1Release 3.4 (honister)
2======================
3
4Migration notes for 3.4 (honister)
5----------------------------------
6
7This section provides migration information for moving to the Yocto
8Project 3.4 Release (codename "honister") from the prior release.
9
10Override syntax changes
11~~~~~~~~~~~~~~~~~~~~~~~
12
13In this release, the ``:`` character replaces the use of ``_`` to
14refer to an override, most commonly when making a conditional assignment
15of a variable. This means that an entry like::
16
17   SRC_URI_qemux86 = "file://somefile"
18
19now becomes::
20
21   SRC_URI:qemux86 = "file://somefile"
22
23since ``qemux86`` is an override. This applies to any use of override
24syntax, so the following::
25
26   SRC_URI_append = " file://somefile"
27   SRC_URI_append_qemux86 = " file://somefile2"
28   SRC_URI_remove_qemux86-64 = " file://somefile3"
29   SRC_URI_prepend_qemuarm = "file://somefile4 "
30   FILES_${PN}-ptest = "${bindir}/xyz"
31   IMAGE_CMD_tar = "tar"
32   BASE_LIB_tune-cortexa76 = "lib"
33   SRCREV_pn-bash = "abc"
34   BB_TASK_NICE_LEVEL_task-testimage = '0'
35
36would now become::
37
38   SRC_URI:append = " file://somefile"
39   SRC_URI:append:qemux86 = " file://somefile2"
40   SRC_URI:remove:qemux86-64 = " file://somefile3"
41   SRC_URI:prepend:qemuarm = "file://somefile4 "
42   FILES:${PN}-ptest = "${bindir}/xyz"
43   IMAGE_CMD:tar = "tar"
44   BASE_LIB:tune-cortexa76 = "lib"
45   SRCREV:pn-bash = "abc"
46   BB_TASK_NICE_LEVEL:task-testimage = '0'
47
48This also applies to
49:ref:`variable queries to the datastore <bitbake:bitbake-user-manual/bitbake-user-manual-metadata:functions for accessing datastore variables>`,
50for example using ``getVar`` and similar so ``d.getVar("RDEPENDS_${PN}")``
51becomes ``d.getVar("RDEPENDS:${PN}")``.
52
53Whilst some of these are fairly obvious such as :term:`MACHINE` and :term:`DISTRO`
54overrides, some are less obvious, for example the packaging variables such as
55:term:`RDEPENDS`, :term:`FILES` and so on taking package names (e.g. ``${PN}``,
56``${PN}-ptest``) as overrides. These overrides are not always in
57:term:`OVERRIDES` but applied conditionally in specific contexts
58such as packaging. ``task-<taskname>`` is another context specific override, the
59context being specific tasks in that case. Tune overrides are another special
60case where some code does use them as overrides but some does not. We plan to try
61and make the tune code use overrides more consistently in the future.
62
63There are some variables which do not use override syntax which include the
64suffix to variables in ``layer.conf`` files such as :term:`BBFILE_PATTERN`,
65:term:`SRCREV`\ ``_xxx`` where ``xxx`` is a name from :term:`SRC_URI` and
66:term:`PREFERRED_VERSION`\ ``_xxx``. In particular, ``layer.conf`` suffixes
67may be the same as a :term:`DISTRO` override causing some confusion. We do
68plan to try and improve consistency as these issues are identified.
69
70To help with migration of layers, a script has been provided in OE-Core.
71Once configured with the overrides used by a layer, this can be run as::
72
73   <oe-core>/scripts/contrib/convert-overrides.py <layerdir>
74
75.. note::
76
77   Please read the notes in the script as it isn't entirely automatic and it isn't
78   expected to handle every case. In particular, it needs to be told which overrides
79   the layer uses (usually machine and distro names/overrides) and the result should
80   be carefully checked since it can be a little enthusiastic and will convert
81   references to ``_append``, ``_remove`` and ``_prepend`` in function and variable
82   names.
83
84For reference, this conversion is important as it allows BitBake to more reliably
85determine what is an override and what is not, as underscores are also used in
86variable names without intending to be overrides. This should allow us to proceed
87with other syntax improvements and simplifications for usability. It also means
88BitBake no longer has to guess and maintain large lookup lists just in case
89e.g. ``functionname`` in ``my_functionname`` is an override, and thus should improve
90efficiency.
