Hi there,

I have a dual-boot between windows and fedora on a 20gig laptop, and since I need to use frequently both OSes, I have setup thunderbird on both, and symlinked to my inbox folder on windows for the linux based thunderbird (so that I don't use twice the HD space for emails).

The system works relatively fine, the folders work properly in both cases. However, every time I start Thunderbird on windows after I used it on linux, and vice versa, it scans through my plugins, checking for compatibility etc. It takes roughly 3-4minutes for it to start up, which is way too long. On top of that, on linux, thunderbird is impressively slow (while being quite fast on windows) I guess this has to do with the fact I use KDE, and thunderbird seems to use GKT (am I correct?). Plus Thunderbird, compared to Kmail, is very rough on the eye.

Therefore I have been looking around to see if I could tell kmail to use the thunderbird inbox folder (on my fat32 partition) and save emails and so on in the same format as thunderbird does. The most important thing being compatibility between linux and windows.

Any suggestions welcome (including advice about speeding up thunderbird if possible)

Thanks,
mtess

I just switched from Thunderbird on Windows XP to Thunderbird on Linux with KDE on a slightly faster computer with a LOT more RAM.

I noticed Thunderbird starts and runs significantly faster on Linux. I expect that is because of the extra ram. But if there were some performance problem with Thunderbird/KDE I think I would have noticed.

When I moved my massive inbox from XP to Linux, I run it through the dos2unix utility, which removed all the CR characters that are in text files on Windows and not in text files on Linux. I assumed I needed to do that. I assume your shared inbox still has all those CR characters and Thunderbird on Linux is OK with them. However, I wonder if performance is affected by those CR characters.

In some cygwin programs using large text files with Windows CR characters performance goes way down. The reason is the program is using a file I/O layer that hides the CR characters from higher I/O layers, including hiding them from seek operations. But hiding CR characters from seek operations is an incredibly inefficient thing to do. I wonder if Thunderbird turns on some similarly lame method of ignoring the CR characters in the inbox if it is running on Linux with a Windows inbox.