1.. SPDX-License-Identifier: CC-BY-2.5
2
3=====================
4File Download Support
5=====================
6
7|
8
9BitBake's fetch module is a standalone piece of library code that deals
10with the intricacies of downloading source code and files from remote
11systems. Fetching source code is one of the cornerstones of building
12software. As such, this module forms an important part of BitBake.
13
14The current fetch module is called "fetch2" and refers to the fact that
15it is the second major version of the API. The original version is
16obsolete and has been removed from the codebase. Thus, in all cases,
17"fetch" refers to "fetch2" in this manual.
18
19The Download (Fetch)
20====================
21
22BitBake takes several steps when fetching source code or files. The
23fetcher codebase deals with two distinct processes in order: obtaining
24the files from somewhere (cached or otherwise) and then unpacking those
25files into a specific location and perhaps in a specific way. Getting
26and unpacking the files is often optionally followed by patching.
27Patching, however, is not covered by this module.
28
29The code to execute the first part of this process, a fetch, looks
30something like the following::
31
32   src_uri = (d.getVar('SRC_URI') or "").split()
33   fetcher = bb.fetch2.Fetch(src_uri, d)
34   fetcher.download()
35
36This code sets up an instance of the fetch class. The instance uses a
37space-separated list of URLs from the :term:`SRC_URI`
38variable and then calls the ``download`` method to download the files.
39
40The instantiation of the fetch class is usually followed by::
41
42   rootdir = l.getVar('WORKDIR')
43   fetcher.unpack(rootdir)
44
45This code unpacks the downloaded files to the specified by ``WORKDIR``.
46
47.. note::
48
49   For convenience, the naming in these examples matches the variables
50   used by OpenEmbedded. If you want to see the above code in action,
51   examine the OpenEmbedded class file ``base.bbclass``
52   .
53
54The :term:`SRC_URI` and ``WORKDIR`` variables are not hardcoded into the
55fetcher, since those fetcher methods can be (and are) called with
56different variable names. In OpenEmbedded for example, the shared state
57(sstate) code uses the fetch module to fetch the sstate files.
58
59When the ``download()`` method is called, BitBake tries to resolve the
60URLs by looking for source files in a specific search order:
61
62-  *Pre-mirror Sites:* BitBake first uses pre-mirrors to try and find
63   source files. These locations are defined using the
64   :term:`PREMIRRORS` variable.
65
66-  *Source URI:* If pre-mirrors fail, BitBake uses the original URL (e.g
67   from :term:`SRC_URI`).
68
69-  *Mirror Sites:* If fetch failures occur, BitBake next uses mirror
70   locations as defined by the :term:`MIRRORS` variable.
71
72For each URL passed to the fetcher, the fetcher calls the submodule that
73handles that particular URL type. This behavior can be the source of
74some confusion when you are providing URLs for the :term:`SRC_URI` variable.
75Consider the following two URLs::
76
77   https://git.yoctoproject.org/git/poky;protocol=git
78   git://git.yoctoproject.org/git/poky;protocol=http
79
80In the former case, the URL is passed to the ``wget`` fetcher, which does not
81understand "git". Therefore, the latter case is the correct form since the Git
82fetcher does know how to use HTTP as a transport.
83
84Here are some examples that show commonly used mirror definitions::
85
86   PREMIRRORS ?= "\
87      bzr://.*/.\*  http://somemirror.org/sources/ \
88      cvs://.*/.\*  http://somemirror.org/sources/ \
89      git://.*/.\*  http://somemirror.org/sources/ \
90      hg://.*/.\*   http://somemirror.org/sources/ \
91      osc://.*/.\*  http://somemirror.org/sources/ \
92      p4://.*/.\*   http://somemirror.org/sources/ \
93     svn://.*/.\*   http://somemirror.org/sources/"
94
95   MIRRORS =+ "\
96      ftp://.*/.\*   http://somemirror.org/sources/ \
97      http://.*/.\*  http://somemirror.org/sources/ \
98      https://.*/.\* http://somemirror.org/sources/"
99
100It is useful to note that BitBake
101supports cross-URLs. It is possible to mirror a Git repository on an
102HTTP server as a tarball. This is what the ``git://`` mapping in the
103previous example does.
104
105Since network accesses are slow, BitBake maintains a cache of files
106downloaded from the network. Any source files that are not local (i.e.
