never saw the message before. 

why on earth would any one handicap themselves by using qbsp256? O_o 

http://celephais.net/board/view_thread.php?id=596&start=132

Using my ConvMap utility, you can actually do the same grid-snapping in a pre-compiler step, just run:
convmap -l2 -i mapname
txqbsp mapname.out

ConvMap will just cut off the fractional part and keep the integer, essentially what qbsp256 does. See ConvMap readme for more details. 

i got your message and replied just now. 

at http://user.tninet.se/~xir870k . Minor improvements and bug fixes in several tools and engine. Please see readmes for more details.

Any comments are welcome. 

i dont know if this has been asked before, looked on this thread and i couldnt find it so here. is are there any tools to convert a doom wad file into a bsp or map file for quake1? or at least, can it be done?

thats it. 

of wad2map and wad2bsp at Google revealed some info that might be useful. 

thanks man 

I could never figure it out, what�s the exact difference between TxQBSP and TreeQBSP? 

there used to be a world of difference, but I've crossfed features between them so now they produce near identical result on any map. If in doubt use TxQBSP, that's what I use all the time.

Please note that Tx has watervis default ON, while Tree's got it OFF. Both can be switched either way via options, of course. 

aguirRe, I use Hammer and TreeQBSP, do I have any reason to switch to Tx? 

As I said, they both produce near identical result (if watervis is set the same). AFAIK, there's no particular reason to use Tree anymore, but Tx has got a nice feature to select starting hull which is very useful in big maps when hunting leaks in hulls 1/2.

Try it, you might like it ...;) 

I'm curious: what is the reason for having the two different versions? 

two versions to begin with, Tree which is made by Greg Lewis and Tx that is made by Armin Rigo (also the man behind QuArK), both building upon the work of the original tools from id software.

I just started three years ago to improve and fix the bugs that I found especially annoying then (like leak handling).

Since Tx was the only compiler at the time that I knew could handle QuArK's float maps and Tree being the most memory efficient and otherwise overall better compiler, I started to work with both tools in parallel.

Over time, they became more and more similar due to my crossfeeding of features and adding work by Tyrann and ideas of my own to the mix.

So far it's been a help for me to have two implementations that basically do the same thing but in a slightly different manner, it's like a 2nd opinion. 

are you considering merging them at some point in the future? 

point in merging them. If anything will happen it's me dropping Tree, but that's also a bit unlikely, since I like both and they serve a purpose. 

aguiRe, my comment is that I'd really prefer you to put the readme files inside the zips... :)

Having the readme also available seperately is fine and good, but by leaving it out of the zip files you're only saving a few kilobytes. I find myself having to download the zip file and then also download the txt seperately.

That's hardly a big deal and it won't kill me, but nevertheless I'd prefer if the readme were included in the zip, unless there's some compelling reason for not doing so. 

that they make for good reading 

My main excuse for not including the readmes in the zips are for easy updating of the readmes without having to upload the zips again. I'm still on dialup and that means max 33k upstream. 

I updated my flow yesterday with your new TxQBSP (and RVIS tool), and what a surprise !! Some non understood texture alignement warnings disappear with your TxQBSP new release (there was some r_CuteNodePortal warnings, pointing some textures location out of polys !!!)... I don't know what you exactly did, but it seems to work better with this new version !!

About readme file not included into zip file, I don't think it's a real problem... Downloading some kBytes files is not a big challenge, even with a standard 56k modem... it took approximatively ... 10 seconds at worst... I really it's not able to increase your phone bill, unless you are a little stingy...

Thanks a lot aguirRe..

Bye.. 

I don't know why the latest minor Tx update seems to make such an improvement for you, but I'm glad it did. 

With last TxQBSP, it appeared some r_CuteNodePortal warnings, giving a texture location out of any poly face... so it becomes hard to find from where the problem comes (regarding the read of your Q1ToolTips text file: it comes for sure from a unaligned textures..) The more when QuArK "search holes in map" tool found nothing... (in most of case, QuArK "search holes in map" tool can point up problem sometimes not detected, or not clearly detailed with TxQBSP (like poly unaligned location, or texture unaligned, or others..).. Anyway, trying the new one, these warnings disappear... so that's why I'm rather happy the problem seems to be solved... I will see later if they will reborn later (hoping not)

Bye.. 

warning has nothing to do with textures or their alignment; it's caused by brushes being off-grid or misaligned to each other.

The reason why the coordinates sometimes seem to be in empty space, is because the warning then appears in hulls 1/2, where the brushes are invisibly expanded (see also ToolTips for more info).

To find the culprit, look in the nearby area for non-axial or complex brushes and perform force-to-grid actions on them one by one until you see the warning disappear. It can be a bit cumbersome ...

Also, the QuArK built-in leak detection is not very good; it often finds a leak where there is none or fail to find one that really exists. 

Thx for these precisions... But then, why these warnings disappeared from old TxQBSP version to latest verison ??? Have you an idea ?? 

by old version, previous i.e. 1.07?

The only thing I can think of is that I know QuArK is shuffling all brush faces around within each brush each time you perform a save-load-build sequence. That can affect all kinds of things like warnings/leaks etc.

However, in both my compilers since a long time, I automatically perform a deterministic face sort to avoid these build variations. It can be disabled by using the -nosortface option. Normally I wouldn't recommend using that option.

I know GlassMan had a similar problem in GtkRadiant when building his gmsp3 map. By manipulating a light entity, a leak would appear/disappear nearby. The editor is probably shuffling the brushes/entities around and this can affect build results slightly.

Otherwise I don't really know.