doc/design.dbk

   1 <?xml version="1.0" encoding="UTF-8"?>
   2 <!-- -*- DocBook -*- -->
   3 <!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
   4   "http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [
   5 <!ENTITY % aptverbatiment SYSTEM "apt-verbatim.ent"> %aptverbatiment;
   6 ]>
   7
   8 <book lang="en">
   9
  10 <title>The APT project design document</title>
  11
  12 <bookinfo>
  13
  14 <authorgroup>
  15   <author>
  16     <personname>Manoj Srivastava</personname><email>srivasta@debian.org</email>
  17   </author>
  18 </authorgroup>
  19
  20 <releaseinfo>Version &apt-product-version;</releaseinfo>
  21
  22 <abstract>
  23 <para>
  24 This document is an overview of the specifications and design goals of the APT
  25 project. It also attempts to give a broad description of the implementation
  26 as well.
  27 </para>
  28 </abstract>
  29
  30 <copyright><year>1997</year><holder>Manoj Srivastava</holder></copyright>
  31
  32 <legalnotice>
  33 <title>License Notice</title>
  34 <para>
  35 APT, including this document, is free software; you may redistribute it and/or
  36 modify it under the terms of the GNU General Public License as published by the
  37 Free Software Foundation; either version 2, or (at your option) any later
  38 version.
  39 </para>
  40 <para>
  41 This is distributed in the hope that it will be useful, but <emphasis>without
  42 any warranty</emphasis>; without even the implied warranty of merchantability
  43 or fitness for a particular purpose. See the GNU General Public License for
  44 more details.
  45 </para>
  46 <para>
  47 You should have received a copy of the GNU General Public License with your
  48 Debian system, in <literal>/usr/share/common-licenses/GPL</literal>, or with
  49 the <command>debiandoc-sgml</command> source package as the file
  50 <literal>COPYING</literal>. If not, write to the Free Software Foundation,
  51 Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  52 </para>
  53 </legalnotice>
  54
  55 </bookinfo>
  56
  57 <chapter id="introduction"><title>Introduction</title>
  58 <para>
  59 APT is supposed to be a replacement for dselect, and not a replacement for
  60 dpkg. However, since addition functionality has been required for APT, and
  61 given the fact that this is very closely related to dpkg, it is not
  62 unreasonable to expect that additional functionality in the underlying dpkg
  63 would also be requested.
  64 </para>
  65 <para>
  66 Deity/dselect are the first introduction that people have to Debian, and
  67 unfortunately this first impression contributes greatly to the public
  68 perception of the distribution. It is imperative that this be a showcase for
  69 Debian, rather than frighten novices away (which has been an accusation often
  70 levelled at the current system)
  71 </para>
  72 </chapter>
  73
  74 <chapter id="ch2"><title>Requirements</title>
  75 <orderedlist numeration="arabic">
  76 <listitem>
  77 <para>
  78 APT should be a replacement for dselect. Therefore it should have all the
  79 functionality that dselect has currently. This is the primary means of
  80 interaction between the user and the package management system, and it should
  81 be able to handle all tasks involved in installing, upgrading, and routine
  82 management without having the users take recourse to the underlying management
  83 system.
  84 </para>
  85 </listitem>
  86 <listitem>
  87 <para>
  88 It should be easier to use and less confusing for novice users. The primary
  89 stimulus for the creation of APT was the perceived intractability, complexity,
  90 and non-intuitive behavior of the existing user interface, and as such, human
  91 factors must be a primary mandate of APT.
  92 </para>
  93 </listitem>
  94 <listitem>
  95 <para>
  96 It should be able to group packages more flexibly, and possibly allow
  97 operations based on a group. One should be able to select, or deselect,
  98 a coherent group of related packages simultaneously, allowing one to add,
  99 remove, or upgrade functionality to a machine as one step.
 100 </para>
 101 </listitem>
 102 <listitem>
 103 <para>
 104 This would allow APT to handle <emphasis>standard installations</emphasis>,
 105 namely, one could then install a set of packages to enable a machine to
 106 fulfill specific tasks. Define a few standard installations, and which
 107 packages are included therein. The packages should be internally consistent.