107downloaded from the Internet) are placed into the download directory,
108which is specified by the :term:`DL_DIR` variable.
109
110File integrity is of key importance for reproducing builds. For
111non-local archive downloads, the fetcher code can verify SHA-256 and MD5
112checksums to ensure the archives have been downloaded correctly. You can
113specify these checksums by using the :term:`SRC_URI` variable with the
114appropriate varflags as follows::
115
116   SRC_URI[md5sum] = "value"
117   SRC_URI[sha256sum] = "value"
118
119You can also specify the checksums as
120parameters on the :term:`SRC_URI` as shown below::
121
122  SRC_URI = "http://example.com/foobar.tar.bz2;md5sum=4a8e0f237e961fd7785d19d07fdb994d"
123
124If multiple URIs exist, you can specify the checksums either directly as
125in the previous example, or you can name the URLs. The following syntax
126shows how you name the URIs::
127
128   SRC_URI = "http://example.com/foobar.tar.bz2;name=foo"
129   SRC_URI[foo.md5sum] = 4a8e0f237e961fd7785d19d07fdb994d
130
131After a file has been downloaded and
132has had its checksum checked, a ".done" stamp is placed in :term:`DL_DIR`.
133BitBake uses this stamp during subsequent builds to avoid downloading or
134comparing a checksum for the file again.
135
136.. note::
137
138   It is assumed that local storage is safe from data corruption. If
139   this were not the case, there would be bigger issues to worry about.
140
141If :term:`BB_STRICT_CHECKSUM` is set, any
142download without a checksum triggers an error message. The
143:term:`BB_NO_NETWORK` variable can be used to
144make any attempted network access a fatal error, which is useful for
145checking that mirrors are complete as well as other things.
146
147If :term:`BB_CHECK_SSL_CERTS` is set to ``0`` then SSL certificate checking will
148be disabled. This variable defaults to ``1`` so SSL certificates are normally
149checked.
150
151.. _bb-the-unpack:
152
153The Unpack
154==========
155
156The unpack process usually immediately follows the download. For all
157URLs except Git URLs, BitBake uses the common ``unpack`` method.
158
159A number of parameters exist that you can specify within the URL to
160govern the behavior of the unpack stage:
161
162-  *unpack:* Controls whether the URL components are unpacked. If set to
163   "1", which is the default, the components are unpacked. If set to
164   "0", the unpack stage leaves the file alone. This parameter is useful
165   when you want an archive to be copied in and not be unpacked.
166
167-  *dos:* Applies to ``.zip`` and ``.jar`` files and specifies whether
168   to use DOS line ending conversion on text files.
169
170-  *striplevel:* Strip specified number of leading components (levels)
171   from file names on extraction
172
173-  *subdir:* Unpacks the specific URL to the specified subdirectory
174   within the root directory.
175
176The unpack call automatically decompresses and extracts files with ".Z",
177".z", ".gz", ".xz", ".zip", ".jar", ".ipk", ".rpm". ".srpm", ".deb" and
178".bz2" extensions as well as various combinations of tarball extensions.
179
180As mentioned, the Git fetcher has its own unpack method that is
181optimized to work with Git trees. Basically, this method works by
182cloning the tree into the final directory. The process is completed
183using references so that there is only one central copy of the Git
184metadata needed.
185
186.. _bb-fetchers:
187
188Fetchers
189========
190
191As mentioned earlier, the URL prefix determines which fetcher submodule
192BitBake uses. Each submodule can support different URL parameters, which
193are described in the following sections.
194
195.. _local-file-fetcher:
196
197Local file fetcher (``file://``)
198--------------------------------
199
200This submodule handles URLs that begin with ``file://``. The filename
201you specify within the URL can be either an absolute or relative path to
202a file. If the filename is relative, the contents of the
203:term:`FILESPATH` variable is used in the same way
204``PATH`` is used to find executables. If the file cannot be found, it is
205assumed that it is available in :term:`DL_DIR` by the
206time the ``download()`` method is called.
207
208If you specify a directory, the entire directory is unpacked.