91
92New host dependencies
93~~~~~~~~~~~~~~~~~~~~~
94
95The ``lz4c``, ``pzstd`` and ``zstd`` commands are now required to be
96installed on the build host to support LZ4 and Zstandard compression
97functionality. These are typically provided by ``lz4`` and ``zstd``
98packages in most Linux distributions. Alternatively they are available
99as part of ``buildtools-tarball`` if your distribution does not provide
100them. For more information see
101:ref:`ref-manual/system-requirements:required packages for the build host`.
102
103Removed recipes
104~~~~~~~~~~~~~~~
105
106The following recipes have been removed in this release:
107
108- ``assimp``: problematic from a licensing perspective and no longer
109  needed by anything else
110- ``clutter-1.0``: legacy component moved to meta-gnome
111- ``clutter-gst-3.0``: legacy component moved to meta-gnome
112- ``clutter-gtk-1.0``: legacy component moved to meta-gnome
113- ``cogl-1.0``: legacy component moved to meta-gnome
114- ``core-image-clutter``: removed along with clutter
115- ``linux-yocto``: removed version 5.4 recipes (5.14 and 5.10 still
116  provided)
117- ``mklibs-native``: not actively tested and upstream mklibs still
118  requires Python 2
119- ``mx-1.0``: obsolete (last release 2012) and isn't used by anything in
120  any known layer
121- ``packagegroup-core-clutter``: removed along with clutter
122
123Removed classes
124~~~~~~~~~~~~~~~
125
126- ``clutter``: moved to meta-gnome along with clutter itself
127- ``image-mklibs``: not actively tested and upstream mklibs still
128  requires Python 2
129- ``meta``: no longer useful. Recipes that need to skip installing
130  packages should inherit ``nopackages`` instead.
131
132Prelinking disabled by default
133~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
134
135Recent tests have shown that prelinking works only when PIE is not
136enabled (see `here <https://rlbl.me/prelink-1>`__ and `here <https://rlbl.me/prelink-2>`__),
137and as PIE is both a desirable security feature, and the only
138configuration provided and tested by the Yocto Project, there is
139simply no sense in continuing to enable prelink.
140
141There's also a concern that no one is maintaining the code, and there
142are open bugs (including :yocto_bugs:`this serious one </show_bug.cgi?id=14429>`).
143Given that prelink does intricate address arithmetic and rewriting
144of binaries the best option is to disable the feature. It is recommended
145that you consider disabling this feature in your own configuration if
146it is currently enabled.
147
148Virtual runtime provides
149~~~~~~~~~~~~~~~~~~~~~~~~
150
151Recipes shouldn't use the ``virtual/`` string in :term:`RPROVIDES` and
152:term:`RDEPENDS` - it is confusing because ``virtual/`` has no special
153meaning in :term:`RPROVIDES` and :term:`RDEPENDS` (unlike in the
154corresponding build-time :term:`PROVIDES` and :term:`DEPENDS`).
155
156Tune files moved to architecture-specific directories
157~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
158
159The tune files found in ``conf/machine/include`` have now been moved
160into their respective architecture name directories under that same
161location; e.g. x86 tune files have moved into an ``x86`` subdirectory,
162MIPS tune files have moved into a ``mips`` subdirectory, etc.
163The ARM tunes have an extra level (``armv8a``, ``armv8m``, etc.) and
164some have been renamed to make them uniform with the rest of the tunes.
165See :yocto_git:`this commit </poky/commit/?id=1d381f21f5f13aa0c4e1a45683ed656ebeedd37d>`
166for reference.
167
168If you have any references to tune files (e.g. in custom machine
169configuration files) they will need to be updated.
170
171Extensible SDK host extension
172~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
173
174For a normal SDK, some layers append to :term:`TOOLCHAIN_HOST_TASK`
175unconditionally which is fine, until the eSDK tries to override the
176variable to its own values. Instead of installing packages specified
177in this variable it uses native recipes instead - a very different
178approach. This has led to confusing errors when binaries are added
179to the SDK but not relocated.
