Get Gentoo!
gentoo.org sites
gentoo.org Wiki Bugs Forums Packages
Planet Archives Sources
Infra Status
summaryrefslogtreecommitdiff
path: root/dependencies.tex
Commit message (Collapse)AuthorAgeFilesLines
* Use sentence case throughoutUlrich Müller2024-05-081-3/+3
| | | | | | | | | | | | Subsection headings were already changed from title case to sentence case in commit ca463ce. This left chapter and section headings alone, which is inconsistent (although it is specified by some style guides like APA). This commit changes headings to sentence case throughout, following most other Gentoo documentation, e.g. wiki and devmanual. Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* WhitespaceUlrich Müller2023-01-181-2/+2
| | | | Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Don't use teletype for EAPIsUlrich Müller2022-12-091-1/+1
| | | | Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* EAPI 8 has install-time dependencies (IDEPEND)Michał Górny2021-06-051-8/+17
| | | | | | | Bug: https://bugs.gentoo.org/660306 Signed-off-by: Michał Górny <mgorny@gentoo.org> [Updated as discussed in -pms mailing list] Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Add an EAPI 8 identical to EAPI 7Ulrich Müller2021-06-051-13/+13
| | | | Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Fix bad page breaks before listsUlrich Müller2021-05-211-2/+2
| | | | | | Taken from https://tex.stackexchange.com/a/2645. Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* pms.cls: Do not define \e for \emphUlrich Müller2021-04-301-2/+2
| | | | | | | | This is only used a few times, so a shorthand is not needed. (We really should get rid of \i and \t as well, because redefining LaTeX internal commands sucks.) Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Correct the definition of ESYSROOT as EPREFIX isn't always applicableJames Le Cuirot2020-06-281-1/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | It was originally envisaged (but not stated in PMS) that SYSROOT would only ever need to equal / or ROOT as a distinct SYSROOT would have no benefit. A check was added to Portage to ensure this held. Myself, the ChromiumOS team, and others have since been caught out by this check when trying to bootstrap brand new systems from scratch. You cannot bootstrap with no headers at all! The check will therefore be adjusted to merely ensure that SYSROOT is / when ROOT is /. There were differing assumptions about how prefixes applied to the above. EPREFIX is traditionally something the user sets so some thought that it would be applied to SYSROOT, regardless of the latter's value. In order to honor the rule about there being no distinct SYSROOT, this would mean that if SYSROOT is / then EPREFIX would have to match BROOT. Despite that limitation, ESYSROOT was written into PMS with a fixed value of ${SYSROOT}${EPREFIX}. Being somewhat unfamiliar with prefix at the time, I didn't realise that this view didn't align with what I'd had in mind and it was only when I came to need a distinct SYSROOT that I realised there was a problem. crossdev toolchains are installed to ${EPREFIX}/usr/${CHOST} but have no further prefix appended and packages subsequently installed with cross-emerge are placed in this location by setting ROOT. Bug #642604 recently revealed that the build system's prefix was being erroneously duplicated on the end but I have now fixed this. What if we want to bootstrap a brand new prefixed system using the crossdev system as SYSROOT? This is the distinct SYSROOT case. The problem is that there is no distinct variable for SYSROOT's prefix and, as already stated, ESYSROOT is always ${SYSROOT}${EPREFIX}. We therefore cannot do it! If the crossdev prefix is blank then ROOT's must be blank too. I also never intended to have the aforementioned limitation where EPREFIX must match BROOT when SYSROOT is /. These are both entirely artificial restrictions. So how should it work instead? We originally intended for SYSROOT to equal either / or ROOT so I imagined the prefix would automatically be adjusted to match the prefix applicable at the matching location, namely BROOT or EPREFIX. This is obviously more flexible than forcing it to match EPREFIX. What about the distinct SYSROOT case? With no distinct variable, we have no way to explicitly set a prefix but this is likely only needed when bootstrapping