On Tue, May 24, 2016 at 04:58:36AM -0400, Frediano Ziglio wrote: > > > > Hi, > > > > On Mon, 2016-05-23 at 16:19 -0500, Jonathon Jongsma wrote: > > > On Mon, 2016-05-23 at 15:17 -0500, Jonathon Jongsma wrote: > > > > On Sun, 2016-05-22 at 04:19 -0400, Frediano Ziglio wrote: > > > > > > > > > > /* FIXME: move back to display-channel.h (once structs are private) > > > > > > */ > > > > > > typedef struct Drawable Drawable; > > > > > > -typedef struct DisplayChannelClient DisplayChannelClient; > > > > > > > > > > > > > > > > We definitively should consider using struct Xxxx instead of moving > > > > > these typedef everytime in order to compile... > > > > > > > > Or create a spice-types.h header that contains forward declarations for > > > > all > > > > structs and include that in every other file. The includes really are > > > > quite > > > > a > > > > tangled mess and it's not an easy problem to unravel. I spent some time > > > > trying, > > > > but it's very difficult to do in a gradual way. I've resisted doing > > > > something > > > > like this in the past, but I'm starting to warm up to it... > > > > > > > > Jonathon > > > > > > > > > So, I chatted (very briefly) with Frediano about this on IRC after sending > > > this > > > email and I thought we should gauge opinions on options here. > > > > > > This typedef stuff really does make refactoring more difficult. When we > > > move > > > stuff around, we have to keep adding temporary typedefs to make things > > > compile. > > > Or we have to remove these typedefs to avoid redefining the same struct > > > multiple > > > types. So there are several options: > > > > > > 1. Use 'struct Foo' throughout the code instead of 'Foo' so that we don't > > > have > > > to deal with the typedef mess. > > > > yes, +1 > > The issue here is having a lot of "struct Xxxx". > > I would propose a half way to this: > typedef only in the public relative header (for instance DisplayChannel in > display-challel.h). This should avoid most typedefs in .c files. Also is > supposed that most function dealing with that type are declared in that file > too so this way you can reuse the typedef. Other headers will use > struct Xxxx; (declaration) and struct Xxxx when referenced. > > +1. +1 as well to use only typedefs on headers if that solves the issue. > > > > > > > 2. the approach listed above (a spice-types.h file that simply typedefs all > > > structs). This is a bit of a maintenance nightmare and needs to be kept up > > > to > > > date and for a lot of types that are relatively self-contained it's not > > > even > > > necessary to put them in a global header like this... > > > > I am afraid that would be easy to forget about updating the file > > There are some issues of this approach: > - scalability, if all types get into a single file every type addition > will cause all files to be recompiled; > - prone to exceptions: > - public headers have to keep their types; > - why moving self contained (like glz or encoders) into this file? > > > > > > > 3. continue with the fragile approach we now have > > > > I think fragile is too strong word > > Surely not! > -1! > > > > > > > 4. use C++ :D > > > > I would wait with this > > Yes, I think this is a weak argument, there are lot more useful feature > we can use for other stuff. > > > > > > > Anything else I'm not thinking of? I'd like to hear opinions from as many > > > regular contributors as possible so we have a rough consensus, but I'd > > > really > > > like to do something about it. > > > > Personally I would go for the first option, but it would be nice to avoid > > some > > mass rename... > > > > Pavel > > > > Jonathon, is not clear your vote :) > > Frediano > _______________________________________________ > Spice-devel mailing list > Spice-devel@xxxxxxxxxxxxxxxxxxxxx > https://lists.freedesktop.org/mailman/listinfo/spice-devel _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel