Looking over this thread, I see no evidence that sqlite3.exe is even running when whatever form of "does not work" occurs so as to induce such a conclusion. For all I (or anyone else here) can tell from what you have posted, your problem is that, for some reason, sqlite3.exe is not even loaded and executed when your troubles occur. As to why that is, you have provide so few clues that I doubt anybody relying on your meager reporting can reasonably guess what is happening. You are showing yourself to be someone who has never learned [How to Write a Good Bug Report](https://duckduckgo.com/?t=ffhp&q=how+to+write+a+bug+report&atb=v232-1&ia=web) and refuses to do so. Here is what I know now of the "bug" reproduction steps: 1. Obtain a PC claimed by "anonymous" to fail in strange and sometimes intermittent ways when some mysterious invocation of SQLite is attempted. (This is necessary because no clue whatsoever is provided as to how things are installed or configured.) 2. Persuade "anonymous" to do whatever clicking, typing, or RPCs lead to the result that disappoints "anonymous" or his customers who may or may not be doing the same thing with the same setup. (This is necessary because no clue whatsoever is provided as to the manner in which SQLite code is supposed to be run.) 3. Observe what does or does not happen. Then, probably, interrogate "anonymous" as to what he wished would have happened. I am quite sure that if I had ever submitted reproduction steps like that in a bug report during my development career, for an employer or client, my work as a developer for them would ended soon afterward.