActiveObject could extend ParticleSpawner to spawn particles. Or ParticleManager could also support ActiveObjects as spawners. Or you could create ParticleSpawner::removeSpawner() which the ActiveObject calls when it is unloaded or deleted.
I hope I could help.

raymoo · 2016-08-04T21:21:49Z

Thanks @juhdanad. I've already started implementing that part, and have gone the route of having the server manage the lifetime of attached particle spawners.

raymoo · 2016-08-04T21:49:42Z

Server-side lifetime handling implemented.

raymoo · 2016-08-04T21:54:35Z

I forgot to document this.

EDIT: documented

Bremaweb · 2016-08-05T02:54:28Z

I love it, seems to work really well.

raymoo · 2016-08-05T04:30:18Z

I noticed that you can use 0 as a "null object id" so I changed the code to use u16 instead of s32 for sending and handling ids.

raymoo · 2016-08-05T20:10:44Z

Rebased and squashed

sofar · 2016-08-08T18:23:52Z

Question: Can we get an interface that is similar to player:set_attach() and player:detach() ? Maybe that will make it easier to avoid the "0" id problem? Just a thought.

raymoo · 2016-08-08T21:20:22Z

@sofar It would only avoid it in the add_particlespawner packet, which I don't think is a big deal. It could already be avoided in the client by keeping a table of attachments, as it is done on the server, but I didn't do that because handling and tracking the offset of particles logically belongs in the particle spawner code.

raymoo · 2016-08-08T23:48:33Z

Also using a set_attach would either require an API change to add_particlespawner, which currently returns a number, or adding a set_attach to the metatable for numbers, which doesn't work if we find another "id-represented thing" that we want to add a set_attach to.

sofar · 2016-09-16T21:53:41Z

src/content_sao.cpp

@@ -156,6 +156,13 @@ LuaEntitySAO::~LuaEntitySAO()
 	if(m_registered){
 		m_env->getScriptIface()->luaentity_Remove(m_id);
 	}
+
+	std::set<u32>::iterator it;
+	std::set<u32>::iterator begin = m_attached_particle_spawners.begin();


I think in most of the code you'll find that begin/end are just omitted and m_attached_particle_spawners.begin() is just used directly in the for( loop.

I'll put them inline but I'm not going to make a line with 100+ columns

sofar · 2016-09-16T21:54:12Z

src/environment.cpp

+	u32 id = addParticleSpawner(exptime);
+	m_particle_spawner_attachments[id] = attached_id;
+	ServerActiveObject *obj = getActiveObject(attached_id);
+	if (obj != NULL) {


I prefer if (obj) {

You just said to use iterator methods inline since the rest of the code does it, but in this file the code typically uses != NULL.

sofar · 2016-09-16T21:55:13Z

src/particles.cpp

+	v3f attached_offset = v3f(0,0,0);
+	if (m_attached_id != 0) {
+		ClientActiveObject *attached =
+		env->getActiveObject(m_attached_id);


fits on one line?

it goes a bit above 80 columns, but I guess that the style guide says that's ok

sofar · 2016-09-16T21:55:28Z

src/particles.cpp

+	if (m_attached_id != 0) {
+		ClientActiveObject *attached =
+		env->getActiveObject(m_attached_id);
+		if (attached != NULL) {


if (attached) {

sofar · 2016-09-16T21:57:34Z

src/server.cpp

@@ -3200,11 +3202,19 @@ u32 Server::addParticleSpawner(u16 amount, float spawntime,
 		peer_id = player->peer_id;
 	}

-	u32 id = m_env->addParticleSpawner(spawntime);
+	u16 attached_id = attached == NULL ? 0 : attached->getId();


u16 attached_id = attached ? attached->getId() : 0; ?

sofar

Looks pretty good.

raymoo · 2016-09-17T20:06:04Z

I've made your changes, just making sure it builds

raymoo · 2016-09-17T20:11:59Z

actually hold on, I missed one.

raymoo · 2016-09-20T00:15:20Z

Here is a demo with particles used to add effects to a projectile: https://a.desu.sh/fywvhg.webm

raymoo · 2016-10-06T06:14:11Z

Rebased.

