doc/method.sgml

   1 <!-- -*- mode: sgml; mode: fold -*- -->
   2 <!doctype debiandoc  PUBLIC  "-//DebianDoc//DTD DebianDoc//EN">
   3 <book>
   4 <title>APT Method Interface </title>
   5
   6 <author>Jason Gunthorpe <email>jgg@debian.org</email></author>
   7 <version>$Id: method.sgml,v 1.10 2003/02/12 15:05:46 doogie Exp $</version>
   8
   9 <abstract>
  10 This document describes the interface that APT uses to the archive
  11 access methods.
  12 </abstract>
  13
  14 <copyright>
  15 Copyright &copy; Jason Gunthorpe, 1998.
  16 <p>
  17 "APT" and this document are free software; you can redistribute them and/or
  18 modify them under the terms of the GNU General Public License as published
  19 by the Free Software Foundation; either version 2 of the License, or (at your
  20 option) any later version.
  21
  22 <p>
  23 For more details, on Debian systems, see the file
  24 /usr/share/common-licenses/GPL for the full license.
  25 </copyright>
  26
  27 <toc sect>
  28
  29 <chapt>Introduction
  30 <!-- General                                                           {{{ -->
  31 <!-- ===================================================================== -->
  32 <sect>General
  33
  34 <p>
  35 The APT method interface allows APT to acquire archive files (.deb), index
  36 files (Packages, Release, Mirrors) and source files (.tar.gz, .diff). It
  37 is a general, extensible system designed to satisfy all of these
  38 requirements:
  39
  40 <enumlist>
  41 <item>Remote methods that download files from a distant site
  42 <item>Resume of aborted downloads
  43 <item>Progress reporting
  44 <item>If-Modified-Since (IMS) checking for index files
  45 <item>In-Line MD5 generation
  46 <item>No-copy in-filesystem methods
  47 <item>Multi-media methods (like CD's)
  48 <item>Dynamic source selection for failure recovery
  49 <item>User interaction for user/password requests and media swaps
  50 <item>Global configuration
  51 </enumlist>
  52
  53 Initial releases of APT (0.1.x) used a completely different method
  54 interface that only supported the first 6 items. This new interface
  55 deals with the remainder.
  56 </sect>
  57                                                                   <!-- }}} -->
  58 <!-- Terms                                                             {{{ -->
  59 <!-- ===================================================================== -->
  60 <sect>Terms
  61
  62 <p>
  63 Several terms are used through out the document, they have specific
  64 meanings which may not be immediately evident. To clarify they are summarized
  65 here.
  66
  67 <taglist>
  68 <tag>source<item>
  69 Refers to an item in source list. More specifically it is the broken down
  70 item, that is each source maps to exactly one index file. Archive sources
  71 map to Package files and Source Code sources map to Source files.
  72
  73 <tag>archive file<item>
  74 Refers to a binary package archive (.deb, .rpm, etc).
  75
  76 <tag>source file<item>
  77 Refers to one of the files making up the source code of a package. In
  78 debian it is one of .diff.gz, .dsc. or .tar.gz.
  79
  80 <tag>URI<item>
  81 Universal Resource Identifier (URI) is a super-set of the familiar URL
  82 syntax used by web browsers. It consists of an access specification
  83 followed by a specific location in that access space. The form is
  84 &lt;access&gt;:&lt;location&gt;. Network addresses are given with the form
  85 &lt;access&gt;://[&lt;user&gt;[:&lt;pas&gt;]@]hostname[:port]/&lt;location&gt;.
  86 Some examples:
  87 <example>
  88 file:/var/mirrors/debian/
  89 ftp://ftp.debian.org/debian
  90 ftp://jgg:MooCow@localhost:21/debian
  91 nfs://bigred/var/mirrors/debian
  92 rsync://debian.midco.net/debian
  93 cdrom:Debian 2.0r1 Disk 1/
  94 </example>
  95
  96 <tag>method<item>
  97 There is a one to one mapping of URI access specifiers to methods. A method
  98 is a program that knows how to handle a URI access type and operates according
  99 to the specifications in this file.
 100
 101 <tag>method instance<item>
 102 A specific running method. There can be more than one instance of each method
 103 as APT is capable of concurrent method handling.
 104
 105 <tag>message<item>
 106 A series of lines terminated by a blank line sent down one of the
 107 communication lines. The first line should have the form xxx TAG
 108 where xxx are digits forming the status code and TAG is an informational
 109 string
 110
 111 <tag>acquire<item>
 112 The act of bring a URI into the local pathname space. This may simply
 113 be verifying the existence of the URI or actually downloading it from
 114 a remote site.
 115
 116 </taglist>
 117
 118 </sect>
 119                                                                   <!-- }}} -->
 120 <chapt>Specification
 121 <!-- Overview                                                          {{{ -->
 122 <!-- ===================================================================== -->
 123 <sect>Overview
 124
 125 <p>
 126 All methods operate as a sub process of a main controlling parent. 3 FD's
 127 are opened for use by the method allowing two way communication and
 128 emergency error reporting. The FD's correspond to the well known unix FD's,
 129 stdin, stdout and stderr.