against crossdev systems, which are unprefixed by nature. We therefore simply assume that the prefix is blank in this case. What about the cross-prefix case? Here, SYSROOT matches both / and ROOT so which prefix do we choose? The bootstrap-prefix.sh script sets flags to build against the target prefix so EPREFIX is used in this case. This happens to fit the current definition of ESYSROOT anyway. Legitimate concerns have been raised about building for a system with a different prefix to the one you're building against. The only binaries that leak from SYSROOT to ROOT are static libraries. Headers from SYSROOT will obviously also influence how ROOT's binaries are built. It is entirely possible that SYSROOT's prefix may leak through a header but grepping /usr/include on my own main system reveals only a few paths from a small handful of packages. pkg-config files invariably include paths but these are almost always used at build time, not runtime. A differing prefix would likely only occur in cases involving core packages like the libc and kernel headers anyway. Also consider that we have never prevented this from happening in the past. It has always been possible to do "EPREFIX=/foo emerge bar" from some system with a different prefix or no prefix at all. All we're doing here is including the prefix (if any) in the ESYSROOT variable. Should this warrant a new EAPI? I don't think so. All existing usage of ESYSROOT that I have seen still fits with this new definition and most of that usage has come from me. We're not even changing what the variable is used for, just loosening the constraints around what it can be set to. If you have doubts about whether this makes sense or actually works in practise, I have experimented with a prefixed system using all the different combinations I could think of, including cross-compiling, and it all worked as expected. Keep in mind that ESYSROOT is not magic and currently isn't used very much. As such, neither the toolchain nor pkg-config will use this sysroot if you don't explicitly tell them to. For the former, I find CC="${CHOST}-gcc --sysroot=${ESYSROOT}" works well. For the latter, crossdev installs a cross-pkg-config wrapper but it is completely lacking prefix support at the moment. I have fixes waiting on this change. Signed-off-by: James Le Cuirot <chewi@gentoo.org> [Replaced "/" by "empty", reworded table cell in ebuild-env-vars.tex] Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* dependencies.tex: Don't mention blocks on provided virtuals.Ulrich Müller2020-01-051-7/+2
| | | | | | | | This is a remnant of old-style virtuals and should have been removed long ago. Fixes: c8ab6b99bffa85bcd686284ba60a30f53c31c8b0 Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* dependencies.tex: Reorder variables to match ebuild-vars section.Ulrich Müller2019-03-241-4/+4
| | | | Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Refer to chapters as chapters.Ulrich Müller2019-03-241-1/+1
| | | | | | | Also rename label prefixes, "ch:" for chapters, "sec:" for sections, as suggested by Michael Orlitzky. Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Promote "Package Dependency Specifications" to section.Ulrich Müller2019-03-241-5/+5
| | | | | | | | Its subsubsections "Operators", "Block operator", "Slot dependencies", and "2-style and 4-style USE dependencies" will become subsections, so the maximum section number depth of the document will be 2. Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Move some subsections out of the "Dependencies" chapter.Ulrich Müller2019-03-241-79/+0
| | | | | | | | | | SRC_URI, REQUIRED_USE, PROPERTIES, and RESTRICT are ebuild-defined variables. Move them to that chapter. Add reference to tab:uri-arrows-table in SRC_URI section. Otherwise, no change of wording. Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* dependencies: Remove the circular dep special case for RDEPENDMichał Górny2018-10-281-2/+1