209
210Here are a couple of example URLs, the first relative and the second
211absolute::
212
213   SRC_URI = "file://relativefile.patch"
214   SRC_URI = "file:///Users/ich/very_important_software"
215
216.. _http-ftp-fetcher:
217
218HTTP/FTP wget fetcher (``http://``, ``ftp://``, ``https://``)
219-------------------------------------------------------------
220
221This fetcher obtains files from web and FTP servers. Internally, the
222fetcher uses the wget utility.
223
224The executable and parameters used are specified by the
225``FETCHCMD_wget`` variable, which defaults to sensible values. The
226fetcher supports a parameter "downloadfilename" that allows the name of
227the downloaded file to be specified. Specifying the name of the
228downloaded file is useful for avoiding collisions in
229:term:`DL_DIR` when dealing with multiple files that
230have the same name.
231
232If a username and password are specified in the ``SRC_URI``, a Basic
233Authorization header will be added to each request, including across redirects.
234To instead limit the Authorization header to the first request, add
235"redirectauth=0" to the list of parameters.
236
237Some example URLs are as follows::
238
239   SRC_URI = "http://oe.handhelds.org/not_there.aac"
240   SRC_URI = "ftp://oe.handhelds.org/not_there_as_well.aac"
241   SRC_URI = "ftp://you@oe.handhelds.org/home/you/secret.plan"
242
243.. note::
244
245   Because URL parameters are delimited by semi-colons, this can
246   introduce ambiguity when parsing URLs that also contain semi-colons,
247   for example::
248
249           SRC_URI = "http://abc123.org/git/?p=gcc/gcc.git;a=snapshot;h=a5dd47"
250
251
252   Such URLs should should be modified by replacing semi-colons with '&'
253   characters::
254
255           SRC_URI = "http://abc123.org/git/?p=gcc/gcc.git&a=snapshot&h=a5dd47"
256
257
258   In most cases this should work. Treating semi-colons and '&' in
259   queries identically is recommended by the World Wide Web Consortium
260   (W3C). Note that due to the nature of the URL, you may have to
261   specify the name of the downloaded file as well::
262
263           SRC_URI = "http://abc123.org/git/?p=gcc/gcc.git&a=snapshot&h=a5dd47;downloadfilename=myfile.bz2"
264
265
266.. _cvs-fetcher:
267
268CVS fetcher (``(cvs://``)
269-------------------------
270
271This submodule handles checking out files from the CVS version control
272system. You can configure it using a number of different variables:
273
274-  :term:`FETCHCMD_cvs <FETCHCMD>`: The name of the executable to use when running
275   the ``cvs`` command. This name is usually "cvs".
276
277-  :term:`SRCDATE`: The date to use when fetching the CVS source code. A
278   special value of "now" causes the checkout to be updated on every
279   build.
280
281-  :term:`CVSDIR`: Specifies where a temporary
282   checkout is saved. The location is often ``DL_DIR/cvs``.
283
284-  CVS_PROXY_HOST: The name to use as a "proxy=" parameter to the
285   ``cvs`` command.
286
287-  CVS_PROXY_PORT: The port number to use as a "proxyport="
288   parameter to the ``cvs`` command.
289
290As well as the standard username and password URL syntax, you can also
291configure the fetcher with various URL parameters:
292
293The supported parameters are as follows:
294
295-  *"method":* The protocol over which to communicate with the CVS
296   server. By default, this protocol is "pserver". If "method" is set to
297   "ext", BitBake examines the "rsh" parameter and sets ``CVS_RSH``. You
298   can use "dir" for local directories.
299
300-  *"module":* Specifies the module to check out. You must supply this
301   parameter.
302
303-  *"tag":* Describes which CVS TAG should be used for the checkout. By
304   default, the TAG is empty.
305
306-  *"date":* Specifies a date. If no "date" is specified, the
307   :term:`SRCDATE` of the configuration is used to
308   checkout a specific date. The special value of "now" causes the
309   checkout to be updated on every build.
310
311-  *"localdir":* Used to rename the module. Effectively, you are
312   renaming the output directory to which the module is unpacked. You
313   are forcing the module into a special directory relative to
314   :term:`CVSDIR`.
315
316-  *"rsh":* Used in conjunction with the "method" parameter.
317
318-  *"scmdata":* Causes the CVS metadata to be maintained in the tarball
319   the fetcher creates when set to "keep". The tarball is expanded into
320   the work directory. By default, the CVS metadata is removed.
