Samuel Tardieu writes: > | - should we build gnat from the gcc-3.1 source at all? > > Sure.
Then I add you to the maintainers list and you subscribe to [EMAIL PROTECTED] > | - package names: I choose gnat-3.15 and libgnat3.15a. Is this ok, or > | should it be gnat-3.1? > > Mmm, at least until we are sure that this version of GNAT is as stable as > the previous one, I would prefer a totally different name (the official > release number will be GNAT 5.0 anyway). > > What about gcc-gnat-<date>? No. this will be from the gcc-3.1 release. > Or gcc-gnat-3.1? If everything gets OK, no > problem for moving to your package. Then again, why not gnat-3.15? Or gnat-3.1? And how about naming the binaries? Currently the are called <tool>-3.15. _But_ the build process expects that the name gnatbind is available? Is there a small package with Ada source packaged for Debian, which can be used as a test case? Maybe not hello-world, but something with more than one module. > | - Do we want to have a versioned Ada package? If yes, the existing > | Ada compiler has to be repackaged. > | > | - If we have versioned names, then we will need to provide the correct > | gnatbind executable for the build process. Currently I only tested > | with gnat-3.14. > | > | - The installation of gnat goes wrong. I had to explicitely call > | the add-install.common target. > | > | - Sam, please could you update the relink script from gnat-3.14 > | for gnat-3.15? Currently all tools are linked statically. > > Unfortunately, I am fully booked for next week, then I leave for a 3 weeks > vacation, then two crazy weeks and again one week of vacation. As a > consequence, I will be totally unable to work on GNAT and unresponsive > for the next 7 weeks at least. well, it works, but uses more space. so it's not urgent. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
