|
From: Solomon P. <pi...@sh...> - 2025-03-06 13:27:01
|
The short version: It's being held up by release blocker bugs.
The current holdup is a nasty margin-related bug that manifests itself
as garbled printouts when producing 2x6" strips, but only with certian
dyesub models and even then, depending on how the printjob is submitted
to CUPS (eg as a pdf vs jpeg) it doesn't occur.
Essentially, the image data is submitted in a portrait orientiation, but
needs to be rotated as the printer needs landscape oriented data, but
under some circumstances (where there is a non-zero non-printable margin
on the LEFT side) this transformation isn't being performed properly,
but the data is read out as if it has.
I was able to work around the problem insofar as the rotation now always
occurs, but I noticed the first colunm (or four) of the output data is
still being garbled. There is clearly still a boundary off-by-one error
going on, and I strongly suspect it has been here for a _long_ time, but
is not normally visible as it falls outside of page margins.
While trying to hunt this one down, I've been cleaning up many of the
spurious build warnings to hopefully expose stuff that matters, but
managed to introduce outright crashes in the injket dithering code along
the way. Turns out there are some real bugs lurking there too, likely
of the "cancel each other out" variety.
(Ah, the joy of technical debt in a 25-year-old codebase..)
- Solomon
--
Solomon Peachy pizza at shaftnet dot org (email&xmpp)
@pizza:shaftnet dot org (matrix)
Dowling Park, FL speachy (libera.chat)
|