On Mon, Feb 7, 2011 at 2:52 AM, ~suv <suv-sf@...58...> wrote:
On 7/2/11 11:36, Nicolas Dufour wrote:
De : ~suv <suv-sf@...58...> Any comments from Windows users (Johan, JazzyNico, Uwesch,...)? The issue is limited to the Windows builds afaict [1] and its impact on usability seems difficult to evaluate without actually using an affected build.
I'm afraid we're going to have the same kind of reaction we had with the persistent markers with 0.48 on GNU/Linux. The problem with UI issues is that even when they are not severe (here it's just flickering, and doesn't prevent Inkscape from working), they are more visible than PDF export bugs. Frankly (if it's not too late), I'd rather keep the old devlibs for 0.48.1 and start using the new one (and try to find a fix or a workaround) in 0.48.2.
As far as I know, the 0.48.1-1 packages already uploaded to sf.net (installer, archives, debug and portable) have been built with the current devlibs (cairo 1.10.2).
Not sure, but it looks like we'll have to bear with the user reactions until the next bug-fix release is ready (0.48.2). Or replace the uploaded files with new packages (0.48.1-2), built with the old devlibs (cairo 1.8.8). The links to the new files have already been spread quickly via twitter et al. - replacing the files now could cause some confusion for users. OTOH the release wasn't yet officially announced.
I see minimal harm with what has spread via twitter that is not official. :) If people file bug reports about it prior to official announcement and say they found out via AwesomeSoftware2011 on twitter, we just need to tell them to watch trusted people (myself, joncruz, prokoudine, or any number of others) for *accurate* information about releases. Confused users that trust third-party sources instead of trusted/official sources are the least of my worries and imho kinda deserve the pain they subject themselves to (like people who get malware infected machines because they pirate via P2P and such).
Enough of that mini-rant. :) I think that we should see about repackaging the win32 versions with the less "annoying" bug for now until fixed libs are released. Uwe & Chris, do you think you'll be able to do that? BTW, has anyone checked if it's a known issue upstream?
Cheers, Josh