 130
 131 <p>
 132 Through operation of the method communication is done via http
 133 style plain text. Specifically RFC-822 (like the Package file) fields
 134 are used to describe items and a numeric-like header is used to indicate
 135 what is happening. Each of these distinct communication messages should be
 136 sent quickly and without pause.
 137
 138 <p>
 139 In some instances APT may pre-invoke a method to allow things like file
 140 URI's to determine how many files are available locally.
 141 </sect>
 142                                                                   <!-- }}} -->
 143 <!-- Message Overview                                                  {{{ -->
 144 <!-- ===================================================================== -->
 145 <sect>Message Overview
 146
 147 <p>
 148 The first line of each message is called the message header. The first
 149 3 digits (called the Status Code) have the usual meaning found in the
 150 http protocol. 1xx is informational, 2xx is successful and 4xx is failure.
 151 The 6xx series is used to specify things sent to the method. After the
 152 status code is an informational string provided for visual debugging.
 153
 154 <list>
 155 <item>100 Capabilities - Method capabilities
 156 <item>101 Log - General Logging
 157 <item>102 Status - Inter-URI status reporting (login progress)
 158 <item>200 URI Start - URI is starting acquire
 159 <item>201 URI Done - URI is finished acquire
 160 <item>400 URI Failure - URI has failed to acquire
 161 <item>401 General Failure - Method did not like something sent to it
 162 <item>402 Authorization Required - Method requires authorization
 163         to access the URI. Authorization is User/Pass
 164 <item>403 Media Failure - Method requires a media change
 165 <item>600 URI Acquire - Request a URI be acquired
 166 <item>601 Configuration - Sends the configuration space
 167 <item>602 Authorization Credentials - Response to the 402 message
 168 <item>603 Media Changed - Response to the 403 message
 169 </list>
 170
 171 Only the 6xx series of status codes is sent TO the method. Furthermore
 172 the method may not emit status codes in the 6xx range. The Codes 402
 173 and 403 require that the method continue reading all other 6xx codes
 174 until the proper 602/603 code is received. This means the method must be
 175 capable of handling an unlimited number of 600 messages.
 176
 177 <p>
 178 The flow of messages starts with the method sending out a
 179 <em>100 Capabilities</> and APT sending out a <em>601 Configuration</>.
 180 After that APT begins sending <em>600 URI Acquire</> and the method
 181 sends out <em>200 URI Start</>, <em>201 URI Done</> or
 182 <em>400 URI Failure</>. No synchronization is performed, it is expected
 183 that APT will send <em>600 URI Acquire</> messages at -any- time and
 184 that the method should queue the messages. This allows methods like http
 185 to pipeline requests to the remote server. It should be noted however
 186 that APT will buffer messages so it is not necessary for the method
 187 to be constantly ready to receive them.
 188 </sect>
 189                                                                   <!-- }}} -->
 190 <!-- Header Fields                                                     {{{ -->
 191 <!-- ===================================================================== -->
 192 <sect>Header Fields
 193
 194 <p>
 195 The following is a short index of the header fields that are supported
 196
 197 <taglist>
 198 <tag>URI<item>URI being described by the message
 199 <tag>Filename<item>Location in the filesystem
 200 <tag>Last-Modified<item>A time stamp in RFC1123 notation for use by IMS checks
 201 <tag>IMS-Hit<item>The already existing item is valid
 202 <tag>Size<item>Size of the file in bytes
 203 <tag>Resume-Point<item>Location that transfer was started
 204 <tag>MD5-Hash<item>Computed MD5 hash for the file
 205 <tag>Message<item>String indicating some displayable message
 206 <tag>Media<item>String indicating the media name required
 207 <tag>Site<item>String indicating the site authorization is required for
 208 <tag>User<item>Username for authorization
 209 <tag>Password<item>Password for authorization
 210 <tag>Fail<item>Operation failed
 211 <tag>Drive<item>Drive the media should be placed in
 212 <tag>Config-Item<item>
 213 A string of the form <var>item</>=<var>value</> derived from the APT
 214 configuration space. These may include method specific values and general
 215 values not related to the method. It is up to the method to filter out
 216 the ones it wants.
 217 <tag>Single-Instance<item>Requires that only one instance of the method be run
 218                           This is a yes/no value.
 219 <tag>Pipeline<item>The method is capable of pipelining.
 220 <tag>Local<item>The method only returns Filename: fields.
 221 <tag>Send-Config<item>Send configuration to the method.
 222 <tag>Needs-Cleanup<item>The process is kept around while the files it returned
 223 are being used. This is primarily intended for CD-ROM and File URIs that need
 224 to unmount filesystems.
 225 <tag>Version<item>Version string for the method
 226 </taglist>
 227
 228 This is a list of which headers each status code can use
 229
 230 <taglist>
 231 <tag>100 Capabilities<item>
 232 Displays the capabilities of the method. Methods should set the
 233 pipeline bit if their underlying protocol supports pipelining. The
 234 only known method that does support pipelining is http.
