
	IP: In Progress
	NA: Not Assigned
	DN: Done


PEX Server Bugs:
---------------

1a DN	mahesh		(phigsivp) edges not showing up
			->was a bundle bug and a text bug. see fix in #10

1b DN 	mahesh		(phigsivp) fill areas in wrong color, 
			->was a bundle bug. see fix in #10

2  DN	paula		(lines, simple, sphere) can't free pex resources
			->lut_strip_entries kept pointers in the lut table to 
			  memory which had already been freed
			->lut index in lut entry needed to be unsigned short
			  (otherwise the hash fn gave bad results for 
			  indices > 32K )
			->trashed name set ptr when copying data from 
			  renderer's pc to private pc during PEXBeginRendering

3  DN	fshu		(sphere) no mouse tracking in choice menus
			->for now redirect console output 
			  (ie: cat -u /dev/xcons &)

4  DN	ucode		draw in wrong window (expose events)
			->(mpw,2-24-90) can't reproduce, new ucode must have
			  fixed this

5  NA			(p_demo_p) lighting looks wrong on plant.goo

6  DN	paula		(sphere) tri-strips, index polys, data look funny.
			->missing dispatch table entry.

7  DN	mahesh		(simple) no text
			->sally's bug.

8  DN	dwm		better line types
			->found ucode bug.  fix is pending.
			->ucode bug fixed.

9  DN	dwm		windows get stuck when moved off screen
			->(mpw,2-24-90) can't reproduce, new server code must 
			  have fixed this

10 DN	paula		(bundles) bundle attrs don't change picture.
			->inconsistent usage of pc's asf flag. Now all code 
			  uses 0 for bundle and 1 for individual
		        ->the edge color and edge flag were not being retrieved
			  from the bundle when edge bundles were in effect
			->restoreState dirty flags updated to match renderer's
			  dirty flags
			->the blue component of the color was not being copied
			  into the color table during PEXSetTableEntries

11 DN			(p_demo_p) colors not correct when depth cueing

12 DN	dwm		toggle depth cue off server hangs
			->assume part of server fix

13 DN	paula		view munge shift down
			->the pex viewport is being taken from the drawable and
			  the drawable width and height are incorrect.  Todd is
			  going to fix this.
		        ->new server code fixed this. markers still incorrect
			  in phigs_demo_phigs due to ucode bug.

14 DN	dwm		depth cue too fast
			->was never broken.  Phils demo "dc" draws only one
			  primitive, and depth cue is applied only to vertices
			  of primitives.

15 DN	dwm		no markers
			->ucode will crash if marker extends off-screen.
			  fix in progress.

16 DN	mahesh		(sphere) crash in struct mode FAS with c&n/v,f,no edge
			->microcode bug in execute buffer fixed. Also had a
			  bug in the method of using tm{Begin,End}Capture.

17 DN			text in demo when double buffering
			->(mpw,2-15-90)can no longer reproduce

18 DN	mahesh		(p_demo_p,phigsivp) white text missing in upper rt.view
			->a combination of #10 and #1a

19 DN	fshu		cursor wrong color and shape

20 DN	mahesh		(sphere) distinguish facets can't be turned off
			->microcode bug. Reported fixed now.

21 DN	mahesh		(sphere) No edges drawn in fillareasets with edges
			->fillareaset bug corrected

22 DN	mahesh		(sphere) lots of edge pixels missing
			sphere, fas with data - the edges are stitched
			->microcode bug, Was told it would be fixed when tuning
			->was told that the best results possible were got
			->should be noted as a restriction

23 DN	mahesh		(phigsivp) Some text was going all over the place
			-> stopped after fixes #1a and #10

24 DN	mahesh		(sphere) Fillarea with data drawing crazy shapes
			->phig's PEXlib had a protocol error. Works with ours.

25 DN   paula		(bundles) used to work - now hangs while redrawing
			->problem fixed with latest ucode and server code

26 DN   paula		(pick) crashes server
			->picking now works in immediate and structure mode

27 DN   mahesh          (sphere) set color representation hangs
			->(04-05-90) can no longer reproduce

28 DN   paula           (phigsivp) doesn't work without z-buffer
			->pex now gives correct answer for HLHSR mode in get 
			enum info. need to test again on 3max with no z-buffer