| | | | | | | | | Remove the circular dependency note for RDEPEND. This is confusing at best, and there is no gain from claiming that non-standard, redundant Portage behavior is valid. Signed-off-by: Michał Górny <mgorny@gentoo.org> Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* dependencies: Permit {,B}DEPEND in source-initiated pkg_setupMichał Górny2018-10-281-4/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Account for build dependencies being satisfied in pkg_setup, when executed as part of source build, in order to match the realistic implementation and widespread use in Gentoo. This changes pkg_setup from the previous state of no dependencies being guaranteed. Firstly, this matches the standard package manager implementation better. Since pkg_setup is executed as part of the sequence of successive phases comprising a source build, it is natural that build dependencies are installed already when it's executed, and not between pkg_setup and src_unpack. Secondly, it accounts for widespread use of pkg_setup to determine installed dependencies in order to set the build environment appropriately. This is a case e.g. with python-any-r1 or llvm eclasses, which query installed packages in order to determine the Python or LLVM version to use, appropriately. This behavior can't be changed without major changes to a large number of ebuilds, and without making ebuild development harder. Technically, we could extend the same rule to RDEPEND being satisfied when package is going to be installed. However, this would restrict the flexibility of parallel builds. With regard to the current rules, the package manager can start building package from sources before all of its runtime dependencies are installed. The unclear benefit of having RDEPEND satisfied in pkg_setup does not seem to justify losing this possibility. Signed-off-by: Michał Górny <mgorny@gentoo.org> Signed-off-by: Ulrich Müller <ulm@gentoo.org>
* Fix whitespace in list items.Ulrich Müller2018-08-121-7/+7
|
* pkg-mgr-commands.tex: Join docompress and dostrip sections.Ulrich Müller2018-03-311-2/+2
|
* EAPI 7 has controllable stripping and dostrip.Ulrich Müller2018-03-311-1/+3
| | | | Bug: https://bugs.gentoo.org/203891
* dependencies: Provide a nice summary table for dep APIsMichał Górny2018-03-311-1/+17
|
* EAPI 7 has BDEPENDMichał Górny2018-03-311-3/+30
| | | | Bug: https://bugs.gentoo.org/317337
* EAPI 7: Empty || and ^^ groups no longer count as matched.Ulrich Müller2018-03-301-2/+19
| | | | Bug: https://bugs.gentoo.org/636596
* Add an EAPI 7 identical to EAPI 6.Ulrich Müller2018-03-301-10/+10
|
* pms.cls: Update \note command to suppress vertical space.Ulrich Müller2018-03-291-3/+3
| | | | Use the command correctly throughout; it requires an argument.
* dependencies.tex: Clarify wording for = slot operator.Ulrich Müller2018-03-291-1/+2
|
* dependencies.tex: Describe similar dep groups only once.Ulrich Müller2018-03-291-12/+5
| | | | | Any-of, exactly-one-of, and at-most-one-of groups have similar format, so there is no need to repeat the same description three times.
* Appendix: Document retroactive ban of empty dependency groups.Ulrich Müller2017-10-101-0/+1
| | | | Bug: https://bugs.gentoo.org/370565
* Ban empty groups in dependency specifications.Ulrich Müller2017-10-081-10/+10
| | | | | | | | | | | | | | | | The specification for empty groups was originally added to match historical Portage behaviour. However, Portage was changed in 2011 to treat empty dependency groups as an error, therefore ebuilds cannot rely on any definite behaviour. Consequently, empty groups are not used by any ebuild or eclass in the Gentoo repository. Note that the explicit handling for empty groups is still retained for the case of implicit empty groups resulting from collapsing USE conditionals in any-of groups et al. This is a retroactive change for all EAPIs. Bug: https://bugs.gentoo.org/370565
* Fix wasteful layout of dependency classes table.Ulrich Müller2017-09-291-2/+2
|
* dependencies: Add extra vert spacing in phase-dep tableMichał Górny2017-09-291-0/+3
| | | | | | | | | Add additional vertical space between rows of the 'Dependency classes required to be satisfied for a particular phase function' to improve readbility. For all but the last row, the phase function cell spans multiple lines while the dependency class cell is one line only. Without the extra spacing, it is not apparent at a first glance whether a particular phase function uses dependencies above or below it.
* Delete redundant sentence in at-most-one-of dependency.Ulrich Müller2017-08-301-2/+0
| | | | | | An empty group will automatically fulfill the condition of "at most one immediate child element", therefore no special provision for an empty group is necessary here.
