Created attachment 255102 [details] [diff] [review] proposed fix We have a user who's crashing in NormalMessageEndDownload with a null m_curHdrInfo. I haven't been able to figure out how that's happening, though I've spent a lot of time pouring over logs...anyway, he's been running with this fix and he's much happier. With this fix, we won't crash, and we'll download the headers next time we sync with the folder (because we won't tell the db about the headers in the formerly crashing scenario). I suspect the problem has to do with re-running urls, but I'm not sure. I'm looking through talkback for similar stack traces, but it's taking forever.
fixed on trunk and branch