Atari FTP/ATARI_FTP_0693.zip/ATARI_FTP_0693/Mint/mntlib32.zoo/README

Text File | 1993-06-17 | 8KB | 161 lines

This is the readme for the MiNT library.

To use the library you will nead the header files in mntincNN.zoo (where NN
is the same patchlevel as the version of the library you're using).

As of patchlevel 26, Eric R. Smith has turned over maintenance of the library
to me (entropy).  Like Eric, I only have GCC, so I've recruited a few people
to act as sub-maintainers for the portions of the library I can't test.

Here are the maintainers' addresses:
GNU C and common files:  [e-mail] (Nick Castellano)
Sozobon C/HSC:  [e-mail] (D. Scott Bigham)
Pure C:  [e-mail] (Michael Hohmuth)
Lattice C:  ???

Be sure to read the "changes" file so that you know what's new in this
release of the library.

My priorities for the library (in approximately this order) are:

1.  Bug fixes, of course!
2.  UNIX compatibility
3.  Synchronization with J. Bammi's GCC library
4.  Support for compilers other than GCC
5.  Frills (beautifying code, fixing ugly hacks, etc.)
6.  Documentation

If you notice a problem with the library but don't know how to fix it,
please try to provide as much information as possible so that I can locate
the problem (a short piece of code demonstrating the bug, for instance).

All top-level sources and all mntinc include files are now in UNIX carriage
control format (LF line termination).  If your compiler/editor really wants
carriage returns, you can convert files back to MS-DOS format (CR-LF line
termination) using the program crlf.ttp supplied in the crlf/ subdirectory
(source included).  I made this change because many of the programs I use
get really confused with MS-DOS (TOS) style files.  Also, J. Bammi's GCC
libs are maintained in the UNIX style, so converting everything will make
synchronizing the two libraries easier.

If you're sending in patches for the library, please take into consideration
the following words from Larry Wall, in the manual page for 'patch'.  It
will make my life so much easier.

    NOTES FOR PATCH SENDERS
        There are several things you should bear  in  mind  if  you  are
        going to  be  sending out patches.  First, you can save people a
        lot of grief by keeping a patchlevel.h file which is patched  to
        increment  the  patch  level as the first diff in the patch file
        you send out.  If you put a Prereq: line in with the  patch,  it
        won't  let them apply patches out of order without some warning.
        Second, make sure you've specified the filenames  right,  either
        in a  context  diff  header, or with an Index: line.  If you are
        patching something in a subdirectory, be sure to tell the  patch
        user to specify a  -p switch as needed.  Third, you can create a
        file by sending out a diff that compares a null file to the file
        you want to create.  This will only work if the file you want to
        create doesn't  exist  already in the target directory.  Fourth,
        take care not to send  out  reversed  patches,  since  it  makes
        people wonder  whether  they  already applied the patch.  Fifth,
        while you may be able to get away with putting 582 diff listings
        into one file, it is probably wiser  to  group  related  patches
        into separate files in case something goes haywire.

Of course the notes about keeping a patchlevel.h file don't apply in this
case, because I maintain the patchlev.h file for the library.  But please
take note of the rest of it, especially the last sentence!

As of Patchlevel 31, the MiNT library no longer supports any version of
GCC before 2.0.

The MiNT library cannot be built on a TOS filesystem, due to name conflicts
between several modules (for example, _udivmod.o and _udivmoddi4.o).  It
should work on a cross-compiler or an ST with a filesystem such as minixfs.

My configuration, used to build the library on an extended filename
V2 minixfs with gcc 2.3.1:
  /mint/bin/mfsconf F: -s n -d n -x t -l n
  UNIXMODE=cu/rUx

Special thanks go out to Jeff Weiner and the rest of the umich archive posse
for letting me use terminator for the maintenance of the library.

May the source be with you.  Wish me luck...I'll need it!

Cheers,
entropy

Eric's original readme file follows:
----------------------------------------------------------------------------

*NOTE*: to compile the library you will need the header files in mntinc25.zoo

============================================================================

Here is mintlib, a library for gcc, Sozobon, Pure C, and Lattice C which
produces programs usable under either MiNT or TOS (of course, some features,
e.g. pipe(), are only available under MiNT). This version has been reasonably
well tested under both TOS and MiNT, and with 16 and 32 bit integers, but no
doubt bugs remain. Please report any that you find.

The main library has the GNU C version of the library, plus the common
files. See the lattice, purec, and sozobon directories for compiler
specific information.

The Lattice C support is definitely a bit rough around the edges (I may
very well have messed up some of the patches, and I don't have Lattice to
test it). Similarly, I can't test the Sozobon and Pure C support myself.

There are no docs ("Use the Source, Luke"), but most of the library should
be pretty self-explanatory. If you know ANSI C and/or Posix, then most
of the functions should be pretty clear to you.

MiNT specific features are active when the external variable __mint is
non-zero; this variable is set automatically by the startup code in main.c.

Some things to watch out for:

(1) MiNT has a blocking fork(), i.e. the parent waits for the child to
    relinquish its address space rather than continuing in parallel.
    (Do NOT rely on this, though, since it will be corrected in a future
    version of MiNT!)
(2) Using the spawn functions instead of fork/exec means that your programs
    will work under TOS as well as under MiNT. vfork() also works under
    TOS now, so vfork/exec is another viable alternative.
(3) The longjmp() code has a call to Psigreturn() embedded in it; this means
    that most signal handlers will work without changes, but in some very
    bizarre circumstances this could cause a problem (if the sig handler
    longjmps *within* itself, and then returns, for example).
(4) Under TOS, all terminal reads are assumed to come from the console
(5) You'll note that there is only minimal support for UNIXMODE;
    this is because MiNT 0.9 supports symlinks in the kernel.
(6) A function call, tfork(), is provided to start a new thread of
    execution in the same address space. The declaration is
    int tfork( int (*func)(), long arg ).
    A new process is started, which executes concurrently with the parent
    process and in the same address space (but a different stack).
    The child begins at the address in func, with the stack set up as though
    the call (*func)(arg) was made (in fact, this call *is* made!).
    The child will exit when func returns, with the exit status being the
    return code of func. The parent continues to execute concurrently;
    in the parent, the value returned from tfork() is the process id of the
    child.
(7) The library is not set up to handle multiple threads of execution in the
    same address space, so you'll have to provide semaphores for critical
    library calls (e.g. malloc) yourself if you have multiple threads.

The library is based on the gcc library that Jwahar Bammi and I put together.
Lots of people have contributed to it, including (but not limited to):

Adrian Ashley, Jwahar Bammi, Scott Bigham, Kai-Uwe Bloem, Howard Chu,
John R. Dunning, Doug Gwyn, Dave Gymer, Michael Hohmuth, Alan Hourihane,
Alex Kiernan, Ulf Moeller, Allan Pratt, Arnold D. Robbins, Edgar Roeder,
Rich Salz, Dale Schumacher, Andreas Schwab, Eric Smith, Henry Spencer,
and Stephen Usher.

I've hacked on things quite a bit, so if they're broken don't blame the
original authors, blame me.

Eric R. Smith