A Dissabte, 30 d'octubre de 2010, Benjamin Adler va escriure: > On 10/30/2010 01:30 AM, Albert Astals Cid wrote: > >> because I don't really know what the desired behaviour is when > >> "!useMediaBox&& !crop" is passed. > > > > I think false, true and false, false actually create a very similar if > > nto the same behaviour. > > > > But that leads me to the fact that (even if noone that we know was using > > it), your current patch makes it impossible to get the same behaviour of > > having useMediaBox = true and crop = true. Since that gave you a page as > > big as the mediabox but the contents cropped to the cropbox, so i really > > think the display methods should have two parameters > > > > Page::PageBox box, Page::PageBox cropBox > > > > that mimic exactly what the Gfx constructor parameters do, and we'd > > probably need another value for the PageBox enum that would be something > > like NoBox so you can pass it to boxToCrop if you don't want any > > cropping (passing a null to Gfx). > > > > What do you think? > > Funny, I thought the same thing today. The way it is now really doesn't > make any sense to me. With my patch as it is, the better description > would probably be "clip painting to pagebox". > > I don't think we need a noBox value. When not specifying a crop-value, > we use the box for the pagesize as cropbox (so no paint-cropping > occurs). If no parameters are specified, we just use MediaBox for both. > Of course, it wouldn't make any sense to use e.g. trimbox for pagesize > and mediabox for cropping, but we could trim the cropping-box to the > pagesize box before passing it on. Good?
What does "no parameters are specified" mean? > > Would it be ok for you to commit the patch as it is now, so I can work > from there instead of fighting with git, or would you prefer one patch > trying to get it right? One patch done correctly, please. > > >> and modify callers accordingly. Do you agree? > > > > Please don't do any change to existing code unless it's totally > > necessary. > > Oh. I was thinking about doing some more changes that would - in my eyes > - simplify the code (e.g. not passing null to Gfx c'tor, but passing > mediaBox instead). Are you worried about regressions, or whats the idea > behind this? Both regressions and less changes as possible so that if someday xpdf does a new release it'll be as less painful as possible to merge the changes. > > >> P.S: Sorry for making this a two-patch-mess, I need to learn git. > > > > That makes two of us ;-.) > > Then how do you send patches to poppler? I don't send patches, i'm the maintainer so i commit right away ;-.) Albert _______________________________________________ poppler mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/poppler