 108 </para>
 109 </listitem>
 110 <listitem>
 111 <para>
 112 Make use of a keywords field in package headers; provide a standard list of
 113 keywords for people to use. This could be the underpinning to allow the
 114 previous two requirements to work (though the developers are not constrained
 115 to implement the previous requirements using keywords)
 116 </para>
 117 </listitem>
 118 <listitem>
 119 <para>
 120 Use dependencies, conflicts, and reverse dependencies to properly order
 121 packages for installation and removal. This has been a complaint in the past
 122 that the installation methods do not really understand dependencies, causing
 123 the upgrade process to break, or allowing the removal of packages that left the
 124 system in an untenable state by breaking the dependencies on packages that were
 125 dependent on the package being removed. A special emphasis is placed on
 126 handling pre-dependencies correctly; the target of a predependency has to be
 127 fully configured before attempting to install the pre-dependent package. Also,
 128 <emphasis>configure immediately</emphasis> requests mentioned below should be
 129 handled.
 130 </para>
 131 </listitem>
 132 <listitem>
 133 <para>
 134 Handle replacement of a package providing a virtual package with another (for
 135 example, it has been very difficult replacing <command>sendmail</command> with
 136 <command>smail</command>, or vice versa), making sure that the dependencies are
 137 still satisfied.
 138 </para>
 139 </listitem>
 140 <listitem>
 141 <para>
 142 Handle source lists for updates from multiple sources. APT should also be able
 143 to handle diverse methods of acquiring new packages; local filesystem,
 144 mountable CD-ROM drives, FTP accessible repositories are some of the methods
 145 that come to mind. Also, the source lists can be separated into categories,
 146 such as main, contrib, non-us, non-local, non-free, my-very-own, etc. APT
 147 should be set up to retrieve the Packages files from these multiple source
 148 lists, as well as retrieving the packages themselves.
 149 </para>
 150 </listitem>
 151 <listitem>
 152 <para>
 153 Handle base of source and acquire all Packages files underneath. (possibly
 154 select based on architecture), this should be a simple extension of the
 155 previous requirement.
 156 </para>
 157 </listitem>
 158 <listitem>
 159 <para>
 160 Handle remote installation (to be implemented maybe in a future version, it
 161 still needs to be designed). This would ease the burden of maintaining
 162 multiple Debian machines on a site. In the authors opinion this is a killer
 163 difference for the distribution, though it may be too hard a problem to be
 164 implemented with the initial version of APT. However, some thought must be
 165 given to this to enable APT to retain hooks for future functionality, or at
 166 least to refrain from methods that may preclude remote activity. It is
 167 desirable that adding remote installation not require a redesign of APT from
 168 the ground up.
 169 </para>
 170 </listitem>
 171 <listitem>
 172 <para>
 173 Be scalable. Dselect worked a lot better with 400 packages, but at last count
 174 the number of packages was around twelve hundred and climbing. This also
 175 requires APT to pay attention to the needs of small machines which are low on
 176 memory (though this requirement shall diminish as we move towards bigger
 177 machines, it would still be nice if Debian worked on all old machines where
 178 Linux itself would work).
 179 </para>
 180 </listitem>
 181 <listitem>
 182 <para>
 183 Handle install immediately requests. Some packages, like watchdog, are
 184 required to be working for the stability of the machine itself. There are
 185 others which may be required for the correct functioning of a production
 186 machine, or which are mission critical applications. APT should, in these
 187 cases, upgrade the packages with minimal downtime; allowing these packages to
 188 be one of potentially hundreds of packages being upgraded concurrently may
 189 not satisfy the requirements of the package or the site. (Watchdog, for
 190 example, if not restarted quickly, may cause the machine to reboot in the
 191 midst of installation, which may cause havoc on the machine)
 192 </para>
 193 </listitem>
 194 </orderedlist>
 195 </chapter>
 196
 197 <chapter id="ch3"><title>Procedural description</title>
 198 <variablelist>
 199 <varlistentry>
 200 <term>Set Options</term>
 201 <listitem>
 202 <para>
 203 This process handles setting of user or site options, and configuration of all
 204 aspects of APT. It allows the user to set the location and order of package
 205 sources, allowing them to set up source list details, like ftp site locations,
 206 passwords, etc. Display options may also be set.