29 DN   peterv          client programs hang N10 after several runs
			->cannot reproduce with latest release of server

30 DN                   server stops taking client requests during menu usage 
			->probably due to stamp pass 1 hardware
			->(mpw,2-24-90)can no longer reproduce

31 DN   paula		(sphere,smode) destroy structure returns an error
                        ->di_delete_references was not initializing error 
			return

32 NA 			(quick_updates_standalone) hangs on 2nd run in str mode
			-> program no longer exists

33 DN			(sphere) crashes on tri strip and quad mesh in str mode

34 DN			(phigs_rings_3d) too many mouse events gives XIO error
			(see #60)
			->(2-19-90,mpw) sent mail to Todd and Sam.

35 DN 			(sphere,str mode)on 2nd rendering get results from 1st
			(ie: draw sphere twice with different primitives. 2nd
			 rendering same as first)
			->this was a phigs bug

36 DN 	mahesh		execute buffer hangs n10 when executing the following:
			from elf3:/tmloc/work/extensions/demo/pex/matra2 run
			  matra -surfcolor 1 1 1 -zbuf -cue =1000x1000+0+0
			  /tmloc/work/extensions/demo/pex/matra/temp2
			  -vvol -350 300 -250 350 -1000 1000 -color
			  -rot -80 0 200 -lightenable -lightcolor 1 1 1 
			  -linecolor 1 1 1 -polygon >/dev/null 
			(note .../demo/pex/matra is immediate mode )
			->problem no longer occurs (3/27/90)

37 DN   jeff		(p_demo_p, triv) Positions skewed.
			In non double buffer mode and in MIT double buffer
			mode Positions seem skewed to the right and down.
			Pixmap double buffering seems to get the position
			correctly, but all echoing gets shifted.
			see phigs bug list - there are many examples of this
			->after the window is re-sized, pex is still getting 	
			the old drawable width and height from the vmd

38 IP   dwm		(p_demo_p) Window gets stuck resizing.
			Frequently in either double buffer mode the window
			seems to get stuck in infinite resize mode (usually
			an indication that the sizes we've recieved don't
			jive with how big the X server thinks the window is).
			A series of "osmDoRelocate PVMD of type 2 collided
			with another window" messages is reported by the
			server.  Resizing and using our reset button fixes
			this (which indicates we do have the correct full size).

39 DN   jeff		(auto_box/twotori) Ctrl C during MIT DB kills server.
  			CNTRL c during MIT double buffering gets the server
			killed on an unaligned access, or bad pointer passed
			to FREE. On a newlyrestarted server this only seems
			to happen on the second time the program is run and
			aborted with control c.
			see phigs bug list - there are many examples of this
			->phigs was sending bad request, server needs to be
			  fixed so it doesn't crash when back buffer written

40 DN			(twotori) request23 (23) error 9, in structure mode 3.
			When running in structure mode 3 the errors di_dispatch
			request23 (23) error 9, (bunch of numbers here)
			followed by the error di_dispatch request 24 (24)
			error 129, id 0 (0) (00000000) occur. 
			these errors repeat several times (10 or so).
			A second run after this error has occured sometimes
			kills the server (this behavior is less reproducible).
			->(03-23-90,mpw) cannot reproduce on latest server

41 DN   paula		system hangs drawing polylines w/ indexed vertex colors
			See /tm/qar/1
			->bug in doIndirectPolyline caused start of N10 packet
		        to be overwritten

42 DN	mahesh		light color range is wrong in pex (should be 0 to 1)
			See /tm/qar/2 and /tm/qar/3
			->Not a bug. We do not check for 0 to 1 range.

43 DN	mahesh		Only spot light spread angles 0 and 1 work correctly
			See /tm/qar/4
			->client programming error, and parameters use
			->when invocating the client wasn't good to see
			->changes

44 DN	philip		Pex incorrectly gives bad value for NURB surface
			See /tm/qar/5 (note that the error is from the server
			not Pexlib)
			->data in program violated semantics for nurb surfaces
			->this was not a bug

45 DN	dwm  		triangles from tesselation displayed in hollow quadmesh 
			see /tm/qar/6
			->fixed

46 DN	dwm		server crashes while allocating N10 lights
			reproducible by repeatedly starting koff processes 
			(see /tm/qar/7) or by using koff with test0 (see 
			/tm/qar/16)
			->light table fix checked in.