sfan5 · 2016-10-06T10:22:39Z

src/particles.cpp

+			attached_offset = attached->getPosition() / BS;
+		} else {
+			unloaded = true;
+		}


removal of braces possible:

if (attached) attached_offset = attached->getPosition() / BS; else unloaded = true;

sfan5 · 2016-10-06T10:24:02Z

src/server.cpp

+		id = m_env->addParticleSpawner(spawntime);
+	} else {
+		id = m_env->addParticleSpawner(spawntime, attached_id);
+	}


removal of braces possible

sfan5 · 2016-10-06T10:24:38Z

I have to say:

Really really good job! 👍

raymoo · 2016-10-06T20:22:56Z

@sfan5 Thanks, also made the suggested changes.

sofar · 2016-10-12T00:23:13Z

👍 from me, very happy to see this patch :)

nerzhul · 2016-10-06T06:31:45Z

src/environment.cpp

+{
+	u32 id = addParticleSpawner(exptime);
+	m_particle_spawner_attachments[id] = attached_id;
+	ServerActiveObject *obj = getActiveObject(attached_id);


merge this in the condition

But I use obj a couple lines later, won't I have to repeat getActiveObject then?

no you use it only in the condition before, then you can do this affectation in the condition

nerzhul · 2016-10-06T06:33:23Z

src/server.cpp

@@ -3163,6 +3164,7 @@ void Server::spawnParticle(const std::string &playername, v3f pos,
 	if (playername != "") {
 		Player* player = m_env->getPlayer(playername.c_str());
 		if (!player)
+


you should remove this

nerzhul · 2016-10-06T06:33:44Z

src/serverobject.cpp

@@ -99,3 +99,12 @@ bool ServerActiveObject::setWieldedItem(const ItemStack &item)
 	return false;
 }

+void ServerActiveObject::attachParticleSpawner(u32 id)


please inline this function

nerzhul · 2016-10-06T06:34:09Z

src/serverobject.cpp

+	m_attached_particle_spawners.insert(id);
+}
+
+void ServerActiveObject::detachParticleSpawner(u32 id)


same here, inline by moving definition to header

nerzhul · 2016-10-06T06:34:56Z

src/particles.cpp

+	v3f attached_offset = v3f(0,0,0);
+	if (m_attached_id != 0) {
+		ClientActiveObject *attached = env->getActiveObject(m_attached_id);
+		if (attached) {


merge condition & previous line

I use attached in the next line, won't I have to call env->getActiveObject again?

nerzhul · 2016-10-06T06:36:14Z

src/environment.cpp

+		if (sao != NULL && remove_from_object) {
+			sao->detachParticleSpawner(id);
+		}
+	} catch (...) {}


why there is a try catch here ?

The "at" method throws an exception if something isn't there. Often there won't be, since particle spawners might not be attached

please catch the corresponding exception then not all

nerzhul · 2016-10-06T06:37:54Z

src/environment.cpp

+}
+
+void ServerEnvironment::deleteParticleSpawner(u32 id,
+	bool remove_from_object)


you can move this on the upper line

raymoo · 2016-10-12T06:44:02Z

@nerzhul Made some of your requested changes, responded or asked for clarification on others.

nerzhul · 2016-10-12T06:45:50Z

src/particles.cpp

+	v3f attached_offset = v3f(0,0,0);
+	if (m_attached_id != 0) {
+		ClientActiveObject *attached = env->getActiveObject(m_attached_id);
+		if (attached)


please do the decl/affectation in the condition here too

I use attached on the next line, am I supposed to duplicate the method calls?

no because it's used in the condition, in fact with this method you do the test for the decl/assignment directly without creating a variable out of the needed scope

I didn't know you could put declarations in the condition, thanks.

raymoo · 2016-10-12T07:08:28Z

@nerzhul Ok, did the things.

raymoo · 2016-10-12T18:43:22Z

Switched from using at to find, since at and its associated exception are C++11 features.

nerzhul · 2016-10-12T19:40:03Z

src/environment.cpp

+			sao->detachParticleSpawner(id);
+		}
+	}
+	m_particle_spawner_attachments.erase(id);


i think this call should be in the condition, no?

Yes it should, however .erase() doesn't do anything when the key is not found in the array, so it doesn't matter in practice.