321
322-  *"fullpath":* Controls whether the resulting checkout is at the
323   module level, which is the default, or is at deeper paths.
324
325-  *"norecurse":* Causes the fetcher to only checkout the specified
326   directory with no recurse into any subdirectories.
327
328-  *"port":* The port to which the CVS server connects.
329
330Some example URLs are as follows::
331
332   SRC_URI = "cvs://CVSROOT;module=mymodule;tag=some-version;method=ext"
333   SRC_URI = "cvs://CVSROOT;module=mymodule;date=20060126;localdir=usethat"
334
335.. _svn-fetcher:
336
337Subversion (SVN) Fetcher (``svn://``)
338-------------------------------------
339
340This fetcher submodule fetches code from the Subversion source control
341system. The executable used is specified by ``FETCHCMD_svn``, which
342defaults to "svn". The fetcher's temporary working directory is set by
343:term:`SVNDIR`, which is usually ``DL_DIR/svn``.
344
345The supported parameters are as follows:
346
347-  *"module":* The name of the svn module to checkout. You must provide
348   this parameter. You can think of this parameter as the top-level
349   directory of the repository data you want.
350
351-  *"path_spec":* A specific directory in which to checkout the
352   specified svn module.
353
354-  *"protocol":* The protocol to use, which defaults to "svn". If
355   "protocol" is set to "svn+ssh", the "ssh" parameter is also used.
356
357-  *"rev":* The revision of the source code to checkout.
358
359-  *"scmdata":* Causes the ".svn" directories to be available during
360   compile-time when set to "keep". By default, these directories are
361   removed.
362
363-  *"ssh":* An optional parameter used when "protocol" is set to
364   "svn+ssh". You can use this parameter to specify the ssh program used
365   by svn.
366
367-  *"transportuser":* When required, sets the username for the
368   transport. By default, this parameter is empty. The transport
369   username is different than the username used in the main URL, which
370   is passed to the subversion command.
371
372Following are three examples using svn::
373
374   SRC_URI = "svn://myrepos/proj1;module=vip;protocol=http;rev=667"
375   SRC_URI = "svn://myrepos/proj1;module=opie;protocol=svn+ssh"
376   SRC_URI = "svn://myrepos/proj1;module=trunk;protocol=http;path_spec=${MY_DIR}/proj1"
377
378.. _git-fetcher:
379
380Git Fetcher (``git://``)
381------------------------
382
383This fetcher submodule fetches code from the Git source control system.
384The fetcher works by creating a bare clone of the remote into
385:term:`GITDIR`, which is usually ``DL_DIR/git2``. This
386bare clone is then cloned into the work directory during the unpack
387stage when a specific tree is checked out. This is done using alternates
388and by reference to minimize the amount of duplicate data on the disk
389and make the unpack process fast. The executable used can be set with
390``FETCHCMD_git``.
391
392This fetcher supports the following parameters:
393
394-  *"protocol":* The protocol used to fetch the files. The default is
395   "git" when a hostname is set. If a hostname is not set, the Git
396   protocol is "file". You can also use "http", "https", "ssh" and
397   "rsync".
398
399   .. note::
400
401     When ``protocol`` is "ssh", the URL expected in :term:`SRC_URI` differs
402     from the one that is typically passed to ``git clone`` command and provided
403     by the Git server to fetch from. For example, the URL returned by GitLab
404     server for ``mesa`` when cloning over SSH is
405     ``git@gitlab.freedesktop.org:mesa/mesa.git``, however the expected URL in
406     :term:`SRC_URI` is the following::
407
408       SRC_URI = "git://git@gitlab.freedesktop.org/mesa/mesa.git;branch=main;protocol=ssh;..."
409
410     Note the ``:`` character changed for a ``/`` before the path to the project.
411
412-  *"nocheckout":* Tells the fetcher to not checkout source code when
413   unpacking when set to "1". Set this option for the URL where there is
414   a custom routine to checkout code. The default is "0".
415
416-  *"rebaseable":* Indicates that the upstream Git repository can be
417   rebased. You should set this parameter to "1" if revisions can become
418   detached from branches. In this case, the source mirror tarball is
419   done per revision, which has a loss of efficiency. Rebasing the
420   upstream Git repository could cause the current revision to disappear
421   from the upstream repository. This option reminds the fetcher to
422   preserve the local cache carefully for future use. The default value
423   for this parameter is "0".