47 IP   paula		PEXGetAncestor doesnt work.  see /tm/qar/8. 
			->low priority

48 DN	dwm		(sphere,twotori) Server crashes due to memory leak .. 
		        get 'Out of storage in MALLOC' after repeated runs
			->memory leak behavior seems to have gone away;
			  however, server crash is still possible due to
			  high VM consumption by OSM.

49 DN	mahesh		(koff) errors shading indexed polygons 
		        also errors with automatic normal calculations.  
		   	see /tm/qar/9 
			-> corrected error in  automatic normal calculations

50 DN   paula	 	server crashes when window moved/resized during rend.
		      	may be a colormap bug.  see /tm/qar/10.
			->Abort rendering was not being handled correctly.
			->ROC still being done after abort.

51 DN   paula		PEXDeleteTableEntries for view entries crashes server.
			see /tm/qar/11
			->not reproducible on server built 3-28-90; tried all				->combinations listed in qar several times.
			->reproduced it on 4-10 by running many clients on
			->server before running this program .. pex was not
			->setting up the default-default table entry correctly
			->in some cases

52 DN	mahesh		initialization problems with shading
			see /tm/qar/12
			->same problem as 49

53 DN	mahesh		Inserted edges (from tesselation) being drawn when
			the ignore edge flag is true.
			->error in the way the edge flags were set when passin
			->data to the microcode

54 DN  	mahesh		polylines with color-per-vertex have drawing artifacts 
			see /tm/qar/13
			->Danny Loh corrected a microcode bug

55 DN	mahesh		attenuation problem with point light .. may not be bug
			see /tm/qar/14
			->Not a bug. There was a variation. Appears like
			->it follows the lighting equations

56 DN   paula		PEXSetTableEntries works incorrectly on certain tables
			see /tm/qar/15
			->some errors were due to unsupported data being 
			 stored in the table.  But there were several bugs
			 fixed in tables

57 DN   paula		funny picture with negative attenuation on point light
			see /tm/qar/17
			->not a bug.. attenuation should not be negative

58 DN   paula  		(avs) attribute changes don't work in structure mode 3
			to duplicate: source /bg28/phigs/setup_avs
			        avs -path /pex/avs/avs
			avs will be run from elf23:/pex/avs/kernel/flow.  do
			following in avs:
				select image
				select image processng
				select 3D mesh
				select 3d mesh controls
				set downsize to about 14 (transport buf bug)
				select read data (read stardent.x)
				select geom viewer
				select no lighting/flat shading ->  NO CHANGE 
			Also get protocol errors reported from 
			PEXDeleteElementsBetweenLabels and 
			PEXSetElementPtrAtLabel
			->phigs was not setting the element ptr to the end
			->of the structure when opening it which is what the
			->phigs protocol requires

59 DN	dwm		(avs) N10 dies when rotating lights.
			see bug 58 to see how to start up avs.
			->fix checked in

60 DN	mahesh		in robot loop option then move option gives XIO error
			(see bug # 34 also)
			->probably a PHIGS bug .. looks like problem which
			->occurs when an X call which was doing a write to an 
			->X Buffer was interrupted by another process which 
			->also made a X call to do a write to an X buffer 
			->
			->does not repeat after the PHIGs race conition
			->change

61 DN   mahesh		following program sometimes kills N10
			 fpvalve -f  valve3d.dat -w 0.8 -r
			 it will ask  "Enter nframes, angle :"
			 answer	1000 5.0
			 it asks Skip Structure edit during redraw? (1=yes):"
			 answer 0
			 it asks skip redraws? (1=yes):"
			 answer 0
			 it starts spinning object, after 20-30 seconds ctrl c
			->(04-05-90) can no longer reproduce

62 DN   paula	        PEX reports an error when unsupported attributes set
			(reported by PHIGS - elf23:/tests/quick_updates)
			Running quick_updates causes an X protocol error as 
			a result of using a interior style of PHATCH. PHATCH 
			is defined under the pex standard but not supported 
			by MPEX.  PEX should use the default PHOLLOW
			instead of yielding the protocol error. 
			->fixed - now most attributes with enumerated values 
			->map to default as in PHIGS

63 DN   paula		PEXGetEnumTypeInfo not working properly ...
			 passing ETHighlightMethod causes a protocol error
   			 (check ETEchoMethod also in case it is an MPEX err)
			 for ETlightType only get LightAmbient & LightVector-
   			 don't get LightWcsPoint & LightWcsSpot
			->enum type index needed to be interpreted unsigned.
			->pex_dd_interface.h had wrong values for lights

64 DN   paula		bug in PEXSetTableEntries, Highlight/EchoTable,index 0
			 Setting entry 0 gives protocol error even though the
			 PEX documentation states that the first index is 0 
			 for these tables and that 0 is the default
			->server thought first index was 1 (it is varies 
			->depending on table type). fixed now .. 

