Great! Have you cleaned your PM box yet ?

No, I have a little room but very little. E-mail is still preferable, or just discuss it here.

What's the name of your plugin? (as in, the const char* I'll use for receiver)

CSE should do.

Whoo I guessed right.

OK, that's implemented, though it seems kind of silly to send out another round of alphas just for that. I also can't really test it myself. Anyway, want me to send it to you?

That gets your name on the code wall.

Sure. Before you do, how have you set it up ? This is how I'd want it - Register to receive any messages from my end and populate your URL map at plugin load. On receiving a message of type 'CSEL' (or something else - you decide), dispatch your message. Sounds good ?

Not exactly how I have it but I can do that too.

Do I have your e-mail address?

Does this mean i can actually make use of my sub surface scattering technique through the use of detail maps in nifskope now?

Dunno. Here it is in any case - shademe.here@gmail.com

EDIT: Apparently Winrar doesn′t extract/create empty folders, I created a missing "ni" folder to the meshes folder and it cleared the problem.

I have mailed Dragoon my log, and explained the issue, but it would probably help if others would confirm if NifSE is working for them or not.

My problem is that I get a missing mesh. Does this happen to other testers?



Simply activate the test esp Dragoon included, when it runs it adds a weapon called "Daggle" into your inventory. Everything seems fine, except if you actually equip the weapon, you get a missing mesh. Or do you? This is what happens to me.

I kinda have a feeling this is only happening to me, so it would probably help to track down the cause if people confirmed this in one direction or another.


Is it possible to have other Obse plugins interfering? As for Obse version, I have the latest stable, downloaded and installed today (had beta6 before this).

This doesn't change NifSkope, so if you couldn't before, you can't now. I have no idea what you just said, so... yeah, probably not.


Sent!


E-mailed you.

In case anyone else does have this issue, my current guess (waiting for Skycaptain to get back to me) is that the Data\Meshes\ni\ folder is not present in his install for some reason - which could easily be because the archive may not include empty folders. If that's the case, everyone's going to have to create the folder manually - it's trivial, just add a new folder to the Meshes\ folder named ni\ (obviously the \ is not part of the name).

If that's not the problem... I'll have to figure something else out.

I'm reading through the docs for NifSE on the wiki, and I'm a little confused about one thing: nifs are hierarchial in structure, but the script commands don't seem to expose the hierarchy. In particular, NifGetNumChildren operates on an entire nif, rather than a particular node within one.

Does NifSE traverse the tree internally and present all the nodes as a flat list?

No, currently you only have access to the first branches, the children (and extradata) of the root node, none of the nodes from there. I'm still thinking about how to handle that one.

Sorry for the slow uptake on my end. Trying to wrap up some other scripts/features before I move onto working with this in any major way, that way I"ll have a bit more freedom to tinker. Plus of course I get the benefit of others spotting any oddities over the first few days

HeX

Re: Missing meshes


Yup, it was caused by the missing "ni" folder, it seems that at least Winrar doesn′t create empty folders when extracting the package. I edited by original post.

Creation an empty folder is sth. I bet most users won't do and then start complaining about Missing Meshes. Honestly, it's not very comfortable to let the User create it himself.
Isn't it possible to check if it exists while Starting Oblivion and create it on the fly if it doesn't exist?

Yes, it can, and if just archiving the empty folder along with everything else doesn't work, I'll have to do that. It's not hard, just not automatic, and I have to dig up the code I used to use for it (which was in v0.75 but not in any since).

But, for right now, different issue: I came home today to find my computer sitting at the Windows loading screen (with the blue bar thingy), which went endlessly. It had been left on with Firefox (and other things) open, so this was odd. I restarted, it continued to not load, I tried using Last Known Good Configuration, which didn't work, and I tried Safe Mode, which didn't even get to the loading screen and just spewed some filenames to the screen for no apparent reason.

Sooo... ugh. I have to figure out what the hell is wrong with my computer now.

EDIT: Goody. Re-installing Windows now. Thank god for separate partitions such that the mod is found on a different one.

You should be able to pull that off using the CreateDirectory and Find(First/Next)File API functions.

Heh, why not keep it simple? Just put an (empty) text file called for example "This is a NifSE mesh folder.txt" inside the "ni" folder. If you feel the need, write a sentence explaining of why this text file is there, inside the text file.

Then it′s not empty folder, problem solved, no fancy tricks needed and works 100% sure.

you took the words right out of my mouth SC.

It seems... that my harddrive has died. Entirely.

If that's the case... that's the end of NifSE. I don't have the source code anywhere, and I'm not rewriting it. ARES, too, for that matter. I don't have the money to pay for the (very expensive) data recovery services that are out there, and the harddrive is failing the apparently very critical SMART tests.

Does anyone have any recommendations for good tech help sites? I'll try my damnedest to salvage things, but...

:mellow: :unsure:

http://www.youtube.com/watch?v=WWaLxFIVX1s

Don't Hex_off, showler and skycaptain have a copy? How about in your email sent items?

They have the DLL. That's not the source. I can't update that. I was going to send them the source, but even 7-zip's highest compression setting couldn't get it under 25 MB. I could have split it into separate archives, but it just didn't seem worth it because in all likelihood they never would have even looked at it.

Hmm, I wrote all this intermittently over the course of the past hour while doing other things, then hit preview and saw your post about your hard drive being dead. Might as well post what I wrote, though.


OK, that makes sense.


I'd be wary of going too far forward without it -- you don't want to paint yourself into a corner and have to deprecate a bunch of functions because you've created hierarchy-capable replacements for them, but keep having to support them forever because mods are already relying on them.

One possible approach that comes to mind is to change the Child ID to be a more general Node ID that can represent nodes anywhere in the tree. You can define the root node to be ID #0, and traverse the tree (preorder, postorder, breadth-first, whatever) to assign numbers to all the others. All the existing commands that take a ChildID and NifID would still work in basically the same way, but could operate on any node rather than just top-level children.

What would need to change, to implement that approach, is just the way that scripts learn what ID numbers are valid. Currently you have NifGetNumChildren, which provides one end of a range of numbers whose other end is implicitly zero. That could be changed to something like NifGetChildren, which would take a NifID and a NodeID and return an array listing the IDs of the node's children. Scripts would call it once with a NodeID of zero, then again with the IDs obtained from the first call, and so on.

If you'd rather avoid returning arrays, you could use a breadth-first traversal to produce the node ID numbers, so that sibling nodes are always numbered consecutively. Then you'd just need a pair of commands that return the starting and ending IDs of a node's children.

You might want to try the old trick of putting the hard drive in the freezer for awhile -- that helps, in some cases, to make it work long enough to recover some data from it. (Put it in a sealed bag to keep moisture out, of course.)

Edit: Also, SMART warnings don't necessarily mean a completely dead drive. SMART is supposed to be able to warn you that the drive is likely to fail within the next 24 hours, so it may not actually have failed yet.

What I would do in this situation is boot a Linux liveCD and see whether the drive is detected by the system and at least remotely readable at the sector level, even if it can't be mounted. If it is, then I'd quickly buy a new drive that's at least as big as the old one, and do a byte-for-byte image copy of the old drive into a big file on the new one, and mark the new file read-only as soon as it's done. That'll preserve what's left on the drive so you don't lose any more if its condition worsens, and from there you can run various data-recovery tools on the backup.