424
425-  *"nobranch":* Tells the fetcher to not check the SHA validation for
426   the branch when set to "1". The default is "0". Set this option for
427   the recipe that refers to the commit that is valid for any namespace
428   (branch, tag, ...) instead of the branch.
429
430-  *"bareclone":* Tells the fetcher to clone a bare clone into the
431   destination directory without checking out a working tree. Only the
432   raw Git metadata is provided. This parameter implies the "nocheckout"
433   parameter as well.
434
435-  *"branch":* The branch(es) of the Git tree to clone. Unless
436   "nobranch" is set to "1", this is a mandatory parameter. The number of
437   branch parameters must match the number of name parameters.
438
439-  *"rev":* The revision to use for the checkout. The default is
440   "master".
441
442-  *"tag":* Specifies a tag to use for the checkout. To correctly
443   resolve tags, BitBake must access the network. For that reason, tags
444   are often not used. As far as Git is concerned, the "tag" parameter
445   behaves effectively the same as the "rev" parameter.
446
447-  *"subpath":* Limits the checkout to a specific subpath of the tree.
448   By default, the whole tree is checked out.
449
450-  *"destsuffix":* The name of the path in which to place the checkout.
451   By default, the path is ``git/``.
452
453-  *"usehead":* Enables local ``git://`` URLs to use the current branch
454   HEAD as the revision for use with ``AUTOREV``. The "usehead"
455   parameter implies no branch and only works when the transfer protocol
456   is ``file://``.
457
458Here are some example URLs::
459
460   SRC_URI = "git://github.com/fronteed/icheck.git;protocol=https;branch=${PV};tag=${PV}"
461   SRC_URI = "git://github.com/asciidoc/asciidoc-py;protocol=https;branch=main"
462   SRC_URI = "git://git@gitlab.freedesktop.org/mesa/mesa.git;branch=main;protocol=ssh;..."
463
464.. note::
465
466   When using ``git`` as the fetcher of the main source code of your software,
467   ``S`` should be set accordingly::
468
469       S = "${WORKDIR}/git"
470
471.. note::
472
473   Specifying passwords directly in ``git://`` urls is not supported.
474   There are several reasons: :term:`SRC_URI` is often written out to logs and
475   other places, and that could easily leak passwords; it is also all too
476   easy to share metadata without removing passwords. SSH keys, ``~/.netrc``
477   and ``~/.ssh/config`` files can be used as alternatives.
478
479Using tags with the git fetcher may cause surprising behaviour. Bitbake needs to
480resolve the tag to a specific revision and to do that, it has to connect to and use
481the upstream repository. This is because the revision the tags point at can change and
482we've seen cases of this happening in well known public repositories. This can mean
483many more network connections than expected and recipes may be reparsed at every build.
484Source mirrors will also be bypassed as the upstream repository is the only source
485of truth to resolve the revision accurately. For these reasons, whilst the fetcher
486can support tags, we recommend being specific about revisions in recipes.
487
488.. _gitsm-fetcher:
489
490Git Submodule Fetcher (``gitsm://``)
491------------------------------------
492
493This fetcher submodule inherits from the :ref:`Git
494fetcher<bitbake-user-manual/bitbake-user-manual-fetching:git fetcher
495(\`\`git://\`\`)>` and extends that fetcher's behavior by fetching a
496repository's submodules. :term:`SRC_URI` is passed to the Git fetcher as
497described in the :ref:`bitbake-user-manual/bitbake-user-manual-fetching:git
498fetcher (\`\`git://\`\`)` section.
499
500.. note::
501
502   You must clean a recipe when switching between '``git://``' and
503   '``gitsm://``' URLs.
504
505   The Git Submodules fetcher is not a complete fetcher implementation.
506   The fetcher has known issues where it does not use the normal source
507   mirroring infrastructure properly. Further, the submodule sources it
508   fetches are not visible to the licensing and source archiving
509   infrastructures.
510
511.. _clearcase-fetcher:
512
513ClearCase Fetcher (``ccrc://``)
514-------------------------------
515
516This fetcher submodule fetches code from a
517`ClearCase <http://en.wikipedia.org/wiki/Rational_ClearCase>`__
518repository.
