Vincent van Ravesteijn wrote: > I don't think this is involved in anything that has to do with fullscreen, > and fullscreen still works flawlessly.
>> generally i would be more happy if you let this code as it were then >> fixing the problems above >> and complicating the code. > Fixing the problems you mention above, just means calling "showBar(true)" > before the return calls. I don't feel that that will very much complicate a i dont speaks about number of lines. i speak about the sideeffects you need to consider because of this rather small issue. > Well, I made it somewhat more disturbing by always closing the first tab > while fixing bug #5001 in r29530. After this change, the tabbar looks a bit > buggy with a missing first tab. That's why I decided to hide the tabbar, > and it's not really needed anymore. aha, i havent seen that new behaviour yet ;) > To please you, I can either, leave it as it was (which looks buggy to me), > or close the tabs in a reverse way and then complicating the way the > session file is written or read (also in reverse), i have no real preference in the things above >or add the calls to show > the bar again after cancelling. we have already catched one bug and i still foresee different problematic scenarios - one example - you are in fullscreen and call exit. again you cancel this exit while having somehow set tabbar toggling. iirc fullscreen has no memory about the visibility states - it just toggles. now whatsover you do - either forget to show again or do automatical showing while user has set this off for fullscreen etc - and you switched to wrong state. i havent looked in the code to check these scenarios but i dont like we have to even considering such things because the visual elegance of closing tabbars. pavel
