I have been (for the first time ever) trying to use the search function in Scrivener. My project is about 120,000 ish words. Every time I try to do a search, the app freezes and I have to close it. I’ve tried this on two different computers. I’m using the Windows version, on Win 7. I wanted to ask if this was a known/established bug, or just some idiosyncratic corruption thing I’m experiencing in my particular file.
This isn’t a bug I’m aware of, no. Using 1.6.1, I just ran a test on a project that’s a bit over 104,000 words in the draft (over 300,000 in the project total) and the search worked fine. Given the amount of text you’re searching, I could see it spinning a moment before bringing up the results; if your search.indexes file were corrupted or missing, it would probably take a good deal longer. (The .indexes file is a plain-text document of all the text in the project; searching through that is much speedier than going through every single RTF file in the project, which is what has to be done if the .indexes file is missing.) In that case, Scrivener would be rebuilding the index as it went, so after the initial slow response it should be much faster.
When you say that Scrivener freezes, I’m assuming you mean it gets the “Not Responding” message. Have you let it chug for a minute or two before closing it? Are you running much else on the computer at the same time that might be cutting down the available memory?
The freeze on search is a major annoyance in Scrivener, especially for users with large numbers of documents in the binder, say 5000+ . It’s entirely understandable that the index may need frequent rebuilding, but this should be accomplished in a separate process, via a standalone EXE. There’s no reason to freeze the interactive program while indexing is going on. We should be able to continue with our work until the rebuilt index is ready, and be notified thereupon.
Scrivener is a superb note-keeper as well as a writing tool. I hope developers will remember to test Scrivener on huge projects as well as book-sized ones. Some easily fixable scaling problems would become evident immediately.
Thanks for considering