519
520To use this fetcher, make sure your recipe has proper
521:term:`SRC_URI`, :term:`SRCREV`, and
522:term:`PV` settings. Here is an example::
523
524   SRC_URI = "ccrc://cc.example.org/ccrc;vob=/example_vob;module=/example_module"
525   SRCREV = "EXAMPLE_CLEARCASE_TAG"
526   PV = "${@d.getVar("SRCREV", False).replace("/", "+")}"
527
528The fetcher uses the ``rcleartool`` or
529``cleartool`` remote client, depending on which one is available.
530
531Following are options for the :term:`SRC_URI` statement:
532
533-  *vob*: The name, which must include the prepending "/" character,
534   of the ClearCase VOB. This option is required.
535
536-  *module*: The module, which must include the prepending "/"
537   character, in the selected VOB.
538
539   .. note::
540
541      The module and vob options are combined to create the load rule in the
542      view config spec. As an example, consider the vob and module values from
543      the SRC_URI statement at the start of this section. Combining those values
544      results in the following::
545
546         load /example_vob/example_module
547
548-  *proto*: The protocol, which can be either ``http`` or ``https``.
549
550By default, the fetcher creates a configuration specification. If you
551want this specification written to an area other than the default, use
552the ``CCASE_CUSTOM_CONFIG_SPEC`` variable in your recipe to define where
553the specification is written.
554
555.. note::
556
557   the SRCREV loses its functionality if you specify this variable. However,
558   SRCREV is still used to label the archive after a fetch even though it does
559   not define what is fetched.
560
561Here are a couple of other behaviors worth mentioning:
562
563-  When using ``cleartool``, the login of ``cleartool`` is handled by
564   the system. The login require no special steps.
565
566-  In order to use ``rcleartool`` with authenticated users, an
567   "rcleartool login" is necessary before using the fetcher.
568
569.. _perforce-fetcher:
570
571Perforce Fetcher (``p4://``)
572----------------------------
573
574This fetcher submodule fetches code from the
575`Perforce <https://www.perforce.com/>`__ source control system. The
576executable used is specified by ``FETCHCMD_p4``, which defaults to "p4".
577The fetcher's temporary working directory is set by
578:term:`P4DIR`, which defaults to "DL_DIR/p4".
579The fetcher does not make use of a perforce client, instead it
580relies on ``p4 files`` to retrieve a list of
581files and ``p4 print`` to transfer the content
582of those files locally.
583
584To use this fetcher, make sure your recipe has proper
585:term:`SRC_URI`, :term:`SRCREV`, and
586:term:`PV` values. The p4 executable is able to use the
587config file defined by your system's ``P4CONFIG`` environment variable
588in order to define the Perforce server URL and port, username, and
589password if you do not wish to keep those values in a recipe itself. If
590you choose not to use ``P4CONFIG``, or to explicitly set variables that
591``P4CONFIG`` can contain, you can specify the ``P4PORT`` value, which is
592the server's URL and port number, and you can specify a username and
593password directly in your recipe within :term:`SRC_URI`.
594
595Here is an example that relies on ``P4CONFIG`` to specify the server URL
596and port, username, and password, and fetches the Head Revision::
597
598   SRC_URI = "p4://example-depot/main/source/..."
599   SRCREV = "${AUTOREV}"
600   PV = "p4-${SRCPV}"
601   S = "${WORKDIR}/p4"
602
603Here is an example that specifies the server URL and port, username, and
604password, and fetches a Revision based on a Label::
605
606   P4PORT = "tcp:p4server.example.net:1666"
607   SRC_URI = "p4://user:passwd@example-depot/main/source/..."
608   SRCREV = "release-1.0"
609   PV = "p4-${SRCPV}"
610   S = "${WORKDIR}/p4"
611
612.. note::
613
614   You should always set S to "${WORKDIR}/p4" in your recipe.
615
616By default, the fetcher strips the depot location from the local file paths. In
617the above example, the content of ``example-depot/main/source/`` will be placed
618in ``${WORKDIR}/p4``.  For situations where preserving parts of the remote depot
619paths locally is desirable, the fetcher supports two parameters:
620
621- *"module":*
622    The top-level depot location or directory to fetch. The value of this
623    parameter can also point to a single file within the depot, in which case
624    the local file path will include the module path.
625- *"remotepath":*
626    When used with the value "``keep``", the fetcher will mirror the full depot
627    paths locally for the specified location, even in combination with the
628    ``module`` parameter.