65 DN   dwm/paula	creating too many structures causes the server to die
			 see /tm/qar/18
			->each structure takes up about 80 bytes of memory.
			->after about 200K structures the swap space is  
			->exhausted .. this is not a reasonable thing to do.

66 DN   mahesh  	text either crashes server or yields no output
			 see /tm/qar/19
			->a bug in using the empty font if default couldn't
			->be created

67 DN   peterv 		nurb surface with edges turned on hangs server
			 see /tm/qar/20
			->fixed in ucode (was 'fixed' on 4/18, but an
			->assembler file was left lying around; everyone
			->should remove .../n10/src/3dprim/StampPgonWE.asm
			->and .../n10/src/3dprim/StampPgonWE.omax and
			->rebuild)

68 IP   dwm		server dies running robot with double buffering ..
			(reported by PHIGS)
			 setenv PHIGSdouble_buffering 1
                	 setenv PHIGS_mit_db_enable 1
			 run elf23:/tests/phigs_robot/robot
			 problem occurs on second invocation of robot with 
			 above environment settings.
			 pid 420 (Xtm) was killed on unaligned access, 
			 at pc 0x700013

			this is also reproducable with twotori.

69 DN	mahesh		protocol error incorrectly reported with db on 
			(reported by PHIGS)
			 setenv PHIGSdouble_buffering 1
                         setenv PHIGS_mit_db_enable 0
                         run elf23:/tests/fpvalve/fpvalve

			 Problem occurs on first invocation of fpvalve.
			 di_dispatch: req 24 (PEX_EndRendering), err 129 
				(RendererStateError) id 0 (0) (00000000)
			->race condition ..probably a phigs bug
			->
			-> does not repeat after the above phigs problem
			-> was corrected

70 DN	philip		server crashes when executing nurb surface
			(reported by PHIGS)
			 run elf23:/tests/sphere/sphere.sav
			 select nurb surface
			 server crashes with a seg fault in the procedure 
			 n_k_refine_kv (module k_refine.c).
    			 Philip Schneider thought the problem was caused by 
			 the fact that the rows and columns are reversed. 
			-> Client program is sending down data that violates
			-> the PEX protocol.  Violation cannot be detected
			-> except by attempting to render surface.

71 DN			pick highlighting problems with markers
			(reported by PHIGS)
			 run elf23:/tests/demo/phigs_demo_phigs

			 Pick highlighting of the markers along the sides of 
			 the upper right and lower left views don't appear to
     			 work. Pick highlighting of the markers in the upper 
			 left and lower right views is ok. 
			->phigs bug - were using bad colors for highlighting

72 DN   mahesh/dwm      avs causes microcode to go into infinite loop .. 
			peterv thinks this is an infinite loop problem in 
			the microcode similar to one he is working on.
			To reproduce the problem:
			  log into tmax23
  			  source /bg28/phigs/start_avs
  			  avs -path /pex/avs/avs
   			  Select Network Editor
   			  Pull down the following boxes into the work window 
			  and connect them:
          			read volume
          			generate colormap
          			arbitrary slicer
          			render geometry
          			display pixmap   
			  Select read volume from top level stack on left
			   and select hydrogen.dat
			  Select Geometry Viewer from at top left
			  Select Read Object at middle left
			  Select teapot.geom
			  Go into display pixmap window and rotate and/or 
			   translate teacup by holding down middle and/or 
			   right button on mouse 
			  Repeat last two steps until server goes into loop 
			   (it generally takes 8 to 10 repetitions)
			->can't repeat anymore 5/7/90
			->NOTE: I am under the impression that the reported
			->bug is that the server/ucode
			->went into an infinite loop causing the program to be 
			->waiting. However, clicking on the right/middle button
			->before releasing the middle/right button causes the
			->client to rotate continuously till a button is clicked
			->on the window again. If the latter is the reported
			->bug, then it is a client bug (missing button event)
			->osm bugs observed. which at times crash the server