180
181To avoid these issues, a new :term:`TOOLCHAIN_HOST_TASK_ESDK` variable has
182been created. If you wish to extend what is installed in the host
183portion of the eSDK then you will now need to set this variable.
184
185Package/recipe splitting
186~~~~~~~~~~~~~~~~~~~~~~~~
187
188- ``perl-cross`` has been split out from the main ``perl`` recipe to
189  its own ``perlcross`` recipe for maintenance reasons. If you have
190  bbappends for the perl recipe then these may need extending.
191
192- The ``wayland`` recipe now packages its binaries in a
193  ``wayland-tools`` package rather than putting them into
194  ``wayland-dev``.
195
196- Xwayland has been split out of the xserver-xorg tree and thus is now
197  in its own ``xwayland`` recipe. If you need Xwayland in your image
198  then you may now need to add it explicitly.
199
200- The ``rpm`` package no longer has ``rpm-build`` in its :term:`RRECOMMENDS`;
201  if by chance  you still need rpm package building functionality in
202  your image and you have not already done so then you should add
203  ``rpm-build`` to your image explicitly.
204
205- The Python ``statistics`` standard module is now packaged in its own
206  ``python3-statistics`` package instead of ``python3-misc`` as
207  previously.
208
209Image / SDK generation changes
210~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
211
212- Recursive dependencies on the ``do_build`` task are now disabled when
213  building SDKs. These are generally not needed; in the unlikely event
214  that you do encounter problems then it will probably be as a result of
215  missing explicit dependencies that need to be added.
216
217- Errors during "complementary" package installation (e.g. for ``*-dbg``
218  and ``*-dev`` packages) during image construction are no longer
219  ignored. Historically some of these packages had installation problems,
220  that is no longer the case. In the unlikely event that you see errors
221  as a result, you will need to fix the installation/packaging issues.
222
223- When building an image, only packages that will be used in building
224  the image (i.e. the first entry in :term:`PACKAGE_CLASSES`) will be
225  produced if multiple package types are enabled (which is not a typical
226  configuration). If in your CI system you need to have the original
227  behaviour, use ``bitbake --runall build <target>``.
228
229- The ``-lic`` package is no longer automatically added to
230  :term:`RRECOMMENDS` for every other package when
231  :term:`LICENSE_CREATE_PACKAGE` is set to "1". If you wish all license
232  packages to be installed corresponding to packages in your image, then
233  you should instead add the new ``lic-pkgs`` feature to
234  :term:`IMAGE_FEATURES`.
235
236Miscellaneous
237~~~~~~~~~~~~~
238
239- Certificates are now properly checked when bitbake fetches sources
240  over HTTPS. If you receive errors as a result for your custom recipes,
241  you will need to use a mirror or address the issue with the operators
242  of the server in question.
243
244- ``avahi`` has had its GTK+ support disabled by default. If you wish to
245  re-enable it, set ``AVAHI_GTK = "gtk3"`` in a bbappend for the
246  ``avahi`` recipe or in your custom distro configuration file.
247
248- Setting the ``BUILD_REPRODUCIBLE_BINARIES`` variable to "0" no longer
249  uses a strangely old fallback date of April 2011, it instead disables
250  building reproducible binaries as you would logically expect.
251
252- Setting noexec/nostamp/fakeroot varflags to any value besides "1" will
253  now trigger a warning. These should be either set to "1" to enable, or
254  not set at all to disable.
255
256- The previously deprecated ``COMPRESS_CMD`` and
257  ``CVE_CHECK_CVE_WHITELIST`` variables have been removed. Use
258  ``CONVERSION_CMD`` and ``CVE_CHECK_WHITELIST`` (replaced by
259  :term:`CVE_CHECK_IGNORE` in version 3.5) respectively
260  instead.
261
262- The obsolete ``oe_machinstall`` function previously provided in the
263  :ref:`utils <ref-classes-utils>` class has been removed. For
264  machine-specific installation it is recommended that you use the
265  built-in override support in the fetcher or overrides in general
266  instead.
267
268.. include:: release-notes-3.4.rst
269.. include:: release-notes-3.4.1.rst
270.. include:: release-notes-3.4.2.rst
271