| Age | Commit message (Collapse) | Author |
|---|
|
In commit 446551c8 I changed MarkInstall to discard the candidate if the
candidate can't satisfy the dependency. This breaks interactive solvers
like aptitude which can change the candidate on-the-fly later.
In commit df77d8a5 I introduced this 'early' loop-breaking to begin with
which can't be that helpful for interactive solvers as well, but makes
perfect sense for non-interactives to stop them from exploring trees
which can't be satisfied, but it isn't perfect as ideally we would check
this before auto-installing the first dependency.
This commit therefore moves the loop into its own IsInstallOk hook so
that frontends can override this check if they want to and in exchange
removes the loop-breaking from MarkInstall itself and does it before any
dependency is installed.
Closes: 740750
|
|
Git-Dch: Ignore
Reported-By: gcc -Wsuggest-attribute={pure,const,noreturn}
|
|
Beside being a bit cleaner it hopefully also resolves oddball problems
I have with high levels of parallel jobs.
Git-Dch: Ignore
Reported-By: iwyu (include-what-you-use)
|
|
Reported-By: gcc -Wunused-parameter
Git-Dch: Ignore
|
|
Git-Dch: Ignore
|
|
|
|
We can't remove packages which are held back by the user with a hold, so
marking them (or its dependencies) as garbage will lead our autoremover
into madness – and given that the package is important enough that the
user has held it back it can't be garbage (at least at the moment), so
even if a front-end wants to use the info just for information display
its a good idea to not consider it garbage for them.
Closes: 724995
|
|
Multi-Arch: same packages can be co-installed, but need to have the same
version for all installed packages (aka "siblings"). Otherwise the
unsynced versions will fight against each other and the auto-install as
wel as the problem resolver will later have to decide between holding the
packages or to remove one of the siblings (usually a foreign) taking a
bunch of packages (like the entire foreign setup) with them.
The idea here is now to be more pro-active: MarkInstall will fail for
a package if the siblings aren't synced, so we don't allow a situation
in which a resolver has to decide if to hold or to remove-upgrade under
the assumption that the remove-upgrade decision is always wrong and
doesn't deserve to be explored (expect valid out-of-syncs of course).
Thats a pretty bold move to take for a library which is used by
different solvers so this check is done in IsInstallOk and can be
overridden if front-ends want to.
|
|
The breakage is just to big for now, so guard the change with
#ifndef APT_8_CLEANER_HEADERS and be nice to library users
|
|
|
|
|
|
|
|
fix a cosmetic warning from cppcheck:
[apt-pkg/depcache.h:462] -> [apt-pkg/depcache.h:122]:
(style) Variable 'Purge' hides enumerator with same name
|
|
depcache by caching the install-recommends and install-suggests
values
|
|
|
|
|
|
|
|
- use a boolean instead of an int for Add/Remove in AddStates
similar to how it works with AddSizes
|
|
pseudo-package related and/or private
|
|
to Keep which happens for example if a user decides to "remove" a not
installed package to forbid that it's part of the solution
|
|
based on a very early draft for EDSP by Stefano
APT can now write a scenario as well as load most stuff from it.
|
|
|
|
seperate IsModeChangeOk which checks sanity and dpkg holds
|
|
Multi-Arch; instead, Arch: all packages only satisfy dependencies for
the native arch, except where the Arch: all package is declared
Multi-Arch: foreign. (Closes: #613584)
This has the sideeffect that arch:all packages internally show up as
coming from the native arch - so packages with the architecture "all"
doesn't exist any longer in the pkgcache
|
|
- add SetCandidateRelease() to set a candidate version and
the candidates of dependencies if needed to a specified
release (Closes: #572709)
- change pkg/release behavior to use the new SetCandidateRelease
so installing packages from experimental or backports is easier
|
|
- do not check endpointer packages instead of only those which prevented
NeverAutoRemove settings from having an effect (Closes: #598452)
|
|
- fix autoremove by using correct config-option name and
don't make faulty assumptions in error handling (Closes: #594689)
|
|
- add another round of const& madness as the previous round accidently
NOT override the virtual GetCandidateVer() method (Closes: #587725)
|
|
* apt-pkg/contrib/configuration.{cc,h}:
- add a wrapper to match strings against configurable regex patterns
|
|
- SetCandidateVer for all pseudo packages
- SetReInstall for the "all" package of a pseudo package
|
|
- switch i{Download,Usr}Size from double to (un)signed long long
The biggest reason is that this saves a lot of float point operations
we do in AddSizes() on integers. The only reason i see that this was
a double is that it was 64bit long and can therefore store bigger
values than int/long, but with the availablity of (un)signed long long
we are now also at 64bit and can store sizes more than 8 Exabytes big -
by the time this will be a limit the C/C++ Standard will have bigger types,
hopefully.
|
|
boolean instead to handle the sizes more gracefully.
|
|
in his package choices
The previous implementation tried to install the package for arch A and
if this fails B, C and so on. This results in wrong architecture choices
for packages which depend on other pseudo packages, so he will now try
to install the dependencies first before trying the package itself and
only if this fails he tries the next architecture.
|
|
|
|
but we support the usage of the new ABI so libapt users
can start to prepare for MultiArch (Closes: #536029)
|
|
- remove Auto-Installed information from extended_states
together with the package itself (Closes: #572364)
|
|
installed reverse dependency and which also doesn't provide something.
They cause problems if this pseudo packages get new dependencies.
As a consequence we also need to recheck the dependencies of a killed
pseudo package (and especially the providers of these dependencies)
to really kill all non required packages.
|
|
packages in a group if one is marked.
The auto-installed flag is from now on Architecture bound:
A section without an architecture tag will be treated as applying
to all architectures - the next write operation will take care of
this by creating separate sections for the architectures.
|
|
Arch all packages are now represented by arch depending packages which
all depend on a package with the same name and the special arch "all".
This packages has NO dependencies, but beside this the same information.
It is the only package which has a size, the arch depending ones all
have a zero size. While the arch depending pseudo packages are used
for dependency resolution the arch "all" package is used for downloading
and ordering of the package.
|
|
can be sure that a package comeing in as a dependency of this package
will be of the same architecture as itself (or all). We don't want to break
this, so internal an arch all package is represented as many arch depending
packages. The only problem we have now is that we only know that a arch
all package is installed or not - we don't know for which architecture it
was installed: So we will look at all these broken arch all pseudo packages
and "remove" them.
|
|
they have no effect anyway (thanks Daniel Burrows for noticing)
* improve documentation of the Is{Delete,Install}Ok methods a bit
* add the FromUser boolean to the debug output of the markers
|
|
* honor the dpkg hold state in new Marker hooks (closes: #64141)
Combine the proposed AutoInstOk and IsAutoInstallOk to more general hooks
for MarkInstall (and another one for MarkDelete) with the same parameters
as the call these hooks should check.
|
|
|
|
|
|
type triggers a autoinst (closes: #458389)
* add debug::pkgDepCache::Marker with more detailed debug output
(closes: #87520)
|
|
|
|
Daniel Burrows, closes: #429378)
* fix FTFBFS by changing build-depends to
libcurl4-gnutls-dev (closes: #428363)
|
|
|
|
|
|
|