73 DN   peterv		edges on indexed polys are not solid in most views
			see /tm/qar/21  and /tm/qar/30 and /tm/qar/29
			->This is no longer true for "most views".  The example
			->cited in /tm/qar/21 now works correctly (in fact it
			->may be closed already). However, edge-on views still 
			->have gaps.
			->fixed now .. see .../n10/qar/README

74 DN   paula 		referencing a predefined table entry crashes 2nd time
			see /tm/qar/22
			->fix is same as bug #51

75 DN   mahesh          Unsupported pex requests return BadRequest .. 
			should return BadImplementation.  see /tm/qar/23
			->fixed in di_dispatch.c

76 DN	mahesh		server reports error when ^C done on twotori
			(reported by PHIGS:  elf23:/tests/twotori/twotori)
			Several di_dispatch errors are reported by the 
			server after ^C is typed to stop program execution. 
			 di_dispatch: req 23 (PEX_BeginRendering), 
			  err 9 (<?>) id 9437203 (9437203) (00900013)
 			 di_dispatch: req 24 (PEX_EndRendering), 
			  err 129 (RendererStateError) id 0 (0) (00000000)
			->could not repeat it. It may be related to the child
			->process in phigs that is the cause of the problem.

77 DN   mahesh		selecting object size in phigs_demo_phigs kills N10
			(reported by PHIGS)
			->looks like N10 is getting a bad packet
			-> no longer able to repeat 5/7/90

78 DN   jeff		view skewing is back .. shows up in fpvalve
			(reported by PHIGS)
			 setenv PHIGSwstype 220
			 run elf23:/tests/fpvalve -f valve3d.dat -w 0.8 -r
			    Enter nframes, angle: 500, 5
			    Skip structure edit during redraw? (1=yes): 0
			    Skip redraws? (1=yes): 0

			Running with window size of 1.0 works. When using 
			smaller sizes (eg. -w 0.0 or -w 0.5) the picture is 
			skewed down & background color is wrong (it should be 
		 	black).

			In phigs_demo_phigs the markers in the right portion 
			of the screen get clipped when double buffering is 
			disabled.

79 DN   dwm/paula	very large structures crash server
			see /tm/qar/24
			->each polygon takes up about 32(pex element) + 
			->52(n10 compiled data) = 84 bytes of memory.
			->after about 200K polygons the swap space is  
			->exhausted .. this is not a reasonable thing to do.

80 DN   paula 		bug in GetStructureInfo
			see /tm/qar/25
			->fixed now. length of oc corrupted when storing convex
			->fill area, 2d fill area and fill area with data with
			->direct colors.

81 DN   paula		server crashes deallocating pixmap table
			see /tm/qar/26
			->fixed with new changes to luts (see bug #56)

82 IP   jeff		drawable remnants left on screen after move/resize
			see /tm/qar/27

83 DN	paula		surface doesn't show up until 2nd rendering	
			see /tm/qar/28
			->not reproducible

84 DN   peterv		hollow indexed polys & polys w/ edges are rendered wrong
			happens with hollow indexed polys when z-buffering off.
		        happens with polys with edges. (see table in qar # 36)	
			see /tm/qar/31, /tm/qar/32 and /tm/qar/36 
			->fixed in ucode  - see .../n10/qar/README
			
85 DN   paula		non-solid line types wrong; swpex and hwpex different
			1)short, non-solid lines appear solid or have drop-outs
     			2)rotation and/or scaling of short, non-solid lines 
        		  sometimes results in the entire line NOT being drawn.
     			3)SWPEX line-types of -1 to -16 are pre-defined with 
			  varying line patterns, but are defined as solid on 
			  hardware PEX.
      			4)SWPEX does not have user defined marker types and
		          hardware PEX does.
			see /tm/qar/33
			->fixed so we initialize user defined line types the
			->same as SWPEX.  Nothing to do about the fact that
			->SWPEX doesn't have user defined marker types. dyu
			->working on #1 and #2.
			->#1 and #2 now fixed in ucode. there are still dropouts
			->though .. these are very short patterned lines though. 