 235 Fields: Version, Single-Instance, Pre-Scan, Pipeline, Send-Config,
 236 Needs-Cleanup
 237
 238 <tag>101 Log<item>
 239 A log message may be printed to the screen if debugging is enabled. This
 240 is only for debugging the method.
 241 Fields: Message
 242
 243 <tag>102 Status<item>
 244 Message gives a progress indication for the method. It can be used to show
 245 pre-transfer status for Internet type methods.
 246 Fields: Message
 247
 248 <tag>200 URI Start<item>
 249 Indicates the URI is starting to be transferred. The URI is specified
 250 along with stats about the file itself.
 251 Fields: URI, Size, Last-Modified, Resume-Point
 252
 253 <tag>201 URI Done<item>
 254 Indicates that a URI has completed being transferred. It is possible
 255 to specify a <em>201 URI Done</> without a <em>URI Start</> which would
 256 mean no data was transferred but the file is now available. A Filename
 257 field is specified when the URI is directly available in the local
 258 pathname space. APT will either directly use that file or copy it into
 259 another location. It is possible to return Alt-* fields to indicate that
 260 another possibility for the URI has been found in the local pathname space.
 261 This is done if a decompressed version of a .gz file is found.
 262 Fields: URI, Size, Last-Modified, Filename, MD5-Hash
 263
 264 <tag>400 URI Failure<item>
 265 Indicates a fatal URI failure. The URI is not retrievable from this source.
 266 As with <em>201 URI Done</> <em>200 URI Start</> is not required to precede
 267 this message
 268 Fields: URI, Message
 269
 270 <tag>401 General Failure<item>
 271 Indicates that some unspecific failure has occurred and the method is unable
 272 to continue. The method should terminate after sending this message. It
 273 is intended to check for invalid configuration options or other severe
 274 conditions.
 275 Fields: Message
 276
 277 <tag>402 Authorization Required<item>
 278 The method requires a Username and Password pair to continue. After sending
 279 this message the method will expect APT to send a <em>602 Authorization
 280 Credentials</> message with the required information. It is possible for
 281 a method to send this multiple times.
 282 Fields: Site
 283
 284 <tag>403 Media Failure<item>
 285 A method that deals with multiple media requires that a new media be inserted.
 286 The Media field contains the name of the media to be inserted.
 287 Fields: Media, Drive
 288
 289 <tag>600 URI Acquire<item>
 290 APT is requesting that a new URI be added to the acquire list. Last-Modified
 291 has the time stamp of the currently cache file if applicable. Filename
 292 is the name of the file that the acquired URI should be written to.
 293 Fields: URI, Filename Last-Modified
 294
 295 <tag>601 Configuration<item>
 296 APT is sending the configuration space to the method. A series of
 297 Config-Item fields will be part of this message, each containing an entry
 298 from the configuration space.
 299 Fields: Config-Item.
 300
 301 <tag>602 Authorization Credentials<item>
 302 This is sent in response to a <em>402 Authorization Required</> message.
 303 It contains the entered username and password.
 304 Fields: Site, User, Password
 305
 306 <tag>603 Media Changed<item>
 307 This is sent in response to a <em>403 Media Failure</> message. It
 308 indicates that the user has changed media and it is safe to proceed.
 309 Fields: Media, Fail
 310 </taglist>
 311
 312 </sect>
 313                                                                   <!-- }}} -->
 314 <!-- Method Notes                                                      {{{ -->
 315 <!-- ===================================================================== -->
 316 <sect>Notes
 317
 318 <p>
 319 The methods supplied by the stock apt are:
 320 <enumlist>
 321 <item>cdrom - For Multi-Disc CD-ROMs
 322 <item>copy - (internal) For copying files around the filesystem
 323 <item>file - For local files
 324 <item>gzip - (internal) For decompression
 325 <item>http - For HTTP servers
 326 </enumlist>
 327
 328 <p>
 329 The two internal methods, copy and gzip, are used by the acquire code to
 330 parallize and simplify the automatic decompression of package files as well
 331 as copying package files around the file system. Both methods can be seen to
 332 act the same except that one decompresses on the fly. APT uses them by
 333 generating a copy URI that is formed identically to a file URI. The destination
 334 file is send as normal. The method then takes the file specified by the
 335 URI and writes it to the destination file. A typical set of operations may
 336 be:
 337 <example>
 338 http://foo.com/Packages.gz -> /bar/Packages.gz
 339 gzip:/bar/Packages.gz -> /bar/Packages.decomp
 340 rename Packages.decomp to /final/Packages
 341 </example>
 342
 343 <p>
 344 The http method implements a fully featured HTTP/1.1 client that supports
 345 deep pipelining and reget. It works best when coupled with an apache 1.3
 346 server. The file method simply generates failures or success responses with
 347 the filename field set to the proper location. The cdrom method acts the same
 348 except that it checks that the mount point has a valid cdrom in it. It does
 349 this by (effectively) computing a md5 hash of 'ls -l' on the mountpoint.
 350
 351 </sect>
 352                                                                   <!-- }}} -->
 353
 354 </book>