Richard Heck wrote: > Great work, Enrico. I've posted the patch to the bug report. We'll see > what the reporter has to say.
Great work indeed! > It is very strange that I can't reproduce, if this is the cause, and, > indeed, that people haven't been seeing this bug all along. This is not strange in my experience. Bugs caused by thread unsafety are extremely sensitive to timing issues, and a document showing the bug reliably on one machine may not have problems on 1000 other machines. I can even imagine that the static buffer is used by two threads at the same time without any problem, if the first one just finishes reading the last bytes of a bigger text, and the second one just converst a few characters. Also, people may have experienced this bug, but did not notice, or they noticed, but thought that a certain encoding is broken and just used a different one. Georg
