hex11 Posted November 22, 2011 I was playing around with SLIGE and noticed something strange about some textures. The map was generated the usual way: $ slige -doom2 -levels 7; bsp slige.out And here's the random seed: $ strings slige.out |head -2 PWAD2 SLIGE (490) 1162831338 MAP01 was fine except for a strange looking exit switch (EXITSWIW texture doesn't look at all like a switch...) MAP02 had all kinds of tutti-fruti all over the place, like this: Those are GRAYALT midtextures on 1s sidedefs, but since the GRAYALT in Freedoom has two transparent patches in upper-left and lower-right corners, you get the tutti fruti (in the original Doom II IWAD, all four pieces are solid patches). Also, it's a wood crate texture instead of some concrete material like the other GRAY* textures... 0 Share this post Link to post
GreyGhost Posted November 23, 2011 I can't find those textures, which IWAD are you using? 0 Share this post Link to post
hex11 Posted November 23, 2011 It's the doom2.wad from v0.7: MD5 (freedoom-iwad-0.7/doom2.wad) = 21ea277fa5612267eb7985493b33150e But you're right, those aren't even in the IWAD. They only show up when I also load the slige.out PWAD into Yadex, or run the resulting output from bsp with chocolate-doom (or prboom). Chocolate Doom also shows some warnings on startup: R_Init: Init DOOM refresh daemon - [...........R_GenerateLookup: column without a patch (GRIDRED) R_GenerateLookup: column without a patch (GRIDBLUE@) R_GenerateLookup: column without a patch (GRIDYLW) R_GenerateLookup: column without a patch (GRIDPURP@) R_GenerateLookup: column without a patch (GRIDWHT2@) R_GenerateLookup: column without a patch (EXITSWIT@) R_GenerateLookup: column without a patch (WOODLITE) So I maybe slige is doing something wrong... Okay, after RTFM, it looks like slige inserts some new (custom) textures into the generated PWAD, unless you use the -nocustom option. But the weird thing is, those new textures don't cause any problems when using the real Doom II IWAD (no warnings at startup, no tutti fruti, etc.) In fact, here's what the same MAP02 looks like in this case: This is with the same exact PWAD that I loaded before with the Freedoom IWAD. But in this case, none of the four patches in that GREYALT texture are transparent, and the EXITSWIW also looks sane. 0 Share this post Link to post
hex11 Posted November 23, 2011 Oohhh, I remember... This thing: http://www.doomworld.com/vb/freedoom/55932-weird-patch-texture-bug/ Slige is suffering from lumpy PNAMES syndrome. Whatever that means, I'm not a WAD doctor. ;) But prepending cct.wad to the -file argument does the trick. Btw, this is a great slige mapset, one of the better ones I've played so far. Anybody can generate it themselves with slige if they use the same seed number I posted, but here's the file anyway: http://www.sendspace.com/file/0inuy5 It's only 7 levels, with a boss arena on MAP07. 0 Share this post Link to post
andrewj Posted November 23, 2011 The textures in FreeDoom are not guaranteed to be composed in the same way as the IWAD(s), and this can cause issues with any pwad which adds custom textures and assumes the patches are certain sizes. 0 Share this post Link to post
Catoptromancy Posted November 23, 2011 It is actually a pwad with a texture1 and no pnames lmp. Slige creates a texture1 without a pnames lmp? 0 Share this post Link to post
hex11 Posted November 23, 2011 Both slige.out and the final tmp.wad (from bsp) have only a TEXTURE2 lump and that's it. No TEXTURE1 and no PNAMES. 0 Share this post Link to post
Gez Posted November 23, 2011 Catoptromancy said:It is actually a pwad with a texture1 and no pnames lmp. Slige creates a texture1 without a pnames lmp? It's perfectly possible. That means the IWAD's PNAMES is used. 0 Share this post Link to post
Jon Posted November 26, 2011 andrewj said:The textures in FreeDoom are not guaranteed to be composed in the same way as the IWAD(s), and this can cause issues with any pwad which adds custom textures and assumes the patches are certain sizes. But this is a bug in freedoom, surely. The initial aim was for the TEXTURES lump to be nigh-identical, or at least a superset of the union of DOOM/DOOM2, and for patches to have the same properties (x/y size; whether transparent or not) 0 Share this post Link to post
andrewj Posted November 27, 2011 Jon said:But this is a bug in freedoom, surely. The initial aim was for the TEXTURES lump to be nigh-identical, or at least a superset of the union of DOOM/DOOM2, and for patches to have the same properties (x/y size; whether transparent or not) If strong compability is a goal, then yeah it's a bug. However I remember some talk about trying to avoid any potential copyright issue with the arrangement of the patches in textures (a bit absurb imho). 0 Share this post Link to post
Catoptromancy Posted November 27, 2011 Freedoom has the pnames lmp in a different order than id's. Only a about dozen wads on the archives are affected by this and apparently slige wads. http://www.doomworld.com/vb/showthread.php?s=&postid=984168#post984168 Need to rescript build system so that it order the pnames exactly like id's. http://www.doomworld.com/vb/showthread.php?s=&postid=913439#post913439 0 Share this post Link to post