Ger Hobbelt

Ok. Do note however, that a *few* of those prefixes aren't really prefixes as they are preceded by a "\n" that's part of the same message string, so a function...

BTW: there are two important levels for the tprintf() messages: - Error - Warning and the rest is *not* prefixed (ok, a rare 'INFO:' may be observed; I'm only consistent...

@egorpugin : before I tackle this, there's also `tlog()` and `ERROR::error()` which invokes `fprintf(stderr, ...)` directly, so the question is: what would you like? I'm myself of two minds about...

Ok, if I understood that correctly, you want to keep this mostly as-is. I'll do the tprintf() et al -> log_error() et al migration as a separate pullreq later. BTW:...

Something simple. After all, tesseract (the application) doesn't need the fancy footwork, nor do many uses as a library. Currently what I've got is a blunt, hardcoded override in tprintf.cpp,...

Do you have a sample newspaper scan where this occurs? 🤔 The fundamental issue with the old code is that it circumnavigates the tesseract thresholding/cleanup process using no obvious criteria...

As I'm digging further into the catacombs of tesseract RTFC I realize that *technically* the LSTM engine is engineered to accept raw RGB input (3 channels) or failing that, greyscale...

Killing it; feeding the engine B&W pixels is not the answer. See also the @stweil comment above: https://github.com/tesseract-ocr/tesseract/pull/4111#issuecomment-1666460160

> read_params_file: Can't open /usr/local/share/tessdata/configs/ AFAICT the trailing `/` slash is indicative that the `fopen()` in there was passed this *directory* path and thus *fails* as it expects a *file*...

I see where such a thing would be handy, but when you want to do this at the macro level (rather than on the lexer rules) there's trouble using it...