[Queue-developers] New CVS development version

[W. G. K. <wer...@ya...>]

The ChangeLog below outlines the changes. These are mainly housekeeping
matters in preparation for a 1.40.1 Real Soon Now (perhaps within 24
hours.)

The one coding that change that could impact users has to do with the
"#includes <...>" in the standard Queue .c files which some people
complained about on some platforms. These have been removed (this is GNU
AutoConf's job.)

Hopefully I haven't broken anything (some of the changes to the code
were made by a mechanical script). If so, it'll get fixed tomorrow ....

===== CHANGELOG EXCERPT FOLLOWS ========

*April     10, 2001   W. G. Krebs <wk...@gn...>

        Well, I'm back! It's great to be back. (It's also great to
        have competent help! Thanks QingLong!)

        1. Upped the version number in configure.in to 1.40.1 to prepare
        for new release.

        2. Removed the "#include <arpa/inet.h>" and other
        "#include <>"s from the stuff I've written because it
        broke other people's setup.

        The needed header files are all defined in a line in
./configure.in
        and get automagically defined in "define.h" when ./configure is
        (define.h is included
        via queue.h) so it shouldn't be necessary to include the
        standard stuff. Doing so breaks systems.

        I've noticed some of the GNU files do include standard stuff
        (they don't include define.h or queue.h, however). They only
        include _very_ standard C stuff, and they don't need to worry
        about depedencies within a vendor's include files (the order
        matters on some systems, and changes from system to system,
        believe it or not!)

        Unless we're sure the file exists on ALL platforms, we should
        stick to the practice of putting the include in the line
        in configure.in and then just include either queue.h
        or define.h in the file that needs the header. This way,
        GNU AutoConf will worry about whether the file exists
        on the target and about dependencies.

        3. I've mechanically had the sections were QingLong added 4
        or more lines of code to .c and .h files add a copyright
comment.
        The FSF believes these legal issues are important (should there
        ever be a problem, it's clear what QingLong added.)

        More practical, it makes it much easier to spot what QingLong
        changed (along changes of less than 4 line additions weren't
        commented on.)

        4. I've added copyright notices to QingLong's SysV script.
        The FSF wants copyright notices on EVERY file, asside from
        maybe some trivial one-liners, so I guess I'll be doing the
        .h files that were missed tomorrow.

        This has the practical effect of telling everyone where
        our web site is, anyway.