Changing the state of an alias

Posted by Dakine on Tue 14 Apr 2015 05:51 PM — 7 posts, 26,685 views.

Dakine Tue 14 Apr 2015 05:51 PM #0

Hi,

MUD: Aaardwolf
Very new to Lua and apologies if this is obvious. I may not be searching for the right terms!

I would like to change certain aliases on reaching a certain level.

As an example, I have a primary spell alias which I then put into one of the numberpad keys.

Example Alias:


<alias
   name="mageprimcast"
   match="^csc(?:[ ]+(.*))?$"
   enabled="y"
   regexp="y"
   send_to="10"
   sequence="100"
  >
  <send>c 346 %1</send>
  </alias>

Currently I have to manually change this alias on obtaining a more powerful spell. This gets a bit tedious after awhile.

I think I can use a trigger like this:


<trigger
   enabled="y"
   match="you reach level {*)."
   send_to="10"
   sequence="100"
  >
  <send>if %1 == 46 then
   SetAliasOption ("mageprimcast", "send", "c 347 %1")
end --if
</send>
</trigger>

However, I've not been successful with the above trigger during testing.

These are contained in a plugin. Two questions.

1) Is this the correct method of changing the send info of an alias? Is there a better way?

2) How do I maintain this new value when I close the and reload the mushclient?

thanks
Dak

Nick Gammon Australia Forum Administrator Tue 14 Apr 2015 08:06 PM #1


if %1 == 46 then
   SetAliasOption ("mageprimcast", "send", "c 347 %1")

Your method is pretty good, however the %1 will be 46, not %1 any more. Try this:


if %1 == 46 then
   SetAliasOption ("mageprimcast", "send", "c 347 %%1")

The %%1 becomes %1 after the alias is expanded.

Quote:

2) How do I maintain this new value when I close the and reload the mushclient?

You can't directly change the aliases in a plugin, because plugins are read-only (they can be shared between multiple world files).

The simplest way is probably to "remember" alias changes in a table, serialize that table, and then on plugin load re-apply all the changes.

See: http://www.gammon.com.au/forum/?id=4960

For example:


if %1 == 46 then
   SetAliasOption ("mageprimcast", "send", "c 347 %%1")
   aliasChanges = aliasChanges or { }  -- ensure table exists
   aliasChanges ["mageprimcast"] = "c 347 %%1" -- new value for this one
end

In reply #6 of the link above ( http://www.gammon.com.au/forum/?id=4960&reply=6#reply6 ) you can see how to change your plugin to save that table (change the variable name). eg.


require "serialize"  -- needed to serialize table to string

aliasChanges = {}  -- ensure table exists, if not loaded from variable

function OnPluginSaveState ()
  SetVariable ("aliasChanges", 
               "aliasChanges = " .. serialize.save_simple (aliasChanges))
end -- function OnPluginSaveState

Then in the plugin install put the aliases back:


function OnPluginInstall ()
  assert (loadstring (GetVariable ("aliasChanges") or "")) ()

  -- put aliases back to remember the changes
  for k, v in pairs (aliasChanges) do
    SetAliasOption (k, "send", v)  -- change this alias "k" (key) to send message "v" (value)
  end -- for

end -- function OnPluginInstall

Make sure the plugin is set to "save state". ie. at the start of the plugin:


<plugin
   name="What_Ever"
...
   save_state="y"
...
   >

Re-reading your post it isn't clear how the 346 becomes 347. Should that be another argument to the alias, or is that a typo?

Dakine Tue 14 Apr 2015 08:43 PM #2

Hi Nick,

Thanks for such a quick response and comprehensive reply.

It's a lot to take in but I think I understand the concept. At least as an overview and can see how it would provide the solution. It's something to work towards.

The c 347 wasn't a typo. The c 346 was the current (spell) value of the send of the alias, which I manually go in and edit. The c 347 was the next (spell) value I would edit. Very clunky!

