On 7/24/07, Melchior FRANZ <[EMAIL PROTECTED]> wrote: > * Hans Ulrich Niedermann -- Sunday 22 July 2007: > > As the whole thing is still in development, it makes sense for > > developers to have more than just one of the three, so that one can > > compare their behaviours. > > And how many developers actually do it? My guess: one. That would > be you. But you have that solution on your HD already, anyway, so > we won't have to change CVS. :-P
I appreciate the change, as I do the same thing, and it's even worse on my laptop (see below). > > Total is about 800MB for the CVS solution. > [...] > > Total: About 500MB for the statically linked solution. > [...] > > Total: About 320MB for the dynamically linked solution, i.e. much less > > than half the original size. > > Developers don't care that much about the binary size. They have > debugging symbols and other stuff compiled in, although that blows > up the binary size. Developers happily commit uncompressed aircraft > textures, although this actually concerns *everyone*, not just > developers, and I seem to be the only one who cares. :-} Even developers (in which I include myself in the loosest sense of the word for FG) start to care when things get huge and multiply. My (dirty) source trees are up to 8GB, and while I happen to have the extra space at the moment that's not the normal state for me. > If your argument would have been that we want shared libs (which we > clearly didn't want in the past), then there'd be something to discuss. > > But if it's only about developers wasting less disk space when they > have binaries linked against all available windowing libs on their > disk, then I must say it's totally pointless. In the PLIB branch > we should rather make SDL default, as freeglut is notoriously broken > (<repeatable> keys; reportedly slower FPS(?)). And in the OSG branch > there are plans to drop GLUT and SDL support altogether, as soon as > OSG provides all necessary functionality. It doesn't make sense to > maintain additional dependencies when there's no need to. And given > that osgviewer *almost* works already, I don't see the need to rework > the build system for a short period, when the advantages are rather > questionable. He didn't have an argument. He had a solution that helped him out and he contributed it. If it doesn't hurt the existing system but only enables people to save some disk space and (more importantly) time, then it sounds worthwhile. I don't know how many times I as a bug reporter have been asked to compile with SDL or osgviewer or GLUT, or both that I hadn't already tried. I as a bleeding-edge user (not the same as a release user of course) have played with both plib and osg, trying to see which works better for my needs now, and because I care about the next released product checking whether bugs I find are in plib as well as osg, and because I care about the far future released products checking whether bugs I see in plib are also in osg... > What doesn't yet work in osgviewer is: > > - mouse wrapping in look-around mode (The glut and sdl interfaces > had the same problem. The fix should be as simple in osgviewer as > it was there.) > > - <repeatable> key flags don't work (just as in freeglut) And it crashes in OS X. -- Hans Fugal Fugal Computing ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
