Updated port of SQLite

The one I am using lives at https://sqlitebrowser.org.

Incidentally I have used it to create a small database application with about 25 views and 15 sql-queries, so I know it works well.

SQLite 3.41 was released for other platforms today, so I just want to give a quick update. This project is not abandoned, but due to various factors it might be a month before I get the RISC OS version out, so I just wanted to set some expectations there.

Regarding DB Browser for SQLite, I was hoping that somebody would put their hand up, but alas it’s all gone quiet. I’m not sure what’s involved with a “direct” port of the existing app; I see from the screenshot on its website that it uses a grid/spreadsheet type view and I’m not sure how difficult it is to produce that sort of thing (does the Toolbox have a gadget for it?).

SQLite 3.41 was released for other platforms today, so I just want to give a quick update. This project is not abandoned, but due to various factors it might be a month before I get the RISC OS version out, so I just wanted to set some expectations there.

And, more than a month later, it’s here. I expect that it’ll turn up in PackMan soon, otherwise there’s a link in the first post of this thread.

I’ve finally got around to updating this to SQLite 3.42. It seems that I actually forgot to put 3.41 in PackMan, but 3.42 should arrive in the near future.

There was a note in another thread about how big queries can make Wimp apps freeze due to the lack of Wimp_Poll calls. First of all, is there any “demand” for fixing this? I’m not convinced that many people are actually using SQLite so I’m not sure whether it’s actually a showstopper or just an observation.

In any case, if anyone has any clever ideas around how to resolve this without breaking the documented SQLite C API, then I’m all ears.

SQLite 3.43 will be released “when it’s done”. The usual update process (drop in a new unmodified sqlite3.c and merge any os_unix.c changes into os_riscos.c) has not worked with this version, so I’m going to need some time to figure out what’s going wrong. Everything compiles and appears to run correctly, but some of the automated tests are returning the wrong results. I’ll post an update once I know what the problem is.

The discussion over in the Wiresalmon thread about libraries maintained by one person and abandoned has prompted me to post a little update here. I have yet to get to the bottom of why 3.43 doesn’t work correctly, and don’t really have any further news.

I suspect I’m going to need to use the “divide in half” technique where I grab the source commit from halfway between 3.42 and 3.43 and see whether that works, then go halfway between those and so on until I figure out which commit broke it. Hopefully the problem will then become more obvious!

Of course, the next question is “when?”. I’m snowed under with “real” work at the moment so realistically we might be looking at the Christmas break.

I suspect I’m going to need to use the “divide in half” technique

Commonly known as a Binary Search
It works. I used it quite a lot before I discovered it was a known technique.
People searching for introduced bugs in the RO 5.x beta releases keep doing it – because it works.