On Thu, Feb 17, 2011 at 12:43:39PM +0100, Manuel Pégourié-Gonnard wrote: > Le 16/02/2011 16:26, Khaled Hosny a écrit : > > Fixed on git few days ago, I think it deserves a release. > > I'm glad you fixed it and are about to release the fixed version. > > May I suggest that you add to the code a global counter of how many times > database (re-)generation was attempted and stop trying if it reaches, say 3. > This way, future similar bugs would not trigger an endless loop. > (Interestingly > enough, on my Linux machine, I rapidly got a stack overflow, while on the > windows (virtual) box, I had to kill the process after a few minutes.)
The loop itself was the bug, it shouldn't be looping at all (I was simply calling a function1 that calls function2 but function2 were in turn calling function1 again :) The new version already uploaded to CTAN. > > I wounder why > > everyone is encountering it just now (this is like the fifth report) > > though the bug have been there for ages. > > > The bug happens only when there is no name database at all, so it doesn't > happen > if you used a previous version of luaotfload. It happens only on new > installations and when you remove all cache files. I know, it is just interesting how many people run into this just now though it have been laying there for months. > Maybe a lot of people are either trying it for the first time or using a new > installation? May be. Regards, Khaled -- Khaled Hosny Egyptian
