Hi Kristaps, Kristaps Dzonsons wrote on Tue, May 03, 2016 at 09:24:49PM +0200: > Ingo Schwarze wrote: >> Ted Unangst wrote:
>>> We should provide documentation for at least the programmatic >>> aspects of it. (The full extent of SQL language documentation >>> may be beyond the scope of base, but C APIs are not.) >> I doubt it's worth wasting time on. > I can't speak for sqlite and inclusion (what is this `less bloated' > alternative you have in mind, by the way?), Probably from scratch, likely based on ohash(3), without using any SQL paradigm, but just simple table lookup sped up by elementary use of hashing. > but that time has already been wasted... > https://github.com/kristapsdz/sqlite2mdoc Sorry for being terse to the point of incomprehensibility. I didn't mean to say sqlite2mdoc(1) was a waste of time, it was not. Quite to the contrary, it fits quite well into the big picture of answering the question: How can transitioning of existing high-qualitity documentation from some other format to mdoc(7) be facilitated? That problem arises again and again for many source formats, and more or less structureless HTML is likely one of the more widespread variants. Besides, each solution for one particular source format makes it more likely to come closer to a more general solution - though great care is needed to not make such a more general solution bloated; it can probably be partly automated at best and will still require some code customization for each individual case even when the solution finally grows mature. All i meant to say is that spending system integration work to put sqlite3(3) manuals into OpenBSD base would likely be wasted. Yours, Ingo
