diff options
Diffstat (limited to 'pxl/pxfts.txt')
| -rw-r--r-- | pxl/pxfts.txt | 177 |
1 files changed, 177 insertions, 0 deletions
diff --git a/pxl/pxfts.txt b/pxl/pxfts.txt new file mode 100644 index 000000000..b765e7b04 --- /dev/null +++ b/pxl/pxfts.txt @@ -0,0 +1,177 @@ + + Copyright (C) 1996 Aladdin Enterprises. All rights reserved. + Unauthorized use, copying, and/or distribution prohibited. + + Report on Genoa PCL6 FTS and LaserJet 6MP + +This document presents the results of Aladdin's investigation of +discrepancies between the Genoa PCL6 Functional Test Suite, the published +PCL XL Feature Reference Protocol Class 1.1 specification, and the behavior +of the H-P LaserJet 6MP printer. + + Report on Genoa PCL6 FTS and LaserJet 6MP + +Introduction +============ + +In the course of validating our PCL XL interpreter using the Genoa PCL6 +Functional Test Suite (FTS), we observed a number of discrepancies that, +after careful investigation, we concluded were due to problems in the FTS +and/or the printer firmware. In this document, we present these in detail. + +This document should be read in conjunction with our separate report on +discrepancies between the specification and the firmware independent of the +Genoa FTS. + +Changes in a given revision of this document are marked with the revision +number in [brackets]. Revision history: + first issued December 6, 1996 + rev. [1] December 22, 1996 + rev. [2] February 6, 1997 + +General +======= + +We noticed that in many places, the input data contain an extra PopGS that +doesn't have a matching PushGS. The published specification says that this +is invalid (i.e., the existence of a PushGS is part of the precondition for +PopGS). Apparently the printer ignores the extra PopGS. + +Individual tests +================ + +[2] t101 - BeginSession +----------------------- + +On pages 2 and 3, in the LJ5 printouts, the characters are not distorted +(stretched vertically on page 2, compressed vertically on page 3) to match +the non-uniform UnitsPerMeasure. This is a bug in the LJ5 firmware that has +been corrected in the LJ 6MP. + +t305 - SetColorSpace +-------------------- + +The two lower rectangles in image 03 are stroked with gray, and are not +painted. The comment on this image says "The 2nd and 3rd rectangles should +be stroked and painted with black color." [1] See the entry for +SetColorSpace in the discrepancies report. + +t310 - SetCharShear + +On page 1 panel 1 the character shear setting results in font scaling +factors thar are zero. The Artifex interpreter produces no result, +the HP 4600 produces a thin one pixel line for the character. We do +not attempt to emulate this. + + +t314 - SetFillMode +------------------ + +The lower figure in all 3 images, ostensibly drawn with the non-zero winding +number rule, is clearly not drawn with this rule. [1] See the discussion of +clipping and of SetClipIntersect in the discrepancies report. + + +t321 - SetClipIntersect +----------------------- + +In the lower figure of image 01, the dark lines stop at the edges of an +invisible rectangle rather than extending all the way to the ellipse. This +is consistent with the interpretation of SetClipIntersect documented in the +discrepancies report, but not with the published specification. + +t322 - SetClipRectangle +----------------------- + +In image 03, the rectangle border is a medium gray. However, the input data +are as follows: + + NewPath + 200 @GrayLevel SetBrushSource + 120 @GrayLevel SetPenSource + 40 @PenWidth SetPenWidth + 0 @ClipRegion [1200 400 2000 2200] @BoundingBox SetClipRectangle + [1200 400 2000 2200] @BoundingBox Rectangle + 20 @GrayLevel SetPenSource + 0 2925 @Point SetCursor + 2400 0 @EndPoint LinePath + PaintPath + +The Rectangle operator is supposed to leave the path set to a rectangle, so +the final PaintPath should paint both the rectangle and the line a very dark +gray; however, as noted in the discrepancies report, Rectangle actually +clears the path. + +t325 - SetClipMode +------------------ + +The middle figure of image 03 is a solid disk. However, the input data are +as follows: + + 0 @GrayLevel SetBrushSource + [300 300 4500 2625] @BoundingBox Rectangle + PushGS + 0 @ClipMode SetClipMode + [1800 2325 3000 3525] @BoundingBox EllipsePath + [2100 2625 2700 3225] @BoundingBox EllipsePath + 0 @ClipRegion SetClipReplace + 0 @GrayLevel SetBrushSource + [300 2250 4500 3600] @BoundingBox Rectangle + +Again, if Rectangle left the path set as documented, this would draw (part +of) the rectangle as well as the disk. + +[2] t329 - SetHalftoneMethod +---------------------------- + +The DitherOrigin of the halftone is incorrectly set relative to the origin +of the default (rotated) user coordinate system rather than to the current +user coordinate system. This causes the patterns to be shifted slightly +relative to the Genoa printout. For a test example, see our separate report +on firmware discrepancies. + +t409 - Ellipse +-------------- + +Image 04 is light gray. However, it should actually be painted with an +outer stripe of light gray and an inner strip of dark gray: + + 1 @ColorSpace SetColorSpace + 255 @GrayLevel SetBrushSource + 180 @GrayLevel SetPenSource + 160 @PenWidth SetPenWidth + [300 300 2100 800] @BoundingBox RectanglePath % no-op + [300 1100 2100 1600] @BoundingBox Ellipse + 60 @GrayLevel SetPenSource + 60 @PenWidth SetPenWidth + PaintPath + +The Ellipse command should leave the path set, and the final PaintPath +should erase the interior of the ellipse (including the inner half of the +lighter stripe) and then repaint the annulus with the darker color with the +narrower pen. This is the same problem as in t322 and t325. + +t413 - Pie +---------- + +Image 05 shows the same problem as noted under t409. + +t415 - Rectangle +---------------- + +Image 02 shows the same problem as noted under t322 and t325. + +t202.bin (Protocol 2.0 fts) +--------------------------- + +Panel 01 and others show incorrect spacing when rendering downloaded +TrueType fonts. This only happens with the AGFA scaler and resolutions +lower than printer resolutions. + +t422.bin (Protocol 2.0 fts) +--------------------------- + +A raster pattern of height 1 results in a 0 area pattern which +produces an error. The error only happens if the device resolution is +less than 300 dpi. + |