Zip file/directories name convention under MVS
       ---------------------------------------------------
                           Draft 1.1


1. Translating native file names to Zip filenames.

1.1 Zipping a PDS

On MVS there are directories called PDS (Partition Data Set) which have
the following format :   name1.name2.name3(mname)
for example:             myuserid.unzip.c(unzip)

So as you see the path delimiter is '.'. Each dir name can be max 8
chars long beginning with a number.

Between '(' and ')' there is the so called member name - it is also 8
chars long. This is the actual file name.


1.1.1 Converting MVS PDS name to zip path/filename  (status: not implemented)

The PDS name is converted to zippath as follows:
in the zip :     name1/name2/mname.name3
becomes on MVS:  name1.name2.name3(mname)


1.2 Unzipping as PDS                                (status: implemented)

When you unzip the file name myuserid/unzip/unzip.c the same process
is done backwards, so you get : myuserid.unzip.c(unzip)

Notice that the file extension is used as last dirname!


1.2 Unzipping to a different PDS                    (status: implemented)

You can also use -d option while unzipping for example:
unzip mytest myuserid/unzip/unzip.c -dnewdest.test

then the new name will become:
newdest.test.myuserid.unzip.c(unzip)

              Second example:

unzip mytest myuserid/unzip/*.c -dnewdest.test

then you get a PDS:
newdest.test.myuserid.unzip.c(...)

with all *.c files in it.


1.3 Zipping a Sequential Dataset               (status: not implemented)

  Sequential dataset is a dataset with NO members.
Such a dataset is translated from native MVS to zip format by replacing
the '.' (points) with '/' (backslash).

Example:
on MVS:                  name1.name2.name3
becomes in the zip :     name1/name2/name3

NOTE : The new filename in the zip has NO extension this way it can be
       recognised as a Sequential dataset and not a PDS.
       But this also means that all files in the zip archive that have
       no extension will be unzipped as Sequential datasets!


1.4 Using a DDNAMES for input.                  (status: not implemented)

To use DDNAMES as input file names put a 'dd:' before the ddname:
example: zip myzip dd:name1 dd:name2 dd:sales

In the Zip archive the ddnames are saved as name.DDNAME so if you try
the example above you will get in your zip file (when listing it) :

..size .. date time .. crc .. NAME1.DDNAME
..size .. date time .. crc .. NAME2.DDNAME
..size .. date time .. crc .. SALES.DDNAME


1.4 Using a DDNAMES as zip name                    (status: implemented)

It is allowed to use a DDNAME as zipfile, just put dd: before it
example: unzip dd:myzip *.c
   this will unzip all .c files from ddname myzip

example2:   ZIP DD:MYZIP DD:MANE1 MYSOURCE.C MYDOC.TEXT(ZIPPING)
   this will zip ddname name1 file mysource.c and PDS mydoc.text(zipping)
   into as a zip file in the ddname myzip


2. Converting longer path names (unix like)     (status: not implemented)
   to native MVS names.

When in the zip archive there are dirnames longer that 8 chars they are
chopped at the 8 position. For example
        MyLongZippath/WithLongFileName.text
is translated to:
        MYLONGZI.TEXT(WITHLONG)

Notice that all chars are converted to uppercase.


2.1 Using special characters                     (status: implemented)

Also all '_' (underscore), '+' (plus), '-' (minus), '(' and ')'
from the file name/path in the zip archive are skipped because they
are not valid in the MVS filenames.


2.2 Numeric file names                        (status: not implemented)

On MVS no name can begin with a number, so when a dir/file name begins with
one, a leading letter 'N' is inserted.  For example:
          Contents.512
becomes:
          CONTENTS.N512




        Zip file/directories name convention under VM/CMS
       ---------------------------------------------------

1. Translating native file names to Zip filenames.

On VM/CMS (not ESA ) there are NO directories so you got only disks
and files.

The file names are delimited with spaces. But for use with unzip/zip
you have to use '.' points as delimiters.

For example on your A disk you have file called PROFILE EXEC
if you want to zip it type : zip myzip profile.exec

If the same file is on your F disk you have to type:
zip myzip profile.exec.f

So as you can see the general format is fname.ftype.fmode

In the zipfile the disk from which the file comes is not saved!
So only the fname.ftype is saved.

If you unzip and you want to give a different destination disk just use
the -d option like:

                     unzip mytest *.c -df

This will unzip all *.c files to your F disk.


2. Converting longer path names (unix like) to native VM/CMS names.

When in the zip archive there are dirnames longer that 8 chars they are
chopped at the 8 position. Also the path is removed. For example
        Zippath/WithLongFileName.text
is translated to:
        WITHLONG.TEXT

Notice that all chars are converted to uppercase.

Also all '+' (plus), '-' (minus), '(' and ')'
from the file name/path in the zip archive are skipped because they
are not valid in the VM/CMS filenames.

If there is no extension for the file name in the zip archive, unzip
will add .NONAME for example:
        mypath/dir1/testfile
becomes:
        TESTFILE.NONAME

3. Future?

There is also discussion for a new option on ZIP that you can give
a virtual directory to be added before each file name that is zipped.

For example you want to zip a few .c file and put them in the zip
structure under the directory 'mydir/test', but you can't create dirs on
VM/CMS so you have to the something like:

ZIP myzip file1.c file2.c -dmydir/test

and you get in the zip archive files:

       mydir/test/file1.c
       mydir/test/file2.c

-------------------------------------------------------------------------


NOTE: Not all of those functions are implemented in the first beta
      release of VM/MVS UNZIP/ZIP.

Every ideas/corrections/bugs will be appreciated.
Mail to maillist:  Info-ZIP@LISTS.WKU.EDU

George Petrov