1B E Reset 1B &k2G End-of-line Termination: CR=CR,LF=CR+LF,FF=CR+FF 1B &s1C EOL Wrap: No 1B (s4099T Primary Typeface: Courier 1B (s0P Primary Spacing: Fixed 1B (s12.706H Primary Pitch: 12.706 CPI 1B &a131M Right Margin: 131 Columns 1B &l2A Page Size: Letter 1B &l0S Simplex/Duplex: Simplex 1B &l1O Orientation: Landscape 1B &l5.4545C Vertical Motion Index: 5.4545 X .0208-inch prints 65 lines on the first page. Moving the landscape sequence to the end produces the correct results on all pages.
Created attachment 3298 [details] Test case that fails When sent to HP printer all pages contain 66 lines.
I've been able to duplicate this bug in the current PCL tarball: ghostpcl-r2871+8137. The attached file prints through line 66 on page 1 on my HP LaserJet 4500DN but GhostPCL puts line 66 on the top of the next page. The file has some extra characters, which I'll remove to make a simpler test case.
Created attachment 3331 [details] 689412.pcl Simplified test case. HP LaserJet 4500 prints all 66 lines on first page; GhostPCL prints line 66 on page 2.
Created attachment 3360 [details] a more conspicuous example This example illustrates that the cursor home position should be recalculated when VMI changes. We have not tested the side effects of changing VMI in the middle of the page. The bug is marked bountiable and I am not sure if Ron Hosler wanted to continue working it. The example was changed to use 1 pixel wide rectangles so pixel position can be measured accurately. The first line should appear 1/2" + .75 * VMI from the top of the page.
More experiments show that the cursor position is not updated if either a positioning command was given or the page was marked prior to the vmi command, we suspect other conditions would result in no movement also. An explicit start position should be given in the pcl file. We are not inclined to emulate this, closing as wontfix.