This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: C++'s 'export' Keyword?


Richard Kenner wrote:

The SC isn't going to reject implementation of "export"; RMS will drag
his heels a bit, but he's accepted GNAT's support of dumping and reloading
meta-information about Ada programs,


Ada just dumps the information: it does not reload it. I don't think this
distinction is relevant to the discussion, but I wanted to clarify this anyway.

Right, just so that things are clear. Ada has the capabilility of fully dumping the generated expanded tree with no loss of information. It provides routines for reading this tree back in. The compiler does not have an option for reading back these trees. However, it would indeed be possible for some back end to read the trees and process them, which is the kind of thing that has been considered worrisome in the past. The dumping of trees is in no way essential to the compilation of Ada as such, but is used in the ASIS implementation which allows programs to get hold of semantic information generated by the compiler for compiled programs.



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]