@LordDammus

Intel Core 2 Quad CPU Q9650@3,84 GHz


@Locksley
It's a counter, higher number is better

Perhaps in the "python.exe Script\niftoster.py ..." call are missing some ' or " for the log- part and as a result the spaces in your folder- path make some problems ("Program Files (x86)").

Yes I used the Wrye Python installer and have python installed.
Yes I have PyFFI installed. Version 2.0.0
I have 0 issues running Wrye Bash and I can run PyFFI on individual files by right clicking on a nif.
So, I must have your tool installed incorrectly.
Path to Python: C:\Program Files\python25
Path to PyFFI: C:\Program Files\pyffi

@aellis
Is it possible, that you install python under it's standard folder "C:\python25"?
I think the spaces in the folder- name are a problem, because if I understand the source code right (or not), then the temporary log is saved in your case (setting python_folder = "C:\Program Files\python25") under the "C:\Program Files\python25"- folder with "the evil space".

1) Is niftoaster.py located in "python root"\Scripts\
2) In the ini file is there a "\" at the end of python_folder
3) can you post the contents of tmp_bat_a.bat


(The next version does a lot more checking before running and giving more useful error messages).

-dan

I can't wait for Alpha 4 Release

Thanks for all your hard work kandiedan!

1) Yes
2) It turns out I set the path statement incoorectly. I forgot the "\" after C:
However, after I corrected it, the tool still burps on me. But this time with a different error message.
Exactly as it appears in the error dialog box:

"ERROR 0x01C:\Documents and Settings\Andrew\Desktop\PyFFI Testfolder\altarcloth01.nif"

The current ini is posted below. I changed all the "skips" to 0 for testing:

python_folder 	= "C\:Program Files\python25\"skip_armor	= 0skip_weapons	= 0skip_characters	= 0skip_creatures	= 0skip_fx		= 0skip_clothes	= 0size_test	= .97num_core	= 2

3) tmp_bat_a.bat:

@echo PROC A 1C\cd C\:Program Files\python25

As to the other post which suggested that "evil spaces" were at work here, wouldn't Wrye BASH have issues if that were the case?

Well I can almost telling you what is happening.
My program is trying to copy the source file "C:\Documents and Settings\Andrew\Desktop\PyFFI Testfolder\altarcloth01.nif"
to destination "C\:Program Files\python25\temp_a.nif"

That command is returning an error.
I would check and see if temp_a.nif is in your python folder.
If not, I would guess it is a permissions issues with windows when trying to write a file into the "Program Files" folder, but I don't know that much about windows.

The only suggestion I have is to reinstall python to a different directory or wait until I get a version that uses the temp directory for all the file manipulation.

Sorry that this has not worked out for and thanks for the feedback.

-dan

Not the hottest pc-crack here, but shouldn?t it be c:\program files instead of c\: ?

Hmm didn't even notice that.
That would explain why it can't copy the file

aellis - Can you post the first few lines from file_list.txt

Jeez, I need new glasses. I corrected the ini error in the parh statement.

Still no joy. Only this time error 0x05 showed up again.

First 5 lines of file_list.txt:

C:\Documents and Settings\Andrew\Desktop\PyFFI TestFolder\altarcloth01.nifC:\Documents and Settings\Andrew\Desktop\PyFFI TestFolder\anvilaltar01.nifC:\Documents and Settings\Andrew\Desktop\PyFFI TestFolder\anvilbasemantdoorframe01.nifC:\Documents and Settings\Andrew\Desktop\PyFFI TestFolder\anvilchaindollholder01.nifC:\Documents and Settings\Andrew\Desktop\PyFFI TestFolder\anvilcirclebench01.nif

It seemed to have barfed on the second file - anvilaltar01.nif.
files_skipped.txt and output_log.txt are still 0 bytes.
tmp_bat_a.bat has changed from the previous posting:

@echo PROC A 1C:cd C:\Program Files\python25\python.exe Scripts\niftoaster.py optimize --noninteractive "C:\Program Files\python25\temp_a.nif" > C:\Program Files\python25\tmp_a.log

I did not find a "tmp_a.log" in the python25 directory.
However, I did find a tmp_a.nif and a tmp_b.nif.
AFAIK, XP does not control Program Files with UAC as Vista does. I don't think that is where the problem lies.

Hmmm, I just had a thought. Wrye Bash is open. Maybe there is some issue with 2 python apps running at the same time? I'm going to close it and try again.
Nope. That wasn't it.


Edited to add addtional information

python.exe Scripts\niftoaster.py optimize --noninteractive "C:\Program Files\python25\temp_a.nif" > C:\Program Files\python25\tmp_a.log

python.exe Scripts\niftoaster.py optimize --noninteractive "C:\Program Files\python25\temp_a.nif" > "C:\Program Files\python25\tmp_a.log"

Perhaps this " is missing after the >

I tried both this tool and the other one also released recently on the forums.
I have been using the quad processor version and it's basically worked a treat, though eventually I switched to using just three cores so I could do other things without my processor lagging on the load.
I've optimised my Oblivion - Meshes.bsa contents, rebuilt the bsa and loaded up a few saves to check they all worked ok.

Fab!

As a word for any future design, the other tool released had a very handy feature that allowed the user to stop and resume jobs, which was very handy indeed, but the multi-thread support makes this version a winner for me with a quad-core.

Thanks a lot!

Vac

just about to run this for the first time and noticed a few pages back that you got the ini defaults from the cs wiki, however it says that it doesnt work on helmets. Yet the ini only specifies armour, so if its included under armour is that ok?

and finally does anyone know if this is fine to do with fallout 3 as well?

thanks

it should work with any NetImmerse using game (and definitely will work with fallout)

helmets are fine unless they have an egm, the egm refers to the specific order of the vertices/faces/triangle (not sure which of those it refers to but w/e) so this checks for egms and if they exist skips the nifs associated with them.
Pacific Morrowind

ok thanks very much
and therefore if your looking for more testing or new ideas etc a post in the fallout forums may be a good idea as well

EDIT: just wondering when packing up a bsa how does the compression affect performance? as i notice the vanilla meshes are around 50% compressed, so would it help to simply unpack them in the data folder, an uncompressed bsa or maximum compression for the sake of space saving?

any help greatly appreciated

for most people using an uncompressed bsa is much better for performance. However if your harddrive is slower than your cpu it is the opposite and you should fully compress it.
Pacific Mororwind

right well im guessing my cpu will outdo my hard drive but how do you measure?

most peoples cpu is not as good as the hd, unless you got an ancient harddrive, the only way I know is testing with compressed/uncompressed but I've never really looked into it.
Pacific Morrowind

ok thanks i may go with the uncompressed then and see how it goes

I finally got this to work.
I had to reinstall Python using C:\python25 as the install directory.
Maybe my problems were due to the tool not looking for where python was installed?
Does the tool look for the relative path to python? Or is it coded to expect python to be in C:\python25?

This worked using all four cores. Unfortunately, I didn't use a timer. But using all 4 cores was noticeably quicker on the test folder.

One thing though, tmp_bat_c.bat and tmp_bat_d.bat are not being deleted during the final end routine and cleanup.

Thanks for a new way to help clean up Oblivion.

Any new developement for release?

I've noticed the Quad version doesn't skip "helmet" or "helm" or "hood" nifs.
It does skip those that have an associated egm.

Otherwise, it appears to be working as touted.

Ok Ok. I posted my latest version Alpha four. I was hoping to do more testing. I believe the problems people were have was because I was not quoting the piped log output filename. if someone had whitespace in their python or pyffi path that would cause the error. checking this requires reinstalling python/pyffi a couple times and I have been underinspired to do so. So I posted my latest work that I think *should* fix the problem but without adequte testing. Use at your own discretion.



yea the idea was that it would only skip files with an egm.
The latest release provides for a lot more configuration for skipping files with the "skip_list" entry in the ini

See the changelog in the top entry for complete list of new features

-dan

I am looking forward to trying this out when there is a higher-than-alpha version.


Keep up the hard work!
- Tomlong75210

never going to get there.

Even if I write the perfect utility there is no way to tell if a pyffied mesh is good or not (other than comparing the two with nifscope), you will never get perfect results.
The utility will either work for you or not (and hopefully give you a decent error message about why it didn't run), it will not introduce any new "errors" into your nif files although pyffi very well might.

The real trick is knowing what files you want to pyffi and what you don't.
Based on my experiences, pyffi should be used by the modder or other experienced user. The results should be checked and then uploaded for the average user to download.

-dan

I hope more modders PyFFI there meshes then. Unfortunately, many great quests do not have PyFFIed meshes. Hopefully, someone can get permission to do that for some of them.