* Use simple underscores throughout.Ulrich Müller2017-05-151-34/+34
| | | | | Replacement was done using "sed -i 's/\\_/_/g' *.tex". This does not change the resulting PDF and HTML output.
* More consistent capitalisation of headings.eapi-6-approved-2017-04-29Ulrich Müller2017-04-161-10/+10
| | | | | | | | Use title case for first and second level headings (i.e., \chapter and \section in the main text, \section and \subsection in the EAPI cheat sheet); use sentence case everywhere else. https://archives.gentoo.org/gentoo-pms/message/cef75e03f1a3b692281a5b79065dc2b6
* dependencies.tex: Explicitly require matching package for :=Michał Górny2016-10-311-0/+4
| | | | | | | | | Explicitly indicate that the ebuild is required to provide appropriate DEPEND for := to have a matching package installed. Additionally ban the cases where := can not be reliably used or does not make any sense: PDEPEND (because by intent the dependency can not be in DEPEND), and any-of groups (because the non-binding nature of any-of group collides with binding nature of :=).
* dependencies.tex: Record slot for := at build time, not install timeMichał Górny2016-10-311-1/+1
| | | | | | | | | | | Correct the wording for the 'equals' slot operator to indicate that the slot of the best installed package at the time package is *built* matters. After all, the common use case for := is to indicate that the build binds to a specific package version. This is also more in line with dependency classes where we clearly split the dependencies between build and runtime, and indicate that only DEPEND is guaranteed to be available before src_*.
* Remove unused labels.Ulrich Müller2016-09-131-1/+1
|
* Reformat EAPI tables to be more compact.Ulrich Müller2015-11-081-71/+52
| | | | | Combine all lines with identical behaviour for different EAPIs. Also use consistent indentation for tables in the LaTeX source.
* Add an EAPI 6 identical to EAPI 5.Ulrich Müller2015-11-081-5/+10
|
* Fix some of the "Underfull \hbox" warnings.Ulrich Müller2015-10-291-9/+10
|
* Typographic whitespace fixes.Ulrich Müller2015-10-151-2/+2
|
* The =* dependency operator can only match complete version components.Ulrich Müller2015-09-221-4/+5
| | | | Bug: 560466
* Use the same line width for Emacs and Vim.Ulrich Müller2015-04-111-0/+1
| | | | | | Lines seem to be wrapped at 100 characters throughout, with the exception of eapi-cheatsheet.tex which is wrapped at 70 characters. Assign the "fill-column" file local variable for Emacs accordingly.
* Typographic fixes to make chktex happier.Ulrich Müller2015-03-241-4/+4
|
* Drop RESTRICT=sandbox.Ulrich Müller2015-03-071-1/+0
| | | | | | | | RESTRICT=sandbox is neither used in the tree nor supported by Portage. Drop it retroactively for all EAPIs, bug 161045. The spec still says that "package managers may recognise other tokens" so package managers are free to support the feature.
* Whitespace: Split some very long lines.Ulrich Müller2015-02-181-11/+17
|
* Fix typo in at-most-one-of dependency specification.Ulrich Müller2013-10-011-1/+1
|
* Fix definition of tilde operator, bug 196561.Ulrich Müller2013-08-291-3/+1
|
* Fetch restriction implies mirror restriction.Ulrich Müller2013-04-261-1/+1
| | | | | See also ebuild(5): fetch - like mirror but the files will not be fetched via SRC_URI either.
* EAPI 5: Extend slot-operator-deps with sub-slot supportZac Medico2012-09-131-15/+25
| | | | See bug 424429.
* EAPI 5 supports ?? ( ) groupsCiaran McCreesh2012-09-131-0/+33
| | | | See bug 354219
* Slot operator dependencies are back in EAPI 5Ciaran McCreesh2012-09-131-5/+22
| | | | This reverts commit 409fccc10861c361f37a959195d7581a5c376dd9.