Conversation
* reformatting for markdown * reordered to have common notes before going into details
|
... Just thinking about changes I may applyI see that this reformatting and adjustment seems like a task an insstructed LLM could handle well, so my task would be "only" the instructions and checking the diff... that would at least be faster. |
|
Seems like a good idea. A couple of small hitches, though.
However, the changes/reorganization in comments about attribute bits should be done in I noticed a while back that @wmcbrine has done some work to make the manual more markdown-friendly (table display, for example). We may want to cherry-pick commits such as wmcbrine/PDCurses@aa8d315f, wmcbrine/PDCurses@574895223, wmcbrine/PDCurses@26f39dec9, wmcbrine/PDCurses@2d5ddb75b, and some of their predecessors... he did quite a bit there. I'm deeply skeptical of so-called "AI". But this is a case where I'd think you could give it a try, and the changes would be such that we could look at it and either say "oh, heck no", or grudgingly admit that it was actually okay. I don't think copyright would be an issue. |
|
Should note it's not so much a matter of "the old version was broken concerning the bits"; it's more a matter of "the old version is from a pre-Markdown universe". |
|
Cherry pick doc work sounds good - can you give that a try? |
|
Returning to this issue... the documentation in question has actually gotten a bit worse. I brought in some Markdown changes from William, and added some of my own, and added a discussion of 16-bit I started out by adopting much of what you had. I noticed a few mistakes I'd made ("bits 7-1" in cases where it should have been 7 to 0, plus some outdated text from before the return of 16-bit Text AttributesAll chtypes have bits devoted to character data, attribute data, and color Default, 64-bit chtype, both wide- and 8-bit character builds:
21 character bits (0-20), enough for full Unicode coverage As of 2025 Dec 30, WinGUI, VT, X11 and X11new, Linux framebuffer, 32-bit chtypes with wide characters (CHTYPE_32 and PDC_WIDE are #defined):
16 character bits (0-16), enough for BMP (Unicode below 64K) 32-bit chtypes with narrow characters (CHTYPE_32 #defined, PDC_WIDE is not):
8 character bits (0-7); only 8-bit charsets will work 16-bit chtypes (CHTYPE_16 #defined, must be narrow characters) :
8 character bits (0-7); only 8-bit charsets will work This is normally used only in very low-memory situations. |
There are a bunch of other changes the manual may need an update to be nicely shown as markdown (starting with C code blocks, going on with more structuring, ... but that will take some time (and is a task I may be up to if we can agree on those changes up-front).