86 DN			Polys and Tristrips have wrong color at certain angles
			see /tm/qar/34
			-> not a bug - see .../n10/qar/README

87 DN 	mahesh		facets at near clipping plane are drawn inconsistently
			see /tm/qar/35
			->a floating point limitation. An increment of e-7
			->to the near plane, eliminates the problem

88 DN   philip		nurb surface with zero weight crashes server
			see /tm/qar/37
			->non-positive weights on control points are
			->not permitted for the time being -- clients
			->sending such curves or surfaces will
			->receive a protocol error	

89 DN   mahesh		phigsivp or fpvalve, wstype=220 gives osm error message:
		        "osmReadyVmd: bad vmd size"
			Reported by PHIGS
			->not repeatable.  now there may be phigs bug when 
			->using wstype 220

90 DN   mahesh	 	spheres (9 spheres using different prims) kills n10
			select geom = hemisphere; facet data = color; vertex
			data = color and edge = none.
			Reported by PHIGS
			-> was a microcode bug (typo error)

91 DN   paula 		default color type is wrong (pexlib?)
			see /tm/qar/38
			->the default pexlib colortype was rgbfloat which is
			->what it is supposed to be.  the crashing & hanging
			->behavior was due to a bug using indexed colors for
			->facet data in tristrips and quadmeshs .. fixed now.

92 DN			shading problem with indexed polys ..
			the specular highlights on indexed polygons 
		        (with no vertex normals) don't flat shade the 
 			facet when viewed with vector light.
			see /tm/qar/39
		        ->fixed in ucode.  it was smooth shading 
		        ->whenever the shading was geometry dependent, 
			->i.e. w/ specular, spot lites or point lights 

93 DN	philip		Possible nurb surface approximation problem.
			When approximating nurb surfaces using 
			SurfaceApproxWcsPlanarDev, translating the 
			surface around the screen drastically changes 
			the approximation resolution.  
			see /tm/qar/40
			->no such bug exists

94 DN	philip		Possible nurb surface initialization error
			After running koff twice with the same nurb 
			surface, one of the trim loops fills partially 
			in.  
			see /tm/qar/41
			->no such bug exists

95 NA		 	Severe performance degradation w/ 2 pex clients	
			Running two KOFF scripts causes the server to 
			slow down by a factor 20 as compared to one 
			KOFF script.
			see /tm/qar/42
			-> this depended on the m/c, & environment. Something
			-> funny was going on in some machines. Sometimes for
			-> no apparent reason the speeds looked normal, at
			-> most times there was a very large slow down. -mahesh

96 DN   paula		progs that run on swpex and 24bit 3da crash on 8bit 3da
			see /tm/qar/43
			->dont have any progs that allocate colormap & color
			->approx correctly.  prog in qar was trying turn on
			->hlhsr mode with no zbuffer

97 DN	mahesh		convex indexed Polys rendered wrong w/ shape not convex 
			see /tm/qar/44
			-> a bug in reading in indexed poly data.

98 DN   mahesh		resizing MIT double-buffering window crashes system
			see /tm/qar/45
			->cannot reproduce. also there was bug in client program

99 DN    paula		copy structure wrong for setsurfacecolor 
			see /tm/qar/46
			->bug in client program

100 DN	mahesh		changing vertex coord of FAS changes color of image 
			see /tm/qar/47  (probably related to bug #104)
			->not a repeatable bug. will use #104 instead.

101 DN  paula		server hangs when it exits right after drawing nurb
			see /tm/qar/48
			->this does not happen when 1) the data is copied into
			->the queue instead of using execute buffer or 2) pex
			->sync is on.  Thus this is either due to a queue 
			->problem (the data is freed before being dma'd) or
			->an execute buffer problem.  
			->fixed with latest execute buffer ucode changes (data
			->was actually freed before it was dma'd since ucode
			->was bumping rptr too soon).

102 DN  ucode		yet another edge bug
			see /tm/qar/49

103 NA			Rendering nurb surf with >= 20 trim loops hangs server
			see /tm/qar/50

104 DN	mahesh		certain combination of rd attr causes PEX to shade wrong
			see /tm/qar/51  (probably related to bug #100)
			see /tm/qar/59
			->cannot reproduce in latest server.