Thanks

Nick Gammon Australia Forum Administrator Wed 15 Apr 2015 06:42 AM #3

In that case the "alias that changes an alias" should probably also change the spell number.

Dakine Thu 16 Apr 2015 06:29 PM #4

I'm making progress..It's simpler than I first thought it was after reviewing the code.

However, I have a question if the following still applies.


local t = {}
setfenv (assert (loadstring (aliasChanges)), t) ()

access it by:


t.aliasChanges

See: http://www.gammon.com.au/forum/?id=4960&reply=7#reply7

I'm always one for protecting ones own code and would like to use this if it is still relevant.

My question: does the OnPluginSaveState code below need change by adding the t?


function OnPluginSaveState ()
  SetVariable ("aliasChanges", 
               "aliasChanges = " .. serialize.save_simple (t.aliasChanges))
end -- function OnPluginSaveState

I kept moving where the t should be until I was slightly confused.

addendum - I try to keep code quite clean and palm off code to other .lua files. On further testing I'm unable to update the unserialized table values from these .lua files only from the exampleplugin.xml file.

Is there a way to do this or only from the exampleplugin.xml file? I thought perhaps it was the order in which the requires were loaded, but not found a solution yet.

Thanks

Nick Gammon Australia Forum Administrator Thu 16 Apr 2015 07:52 PM #5

Let's see. If you do a normal serialize, you get what amounts to assignment statements like this:


mobs = {}  -- create mobs table
mobs.kobold = {   name = 'Foo'  }
mobs.worm = {   name = 'Bar'  }

require "serialize"
print ((serialize.save ("mobs")))

Output:


mobs = {}
  mobs.kobold = {}
    mobs.kobold.name = "Foo"
  mobs.worm = {}
    mobs.worm.name = "Bar"

So when you do a "loadstring" it is executing those statements, repopulating your mobs table (in global namespace).

If you want to leave the global namespace "pure" you could serialize into a table like this:


mobs = {}  -- create mobs table
mobs.kobold = {   name = 'Foo'  }
mobs.worm = {   name = 'Bar'  }

saved_mobs = serialize.save ("mobs")

local mud_stuff = {}
setfenv (assert (loadstring (saved_mobs)),  mud_stuff) ()

require "tprint"
tprint (mud_stuff)

Output:


"mobs":
  "kobold":
    "name"="Foo"
  "worm":
    "name"="Bar"

This means that your recovered table is not in mud_stuff.mobs, not just mobs. This might be confusing because you are now a level down from when you did the serializing.

However if you use serialize.save_simple then you can save from a lower-level table like this:


mud_stuff = {}

mud_stuff.mobs = {}  -- create mobs table
mud_stuff.mobs.kobold = {   name = 'Foo'  }
mud_stuff.mobs.worm = {   name = 'Bar'  }

saved_mobs = serialize.save_simple (mud_stuff.mobs)

print ("Saved mobs ...")
print (saved_mobs)

local mud_stuff = {}
setfenv (assert (loadstring ("mobs = " .. saved_mobs)),  mud_stuff) ()

print ("Reloaded mobs ...")
require "tprint"
tprint (mud_stuff)

Output:


Saved mobs ...
{
  kobold = {
    name = "Foo",
    },
  worm = {
    name = "Bar",
    },
  }
Reloaded mobs ...
"mobs":
  "kobold":
    "name"="Foo"
  "worm":
    "name"="Bar"

serialize.save_simple doesn't know the name of the "root" table (it only has its address not its name) so it isn't in the saved data. That's why we have to put the variable assignment into the loadstring function call.

If you are just writing for yourself you probably don't need this extra level of complexity, because presumably you trust your own code.

Dakine Thu 16 Apr 2015 10:31 PM #6

I've got the beginnings of my code working thanks to your replies.

As you mention, it's probably best to keep it simple to start with. At least while learning the fundamentals.

Dak