629
630Here is an example use of the the ``module`` parameter::
631
632   SRC_URI = "p4://user:passwd@example-depot/main;module=source/..."
633
634In this case, the content of the top-level directory ``source/`` will be fetched
635to ``${P4DIR}``, including the directory itself.  The top-level directory will
636be accesible at ``${P4DIR}/source/``.
637
638Here is an example use of the the ``remotepath`` parameter::
639
640   SRC_URI = "p4://user:passwd@example-depot/main;module=source/...;remotepath=keep"
641
642In this case, the content of the top-level directory ``source/`` will be fetched
643to ``${P4DIR}``, but the complete depot paths will be mirrored locally. The
644top-level directory will be accessible at
645``${P4DIR}/example-depot/main/source/``.
646
647.. _repo-fetcher:
648
649Repo Fetcher (``repo://``)
650--------------------------
651
652This fetcher submodule fetches code from ``google-repo`` source control
653system. The fetcher works by initiating and syncing sources of the
654repository into :term:`REPODIR`, which is usually
655``${DL_DIR}/repo``.
656
657This fetcher supports the following parameters:
658
659-  *"protocol":* Protocol to fetch the repository manifest (default:
660   git).
661
662-  *"branch":* Branch or tag of repository to get (default: master).
663
664-  *"manifest":* Name of the manifest file (default: ``default.xml``).
665
666Here are some example URLs::
667
668   SRC_URI = "repo://REPOROOT;protocol=git;branch=some_branch;manifest=my_manifest.xml"
669   SRC_URI = "repo://REPOROOT;protocol=file;branch=some_branch;manifest=my_manifest.xml"
670
671.. _az-fetcher:
672
673Az Fetcher (``az://``)
674--------------------------
675
676This submodule fetches data from an
677`Azure Storage account <https://docs.microsoft.com/en-us/azure/storage/>`__ ,
678it inherits its functionality from the HTTP wget fetcher, but modifies its
679behavior to accomodate the usage of a
680`Shared Access Signature (SAS) <https://docs.microsoft.com/en-us/azure/storage/common/storage-sas-overview>`__
681for non-public data.
682
683Such functionality is set by the variable:
684
685-  :term:`AZ_SAS`: The Azure Storage Shared Access Signature provides secure
686   delegate access to resources, if this variable is set, the Az Fetcher will
687   use it when fetching artifacts from the cloud.
688
689You can specify the AZ_SAS variable as shown below::
690
691   AZ_SAS = "se=2021-01-01&sp=r&sv=2018-11-09&sr=c&skoid=<skoid>&sig=<signature>"
692
693Here is an example URL::
694
695   SRC_URI = "az://<azure-storage-account>.blob.core.windows.net/<foo_container>/<bar_file>"
696
697It can also be used when setting mirrors definitions using the :term:`PREMIRRORS` variable.
698
699.. _gcp-fetcher:
700
701GCP Fetcher (``gs://``)
702--------------------------
703
704This submodule fetches data from a
705`Google Cloud Storage Bucket <https://cloud.google.com/storage/docs/buckets>`__.
706It uses the `Google Cloud Storage Python Client <https://cloud.google.com/python/docs/reference/storage/latest>`__
707to check the status of objects in the bucket and download them.
708The use of the Python client makes it substantially faster than using command
709line tools such as gsutil.
710
711The fetcher requires the Google Cloud Storage Python Client to be installed, along
712with the gsutil tool.
713
714The fetcher requires that the machine has valid credentials for accessing the
715chosen bucket. Instructions for authentication can be found in the
716`Google Cloud documentation <https://cloud.google.com/docs/authentication/provide-credentials-adc#local-dev>`__.
717
718If it used from the OpenEmbedded build system, the fetcher can be used for
719fetching sstate artifacts from a GCS bucket by specifying the
720``SSTATE_MIRRORS`` variable as shown below::
721
722   SSTATE_MIRRORS ?= "\
723       file://.* gs://<bucket name>/PATH \
724   "
725
726The fetcher can also be used in recipes::
727
728   SRC_URI = "gs://<bucket name>/<foo_container>/<bar_file>"
729
730However, the checksum of the file should be also be provided::
731
732   SRC_URI[sha256sum] = "<sha256 string>"
733
734.. _crate-fetcher:
735
736Crate Fetcher (``crate://``)
737----------------------------
738
739This submodule fetches code for
740`Rust language "crates" <https://doc.rust-lang.org/reference/glossary.html?highlight=crate#crate>`__
741corresponding to Rust libraries and programs to compile. Such crates are typically shared
742on https://crates.io/ but this fetcher supports other crate registries too.
