On this page:
11.1 Unpacking API
unpack
fold-plt-archive

11 raco unpack: Unpacking Library Collections

The raco unpack command unpacks a ".plt" archive (see raco pack: Packing Library Collections) to the current directory without attempting to install any collections. Use raco pkg (see Package Management in Racket) to install a ".plt" archive as a package, or use raco setup -A (see raco setup: Installation Management) to unpack and install collections from a ".plt" archive.

Command-line flags:

11.1 Unpacking API

 (require setup/unpack) package: base
The setup/unpack library provides raw support for unpacking a ".plt" file.

procedure

(unpack archive    
  [main-collects-parent-dir    
  print-status    
  get-target-directory    
  force?    
  get-target-plt-directory])  void?
  archive : path-string?
  main-collects-parent-dir : path-string? = (current-directory)
  print-status : (string? . -> . any)
   = (lambda (x) (printf "~a\n" x))
  get-target-directory : (-> path-string?)
   = (lambda () (current-directory))
  force? : any/c = #f
  get-target-plt-directory : 
(path-string?
 path-string?
 (listof path-string?)
 . -> . path-string?)
   = 
(lambda (preferred-dir main-dir options)
  preferred-dir)
Unpacks archive.

The main-collects-parent-dir argument is passed along to get-target-plt-directory.

The print-status argument is used to report unpacking progress.

The get-target-directory argument is used to get the destination directory for unpacking an archive whose content is relative to an arbitrary directory.

If force? is true, then version and required-collection mismatches (comparing information in the archive to the current installation) are ignored.

The get-target-plt-directory function is called to select a target for installation for an archive whose is relative to the installation. The function should normally return one if its first two arguments; the third argument merely contains the first two, but has only one element if the first two are the same. If the archive does not request installation for all uses, then the first two arguments will be different, and the former will be a user-specific location, while the second will refer to the main installation.

procedure

(fold-plt-archive archive    
  on-config-fn    
  on-setup-unit    
  on-directory    
  on-file    
  initial-value)  any/c
  archive : path-string?
  on-config-fn : (any/c any/c . -> . any/c)
  on-setup-unit : (any/c input-port? any/c . -> . any/c)
  on-directory : 
((or/c path-string?
       (list/c (or/c 'collects 'doc 'lib 'include)
               path-string?))
 any/c
 . -> . any/c)
  on-file : 
(or/c ((or/c path-string?
             (list/c (or/c 'collects 'doc 'lib 'include)
                     path-string?))
       input-port?
       any/c
       . -> . any/c)
      ((or/c path-string?
             (list/c (or/c 'collects 'doc 'lib 'include)
                     path-string?))
       input-port?
       (one-of/c 'file 'file-replace)
       any/c
       . -> . any/c))
  initial-value : any/c
Traverses the content of archive, which must be a ".plt" archive that is created with the default unpacking unit and configuration expression. The configuration expression is not evaluated, the unpacking unit is not invoked, and not files are unpacked to the filesystem. Instead, the information in the archive is reported back through on-config, on-setup-unit, on-directory, and on-file, each of which can build on an accumulated value that starts with initial-value and whose final value is returned.

The on-config-fn function is called once with an S-expression that represents a function to implement configuration information. The second argument to on-config is initial-value, and the function’s result is passes on as the last argument to on-setup-unit.

The on-setup-unit function is called with the S-expression representation of the installation unit, an input port that points to the rest of the file, and the accumulated value. This input port is the same port that will be used in the rest of processing, so if on-setup-unit consumes any data from the port, then that data will not be consumed by the remaining functions. (This means that on-setup-unit can leave processing in an inconsistent state, which is not checked by anything, and therefore could cause an error.) The result of on-setup-unit becomes the new accumulated value.

For each directory that would be created by the archive when unpacking normally, on-directory is called with the directory path (described more below) and the accumulated value up to that point, and its result is the new accumulated value.

For each file that would be created by the archive when unpacking normally, on-file is called with the file path (described more below), an input port containing the contents of the file, an optional mode symbol indicating whether the file should be replaced, and the accumulated value up to that point; its result is the new accumulated value. The input port can be used or ignored, and parsing of the rest of the file continues the same either way. After on-file returns control, however, the input port is drained of its content.

A directory or file path can be a plain path, or it can be a list containing 'collects, 'doc, 'lib, or 'include and a relative path. The latter case corresponds to a directory or file relative to a target installation’s collection directory (in the sense of find-collects-dir), documentation directory (in the sense of find-doc-dir), library directory (in the sense of find-lib-dir), or “include” directory (in the sense of find-include-dir).