105 DN  paula		the object is a single color when using indexed colors 
			see /tm/qar/52  (references /tm/qar/1)
			->cannot reproduce

106 DN  paula		part of the image is clipped after resizing the window
			see /tm/qar/53
			->not a bug.. problem with client program

107 DN  kyu		large polyline has dropout if not z-buffered 
			see /tm/qar/54 

108 DN  ucode		Surf edges have large gaps when surface viewed edge-on	
			see /tm/qar/55
			->fixed checked in 5/11

109 DN  dyu 	  	specular highlights different in persp. & ortho views
			see /tm/qar/56 

110 DN	mahesh		PEXText still drawn when window is obscured
			see /tm/qar/57, see /tm/qar/62
			-> cannot reproduce on the latest server. may
			-> have been an osm problem.

111 NA			windows VERY slow on mid 3da
			see /tm/qar/58  (x server bug?)

112 DN   ucode		Long polylines get clipped too much with no zbuffering.
			see /tm/qar/60

113 DN   ucode	 	edges of obscured polys shown when surface edges on
			see /tm/qar/61     
	clarification for qar:

	this was seen in a situation where we had a quad-mesh or nurbs or some
	kind of hump-shaped thing, and it was mostly obscurred by a polygon.
	the edges we're talking about are the edges around the tesselation-
	triangles or quadralaterals or whatever.  the onesthat get drawn on top
	are ones from a part of the surface that the eye-to-object ray is
	_tangent_ to.

114 DN   paula		PEXGetEnumInfo crashes server if enum type has no values
			see /tm/qar/63
                        ->for enum types with no supported values the server
                        ->needs to return 1 word of data -- the count, set to 0.
                        ->not doing this caused pexlib to crash.

115 DN   mahesh		PEX draws to screen after window goes away
			see /tm/qar/64  (probably the same as bug #110)
			->cannot reproduce.

116 NA			X protocal error generated by Euclid 
			see /tm/qar/65 (probably an application bug)
			->phigs race condition? hard to reproduce. glenn is
			->going to see if it ever happens with wstype = 220.

117 DN  paula		circlar structure crash
			see /tm/qar/66
			->not a bug. this is as designed.

118 DN  paula		get wrong results when inquiring default pex pc.
			see /tm/qar/67
			->not a bug. problem with client program.

119 DN	mahesh		koff gets wrong colors on lo 3d
			see /tm/qar/68	
			->not a bug. The hardware does not allow color
			->space  interpolation on an 8 bit frame buffer

120 DN  paula		sometimes text and atext color is white 
			see /tm/qar/69, /tm/qar/70
			The text color is white if only text is defined. If 
			some other primitive type is defined (ie: polylines), 
			the text is rendered with the color of other 
			primitives.  To see this do the following -
				>testlut -li 3 -fi 2
			->this is ucode bug .. they are not using the color
			->per prim but rather the line color in the stateblock
			->fixed in ucode now.

121 NA			server crashes when some nurb surface calls are made
			see /tm/qar/71

122 DN   mahesh		shading  wrong after polys w/ no normals, type unknown	
			see /tm/qar/72

123 DN	mahesh	        error reported due to lost packets 
			see /tm/qar/73
			->pex was not returning the sequence # for the reply
			->packet

124 DN  ucode		surface edges have gaps when viewed edge on (once again)
			see /tm/qar/74

125 IP  ucode		sometimes markers cause bad stamp packet to be sent.
			see /tm/qar/75



	

Phigs bugs:
-----------

1  The phigs code assumes that when wstype 221 is used that z buffering
   should be enabled.  It could be that the 3max doesn't include the
   z buffer hardware.  This problem can be seen by running the phigs_ivp
   program on a 3max with no z buffer and PHIGSwtype set to 221

2  Phigs generates an error if all registered surface edges are not supported.
   3max only supports solid surface edges.

3  (sphere) Immediate mode using the primitive Fill Area Set w/data and per 
   edge flags causes a segmentation fault in the code to draw the primitive
   down in the I3d. 


Phigs demo / test bugs:
------------------------

1  In the bundles program phigs will not send interior style empty if it
   is chosen for the interior bundle entry data.

2  In the bundles program, the valuator which comes up when choosing line width
   or marker size or ... often has the wrong title (it says " color blue ") and 
   phigs prints out an error message

3 (sphere) Attempting to set the ambient light color causes a segflt.


PEXLIB bugs
-----------

1  ffanim -egg -efile eggblue.bin -w 222 overflows the transport buffer.
