8ca8370426
Pos895Test is a console app that drives a POS895 over TCP through the existing EpsonPrinter class, to answer the first question in the feasibility notes: does the printer accept our ESC/POS at all. It builds the whole job as one byte array and sends that exact array, so the hex listing it prints is the wire content by construction rather than a reconstruction that can drift. Current job prints a "|---" column ruler at 48/width characters per row for width magnifications 1x, 2x and 3x at double height, so every row should end at the same right margin. Confirmed against the hardware so far: - DLE EOT 1..4 are answered and follow the Epson bit layout exactly, with the 0x12 fixed bits set on every reply, so EpsonPrinterStatus parses this printer unmodified. - Font A is 48 columns on the 576-dot line, as the manuals claim. - GS ! magnification works, verified at width 1x/2x/3x and height 2x. - ESC ! 0x00 selects Font A. Not yet confirmed: Font B and its claimed 64 columns, the GS I 1 model byte, ESC d n feed (this firmware appears to ignore it, hence the literal LF bytes before the cut), code page 18 being CP852, and raster graphics. Deliberately does not call GS I 1: the model byte is still an open unknown and a smoke test must not be able to fail on one. All settings are constants at the top of Program.cs; there are no arguments. Claude-Session: https://claude.ai/code/session_01E6N3P4nSZLtp4BC2LXVref