743
744The format for the :term:`SRC_URI` setting must be::
745
746   SRC_URI = "crate://REGISTRY/NAME/VERSION"
747
748Here is an example URL::
749
750   SRC_URI = "crate://crates.io/glob/0.2.11"
751
752.. _npm-fetcher:
753
754NPM Fetcher (``npm://``)
755------------------------
756
757This submodule fetches source code from an
758`NPM <https://en.wikipedia.org/wiki/Npm_(software)>`__
759Javascript package registry.
760
761The format for the :term:`SRC_URI` setting must be::
762
763   SRC_URI = "npm://some.registry.url;ParameterA=xxx;ParameterB=xxx;..."
764
765This fetcher supports the following parameters:
766
767-  *"package":* The NPM package name. This is a mandatory parameter.
768
769-  *"version":* The NPM package version. This is a mandatory parameter.
770
771-  *"downloadfilename":* Specifies the filename used when storing the downloaded file.
772
773-  *"destsuffix":* Specifies the directory to use to unpack the package (default: ``npm``).
774
775Note that NPM fetcher only fetches the package source itself. The dependencies
776can be fetched through the `npmsw-fetcher`_.
777
778Here is an example URL with both fetchers::
779
780   SRC_URI = " \
781       npm://registry.npmjs.org/;package=cute-files;version=${PV} \
782       npmsw://${THISDIR}/${BPN}/npm-shrinkwrap.json \
783       "
784
785See :yocto_docs:`Creating Node Package Manager (NPM) Packages
786</dev-manual/packages.html#creating-node-package-manager-npm-packages>`
787in the Yocto Project manual for details about using
788:yocto_docs:`devtool <https://docs.yoctoproject.org/ref-manual/devtool-reference.html>`
789to automatically create a recipe from an NPM URL.
790
791.. _npmsw-fetcher:
792
793NPM shrinkwrap Fetcher (``npmsw://``)
794-------------------------------------
795
796This submodule fetches source code from an
797`NPM shrinkwrap <https://docs.npmjs.com/cli/v8/commands/npm-shrinkwrap>`__
798description file, which lists the dependencies
799of an NPM package while locking their versions.
800
801The format for the :term:`SRC_URI` setting must be::
802
803   SRC_URI = "npmsw://some.registry.url;ParameterA=xxx;ParameterB=xxx;..."
804
805This fetcher supports the following parameters:
806
807-  *"dev":* Set this parameter to ``1`` to install "devDependencies".
808
809-  *"destsuffix":* Specifies the directory to use to unpack the dependencies
810   (``${S}`` by default).
811
812Note that the shrinkwrap file can also be provided by the recipe for
813the package which has such dependencies, for example::
814
815   SRC_URI = " \
816       npm://registry.npmjs.org/;package=cute-files;version=${PV} \
817       npmsw://${THISDIR}/${BPN}/npm-shrinkwrap.json \
818       "
819
820Such a file can automatically be generated using
821:yocto_docs:`devtool <https://docs.yoctoproject.org/ref-manual/devtool-reference.html>`
822as described in the :yocto_docs:`Creating Node Package Manager (NPM) Packages
823</dev-manual/packages.html#creating-node-package-manager-npm-packages>`
824section of the Yocto Project.
825
826Other Fetchers
827--------------
828
829Fetch submodules also exist for the following:
830
831-  Bazaar (``bzr://``)
832
833-  Mercurial (``hg://``)
834
835-  OSC (``osc://``)
836
837-  S3 (``s3://``)
838
839-  Secure FTP (``sftp://``)
840
841-  Secure Shell (``ssh://``)
842
843-  Trees using Git Annex (``gitannex://``)
844
845No documentation currently exists for these lesser used fetcher
846submodules. However, you might find the code helpful and readable.
847
848Auto Revisions
849==============
850
851We need to document ``AUTOREV`` and :term:`SRCREV_FORMAT` here.
852