Right, though it didn't do any harm being outside the condition. (Pushed a new version with the erase inside the condition)

nerzhul · 2016-10-13T15:09:44Z

Okay for me

juhdanad reviewed Aug 4, 2016
View reviewed changes

raymoo force-pushed the attached_spawners branch from bb820a0 to 632398c Compare August 4, 2016 21:10

raymoo changed the title ~~(WIP) Attached Particle Spawners~~ Attached Particle Spawners Aug 4, 2016

raymoo force-pushed the attached_spawners branch 2 times, most recently from be04e1a to 990e26d Compare August 5, 2016 20:10

raymoo force-pushed the attached_spawners branch from 990e26d to 79b7a3e Compare August 5, 2016 23:09

rubenwardy added Enhancement @ Server / Client / Env. @ Script API @ Client / Audiovisuals labels Aug 16, 2016

sofar reviewed Sep 16, 2016

View reviewed changes

raymoo force-pushed the attached_spawners branch from 79b7a3e to 36472b4 Compare September 17, 2016 20:10

raymoo force-pushed the attached_spawners branch from 36472b4 to 853511e Compare September 17, 2016 20:14

raymoo force-pushed the attached_spawners branch from 853511e to 3765527 Compare October 6, 2016 06:12

sfan5 reviewed Oct 6, 2016

View reviewed changes

sfan5 added the One approval ✅ ◻️ label Oct 6, 2016

raymoo force-pushed the attached_spawners branch from 3765527 to eadda20 Compare October 6, 2016 20:22

nerzhul requested changes Oct 12, 2016

View reviewed changes

nerzhul added this to the 0.4.15 milestone Oct 12, 2016

raymoo force-pushed the attached_spawners branch from eadda20 to 41754cb Compare October 12, 2016 06:43

nerzhul reviewed Oct 12, 2016

View reviewed changes

raymoo force-pushed the attached_spawners branch from 41754cb to 9b9dc0b Compare October 12, 2016 07:08

raymoo force-pushed the attached_spawners branch 2 times, most recently from 4c70e11 to 9f60f75 Compare October 12, 2016 18:22

nerzhul reviewed Oct 12, 2016

View reviewed changes

Attached particle spawners

ce3a428

raymoo force-pushed the attached_spawners branch from 9f60f75 to ce3a428 Compare October 12, 2016 20:00

nerzhul approved these changes Oct 13, 2016

View reviewed changes

nerzhul added >= Two approvals ✅ ✅ and removed One approval ✅ ◻️ labels Oct 13, 2016

nerzhul mentioned this pull request Oct 13, 2016

Implement attached particle spawners. #567

Closed

nerzhul merged commit c9e7a27 into minetest:master Oct 13, 2016

raymoo deleted the attached_spawners branch October 13, 2016 17:30

Attached Particle Spawners #4409

Attached Particle Spawners #4409

Conversation

raymoo commented Aug 4, 2016 • edited Loading

Choose a reason for hiding this comment

juhdanad commented Aug 4, 2016

raymoo commented Aug 4, 2016

raymoo commented Aug 4, 2016

raymoo commented Aug 4, 2016 • edited Loading

Bremaweb commented Aug 5, 2016

raymoo commented Aug 5, 2016

raymoo commented Aug 5, 2016

sofar commented Aug 8, 2016

raymoo commented Aug 8, 2016

raymoo commented Aug 8, 2016 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sofar left a comment

Choose a reason for hiding this comment

raymoo commented Sep 17, 2016

raymoo commented Sep 17, 2016

raymoo commented Sep 20, 2016 • edited Loading

raymoo commented Oct 6, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sfan5 commented Oct 6, 2016

raymoo commented Oct 6, 2016

sofar commented Oct 12, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

raymoo Oct 12, 2016 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

raymoo commented Oct 12, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

raymoo commented Oct 12, 2016

raymoo commented Oct 12, 2016 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

raymoo Oct 12, 2016 • edited Loading

Choose a reason for hiding this comment

nerzhul commented Oct 13, 2016

raymoo commented Aug 4, 2016 •

edited

Loading

raymoo commented Aug 4, 2016 •

edited

Loading

raymoo commented Aug 8, 2016 •

edited

Loading

raymoo commented Sep 20, 2016 •

edited

Loading

raymoo Oct 12, 2016 •

edited

Loading

raymoo commented Oct 12, 2016 •

edited

Loading

raymoo Oct 12, 2016 •

edited

Loading