 207 </para>
 208 </listitem>
 209 </varlistentry>
 210 <varlistentry>
 211 <term>Updates</term>
 212 <listitem>
 213 <para>
 214 Build a list of available packages, using source lists or a base location and
 215 trawling for Packages files (needs to be aware of architecture). This may
 216 involve finding and retrieving Packages files, storing them locally for
 217 efficiency, and parsing the data for later use. This would entail contacting
 218 various underlying access modules (ftp, cdrom mounts, etc) Use a backing store
 219 for speed. This may also require downloading the actual package files locally
 220 for speed.
 221 </para>
 222 </listitem>
 223 </varlistentry>
 224 <varlistentry>
 225 <term>Local status</term>
 226 <listitem>
 227 <para>
 228 Build up a list of packages already installed. This requires reading and
 229 writing the local??  status file. For remote installation, this should
 230 probably use similar mechanisms as the Packages file retrieval does. Use
 231 the backing store for speed. One should consider multiple backing stores,
 232 one for each machine.
 233 </para>
 234 </listitem>
 235 </varlistentry>
 236 <varlistentry>
 237 <term>Relationship determination</term>
 238 <listitem>
 239 <para>
 240 Determine forward and reverse dependencies. All known dependency fields should
 241 be acted upon, since it is fairly cheap to do so. Update the backing store
 242 with this information.
 243 </para>
 244 </listitem>
 245 </varlistentry>
 246 <varlistentry>
 247 <term>Selection</term>
 248 <listitem>
 249 <para>
 250 Present the data to the user. Look at Behan Webster's documentation for the
 251 user interface procedures. (Note: In the authors opinion deletions and reverse
 252 dependencies should also be presented to the user, in a strictly symmetric
 253 fashion; this may make it easier to prevent a package being removed that breaks
 254 dependencies)
 255 </para>
 256 </listitem>
 257 </varlistentry>
 258 <varlistentry>
 259 <term>Ordering of package installations and configuration</term>
 260 <listitem>
 261 <para>
 262 Build a list of events. Simple topological sorting gives order of packages
 263 in dependency order. At certain points in this ordering,
 264 predependencies/immediate configure directives cause an break in normal
 265 ordering. We need to insert the uninstall/purge directive in the stream
 266 (default: as early as possible).
 267 </para>
 268 </listitem>
 269 </varlistentry>
 270 <varlistentry>
 271 <term>Action</term>
 272 <listitem>
 273 <para>
 274 Take the order of installations and removals and build up a stream of events
 275 to send to the packaging system (dpkg). Execute the list of events if
 276 successful. Do not partially install packages and leave system in broken
 277 state. Go to The Selection step as needed.
 278 </para>
 279 </listitem>
 280 </varlistentry>
 281 </variablelist>
 282 </chapter>
 283
 284 <chapter id="ch4"><title>Modules and interfaces</title>
 285 <variablelist>
 286 <varlistentry>
 287 <term>The user interface module</term>
 288 <listitem>
 289 <para>
 290 Look at Behan Webster's documentation.
 291 </para>
 292 </listitem>
 293 </varlistentry>
 294 <varlistentry>
 295 <term>Widget set</term>
 296 <listitem>
 297 <para>
 298 Related closely to above Could some one present design decisions of the widget
 299 set here?
 300 </para>
 301 </listitem>
 302 </varlistentry>
 303 <varlistentry>
 304 <term>pdate Module</term>
 305 <listitem>
 306 <para>
 307 Distinct versions of the same package are recorded separately, but if multiple
 308 Packages files contain the same version of a package, then only the first one
 309 is recorded. For this reason, the least expensive update source should be
 310 listed first (local file system is better than a remote ftp site)
 311 </para>
 312 <para>
 313 This module should interact with the user interface module to set and change
 314 configuration parameters for the modules listed below. It needs to record that
 315 information in an on disk data file, to be read on future invocations.
 316 </para>
 317 <orderedlist numeration="arabic">
 318 <listitem>
 319 <para>
 320 FTP methods
 321 </para>
 322 </listitem>
 323 <listitem>
 324 <para>
 325 mount and file traversal module(s)?
 326 </para>
 327 </listitem>
 328 <listitem>
 329 <para>
 330 Other methods ???
 331 </para>
 332 </listitem>
 333 </orderedlist>
 334 </listitem>
 335 </varlistentry>
 336 <varlistentry>
 337 <term>Status file parser/generator</term>
 338 <listitem>
 339 <para>
 340 The status file records the current state of the system, listing the packages
 341 installed, etc. The status file is also one method of communicating with dpkg,
 342 since it is perfectly permissible for the user to use APT to request packages
 343 be updated, put others on hold, mark other for removal, etc, and then run
 344 <literal>dpkg -BORGiE</literal> on a file system.
 345 </para>
 346 </listitem>
 347 </varlistentry>
 348 <varlistentry>
 349 <term>Package file parser/generator</term>
 350 <listitem>
 351 <para>
 352 Related to above. Handle multiple Packages files, from different
 353 sources. Each package contains a link back to the packages file structure
 354 that contains details about the origin of the data.
 355 </para>
 356 </listitem>
 357 </varlistentry>
 358 <varlistentry>
 359 <term>Dependency module</term>
 360 <listitem>
 361 <itemizedlist>
 362 <listitem>
 363 <para>
 364 dependency/conflict determination and linking
 365 </para>
 366 </listitem>
 367 <listitem>
 368 <para>
 369 reverse dependency generator. Maybe merged with above
 370 </para>
 371 </listitem>
 372 </itemizedlist>
 373 </listitem>
 374 </varlistentry>
 375 <varlistentry>
 376 <term>Package ordering Module</term>
 377 <listitem>
 378 <para>
 379 Create an ordering of the actions to be taken.
 380 </para>
 381 </listitem>
 382 </varlistentry>
 383 <varlistentry>
 384 <term>Event generator</term>
 385 <listitem>
 386 <para>
 387 module to interact with dpkg
 388 </para>
 389 </listitem>
 390 </varlistentry>
 391 </variablelist>
 392 </chapter>
 393
 394 <chapter id="ch5"><title>Data flow and conversions analysis.</title>
 395 <screen>
 396                                                           ____________
 397                                                        __\|ftp modules|
 398                                                       /  /|___________|
 399                                     _ ____________   /     ________________
 400                                     |    update   | /     |mount/local file|
 401         |==========================&gt;|   module    |/_____\|  traversals    |
 402         |                           |_____________|      /|________________|
 403         |                             ^        ^
 404         |                             |        |               ______________
 405   ______|_______    _ _____ ______    |   _____v________      \|            |
 406  |Configuration |   |configuration|   |   |Packages Files|  ===|Status file |
 407  |  module      |&lt;=&gt;|    data     |   |   |______________| /  /|____________|
 408  |______________|   |_____________|   |        ^          /
 409          ^                            |        |         /
 410          |                            | _______v_______|/_
 411          |                            | |              |    ________________
 412          |                            | |              |/_\|   Dependency  |
 413          |                            | |backing store |\ /|     Module    |
 414          |                            | |______________|  _|_______________|
 415          |                             \       ^          /|       ^
 416          |                              \      |         /         |
 417          |                              _\|____v_______|/__    ____v_______
 418          |_____________________________\| User interaction|    |    dpkg   |
 419                                        /|_________________|&lt;==&gt;  Invoker  |
 420                                                                |___________|
 421 </screen>
 422 <para>
 423 dpkg also interacts with status and available files.
 424 </para>
 425 <para>
 426 The backing store and the associated data structures are the core of APT. All
 427 modules essentially revolve around the backing store, feeding it data, adding
 428 and manipulating links and relationships between data in the backing store,
 429 allowing the user to interact with and modify the data in the backing store,
 430 and finally writing it out as the status file and possibly issuing directives
 431 to dpkg.
 432 </para>
 433 <para>
 434 The other focal point for APT is the user interface.
 435 </para>
 436 </chapter>
 437
 438 </book>