On Fri, 21 Jun 2013 10:56:54 -0500, Paul Gilmartin wrote: >Reading further, the only mentions of UNIX files are in (abridged): > >11.7.2.2 <FILEDEF> Tag syntax > > README > indicates that the associated file is a sequential text data set or a file > in the UNIX file system that is intended to be viewable text after it is > placed in the package. These files are not placed in archives by GIMZIP and > are not compressed. ... > >SMPHOLD > indicates that the associated data set or file contains SMP/E HOLDDATA > statements. ... > >SMPPTFIN > indicates that the associated data set or file contains SMP/E modification > control statements (MCS). ...
You didn't read far enough, Paul. Immediately following the part you quoted is this: <quote> The type attribute must not be specified when a VSAM data set or a UNIX file system directory is specified on the name attribute. If the type attribute is not specified, then the data set specified on the name attribute can be either a sequential, partitioned, or VSAM data set, or a file or directory in the UNIX file system </quote> If someone were to use GIMZIP for files that are not to be a network package for SMP/E, the only kind of tag that would make sense is the README tag. That tag causes the referenced file to be included in the directory without first being tar'ed. That is, if it is a plain text file, it can be read directly. Any of the supported file types can be included in the archive with no tag specified. There is also this at the beginning of chapter 11.7, describing GIMZIP. <quote> The GIMZIP service routine creates portable packages of software and associated materials. Typically the packages will contain SYSMODs, RELFILE data sets, HOLDDATA, and associated materials such as documentation, samples, and text files. These GIMZIP packages may be transported through a network, processed by the GIMUNZIP service routine, and then processed by the SMP/E RECEIVE command. More specifically, a GIMZIP package consists of a single package definition file, a set of archive files, and text files. The package definition file describes the total package and identifies the archive files and text files contained in the package. An archive file consists of : 1. a portable image of any of the following: --- a sequential data set --- a partitioned data set --- a VSAM data set --- a file in the UNIX file system --- a directory in the UNIX file system 2. and the information necessary to reload the data from the portable image. </quote> The archive is a Unix directory that contains - Any files of type README. These files are not compressed. - GIMPAF.XML and GIMPAF.XSL. These files are also not compressed and they contain control information used to process the archive. These files are also not compressed. - An SMPHOLD directory containing all of the files that were included with type SMPHOLD. All of these files are compressed. - An SMPPTFIN directory containing all of the files that were included with type SMPPTFIN. All of these files are compressed. - An SMPRELF directory containing all of the files that were included with type SMPRELF. All of these files are compressed. - Any other files that were included with no type specified. All of these files are compressed. The SMPHOLD, SMPPTFIN and SMPRELF directories are only included if there was one or more files included in the archive with that type. > >I suppose that if GIMZIP does no syntax checking any UNIX file can be called >either >SMPHOLD or SMPPTFIN. There is no reason to do that unless the file is to be processed as SMPHOLD or SMPPTFIN by SMP/E. >Does this still require specific RACF authorization, as per APAR IO11698? Yes. See the SMP/E user's guide chapter 3.1, "Authorizing use of SMP/E commands and services". >Are you, or anyone, actively using GIMZIP as a generalized interchange >technique, independent of SMP/E? I'm not. I've only used it to create network packages for use by SMP/E. I have, however, included other kinds of files in the package. -- Tom Marchant ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN