To know what AUR packages need updating a rpc request is needed for all
packages. The dep tree is designed to cache everything to minimize the
amount of rpc requests. The downside of this is the dep tree ends up
with all sorts of packages in cache that it doesn't need. Then the
deptree tries to resolve deps for all of thoes packages.
By spliting the sysupgrade from the dep tree this stops this from
happening, it uses one more rpc request but also may lower the amount of
total rpc requests needed lated on.
This fixes a couple of tiny bugs such as triggering providers prompts
and printing AUR out of date messages for packages that are not going
to be installed.
This also fixes another display bug where repo packages from -Su would
not apear when printing the packages to be installed under [Repo].
--redownload is meant to only download the targets the user provides.
If the user enters aur/foo then Yay will find the package foo from the
aur, --redownload will see thats not what the user entered and skips the
download.
This makes it so after the dep searching is done, all db/ prefixes are
dropped.
This commit extends the conflict checking a lot, it adds support for:
Conflicting with provides as well as actual package names
Reverse conflicts
Inner conflicts
Both normal conflicts and inner conflicts are run in parallel.
Messages are now printing when checking conflicts.
This also fixes packages sometimes being listed as conflicting with
themselves.
The inner conflict is a little verbose and could be toned down a little
but I am insure exactly how to tone it down.
Previosly during `yay -Su` Yay would pass
`pacman -S <packages that need upgrade>` to pacman.
Instead pass `pacman -Su --ignore <number menu choices>`
This allows yay to handle replaces and package downgrades `-Suu`
When pkgbuilds are built by makepkg, if the pkgbuild's arch=() array
does not include the current carch set in makepkg.conf, makepkg will
fail to build the package.
Now, Yay detects if a pkgbuild does not support the arch set in
pacman.conf Yay will ask the user about this and ask them if they want
to build anyway, passing `--ignorearch` to makepkg.`
Note that Yay will check against the arch set in pacman.conf which is
what pacman uses to only allow installs of package of that arch. makepkg
will still use carch set in makepkg.conf to build packages. These two
values are expected to be the same otherwise Yay will fail.
The on disk .srcinfo is needed for this as the user should be asked pre
source download. This and pgp checking both use the on disk .srcinfo so
it is no longer a one off. Store the 'stale' srcinfos so they can be
accesed by both functions.
Fix typo where adding to has instead of depStrings
Error correcly when missing packages
Also handle cases where a package is provided multiple times. If one
package provies `foo=1` and another provides `foo=2` before the latter
would just overide the former version. Now both versions will be checked
against.
With the addition of pgp key checking in Yay, the srcinfo parsing was
moved to before the pkgver() bump, leading to outdated pkgbuild
information.
Srcinfo parsing must be done after the pkgver() bump
The pkgver() bump must be done after downloading sources
makepkg's PGP checking is done as the sources download
yays PGP importing requires the srcingo to be parsed
Quite the chicken and egg problem
It is possible to skip the integ checks after the sources download
then parse the srcinfo
do the yay PGP check
then run the integ checks
the problem here is that to generate the srcinfo `makepkg --printsrcingo` is ran
This causes the pkgbuild to be sourced which I am not comftable with
doing without the integ checks.
Instead we parse the on disk .SRRCINFO that downloads with the PKGBUILD
just for the PGP checking. Later on we parse a fresh srcinfo straight
from `makepkg --printsrcingo`. This is a little bit less efficient but
more secure than the other option.
The callback is set to allways silently say yes, When passing to pacman
for the intall pacman will then ask the question giving the user
a chance to answer.
Similar to the --redownload flag, when specifed targets will be rebuilt
even if an up to date version is cached. --rebuildall can be used to
ensure uninstalled dependencies are rebuilt as well.
Additionally, unlike --redownload there is also --rebuildtree. This
causes a rebuild and reinstall of a package and all of it's dependencies
recursivley. This is designed for when a libary updae, breaks an
installed AUR package due to a partial upgrade. polybar is a common
example
--rebuild allows you to easily skip the cache and rebuild against a newer
libary version. --rebuildtree is a more nuclear option where you can
rebuild the whole dependency tree.
When building a package from the AUR for which there are missing keys,
yay will now prompt the user whether it should try to import such keys
using gpg:
[...]
:: Parsing SRCINFO (1/3): libc++ (libc++abi libc++)
:: Parsing SRCINFO (2/3): aurutils
:: Parsing SRCINFO (3/3): cower
==> GPG keys need importing:
487EACC08557AD082088DABA1EB2638FF56C0C53, required by: cower
11E521D646982372EB577A1F8F0871F202119294, required by: libc++ (libc++abi libc++)
B6C8F98282B944E3B0D5C2530FC3042E345AD05D, required by: libc++ (libc++abi libc++)
DBE7D3DD8C81D58D0A13D0E76BC26A17B9B7018A, required by: aurutils
==> Import? [Y/n]
[...]
Default is to try to import the problematic keys ([Y/n]).
Adds the ability to pick which database to install a package from. This
is extended to also support for AUR packages. For example `extra/git`
and `aur/yay` should both work`. When not explicitly requesting
a database repo packages will be choosen over the AUR.
This features extends to yogurt mode, listings where a package shows up
in multiple database/the AUR is now handled.
The aur does not have a real pacman databse like core, extra ect. But
can be accessed as if was one with `aur/name`. Using Yay with a pacman
repository named "aur" is undefined.
If a pkgbuild is already in cache and matches the version on
the aur skip the download.
The version we check comes from the .SRCINFO file on disk which is never
updated. (updates through pkgver() edit the pkgbuild but do not effect
the .SRCINFO). Therefore if the the version of the .SRCINFO matches the
AUR's version there must not be an update.
In the case of the on disk version being newer than the AUR version we
can assume user interaction and they probably do not want it overwitten
so in that case also skip the download.
Use the command `git ls-remote <url> <branch>` to track devel updates
rather than relying on the GitHub API.
This allows devel update to work for every git based source and
elimantes the rate limiting from GitHub.
The yay_vcs.json format has changed to better support packages which
multiple vcs sources and to track the protocols each source uses. And
track the branch that each source tracks in it's fragment.
In a very specific case where the user runs `yay -Syu` then uses the
number menu to ignore all AUR upgrades after the Repo install the user
will still be prompted to install and download packages.
Add ( and ) to "installed" to match groups and out of date
Show popularity as well as voted in yogurt mode
Show download and install size in yogurt mode
Remove printing of white and black
Fix incorrect message on number menu
Yellow is now almost never used
Use `v` instead of `r` when printing version
show when a page was marked out of date on search and info
Previously Installing repo targets and repo dependencies of aur targets
was done in two steps.
doing `yay -S repo1 repo2 aur1 aur2` would lead to yay calling
`pacman -S repo1 repo2`. Then after we find all the dependencies of the
aur packages we call `pacman -S --asdeps <all repo dependencies>`. This
was an easy way to correctly mark dependencies. Since 005635b4 Both
these calls were merged into one command but dependency marking was
forgoten about.
Now correctly mark the dependencies through `pacman -D`
This allows the user to abort the install by telling the editor to exit
with a non zero status. e.g. `:cq` in vim. This should also catch errors
if the editor does actually fail or if the configured editor does not
exist.
Save the VSC Info as soon as the package install finishes. This should
ensure the VSC db does not end up in an incorrect state if an install
fails or is cancelled by the user.
This also adds better support for split packages. When one or more
packages are installed from the same base each individual package is
added to the db not just the base. This allows us to track individual
updates from the same base so that if one package gets updated we don't
assume all packages in the base are updated.
The -s was kept aroung because the package base `python-virtualfish`
failed to build without it. I first blamed this on some aur rpc bug
because it was missing some deps that were listed in python-virtualfish.
As is turned out python-virtualfish actually does some things wrong in
it's package build and if it was formated correctly this wouldnt be
a problem.
I kept the -s in just so it would build even though it might have had
some side affects. makepkg not respecting the --dbpath for example.
From now on python-virtualfish will fail to build when you dont have all
the deps installed. This is their fault and will not be fixed here.
Exit after parsing srcinfo with db do it actually does something.
when using gendb dont bother generating srcinfos.
Improved the formatting for some things mainly downloading and parsing.
This commit mostly focuses on merging the install() and upgradePkgs()
functions to be more integrated. Instead of both making their own aur
queries they now both combine theyre needes into one query. This gives
us a speed up and allows us to easily print orphands, misising and out
of date in one clean block.
Renable conflict checking, was disabled for testing, forgot to reenable
Use pacman --ask to remove conflicting packages, dont do it directly.
Say what each package conflicts with
Check for conflicts before aking for clean build
Reverse conflict are still not checked
Split packages are now grouped together when printing displaying the
package base and the packages inside of the base to beinstalled. If only
one packge is to be installed from a base and the package name matches
the base name print normally
Only build and install once per package base
Only ask questions once per package base for editing pkgbuils and clean
build
Install all deps for aur packages after the user confirms they want to
continue installing. This takes most of the load off of makepkg -s but
the -s is still left in for some edge cases with split packages.
When installing a package might be a dependancy for something and a make
dependancy for something else. This means when prompted to remove make
dependencies yay might also try to remove a package that is actually
needed causing a pacman error.
Adding the -u option will cause pacman to skip needed packages and give
a nice warning as it does so. It does not fix the root issue but works
for now.
I have replaced the old install and dependancy algorithms with a new
design that attemps to be more pacaur like. Mostly in minimizing user
input. Ask every thing first then do everything with no need for more
user input.
It is not yet fully complete but is finished enough so that it works,
should not fail in most cases and provides a base for more contributors
to help address the existing problems.
The new install chain is as follows:
Source info about the provided targets
Fetch a list of all dependancies needed to install targets
I put alot of effort into fetching the dependancy tree
while making the least amount of aur requests as
possible. I'm actually very happy with how it turned out
and yay wil now resolve dependancies noticably faster
than pacaur when there are many aur dependancies.
Install repo targets by passing to pacman
Print dependancy tree and ask to confirm
Ask to clean build if directory already exists
Download all pkgbuilds
Ask to edit all pkgbuilds
Ask to continue with the install
Download the sources for each packagebuild
Build and install every package
using -s to get repo deps and -i to install
Ask to remove make dependancies
There are still a lot of things that need to be done for a fully working
system. Here are the problems I found with this system, either new or
existing:
Formating
I am not so good at formatting myself, I thought best to
leave it until last so I could get feedback on how it
should look and help implementing it.
Dependancy tree
The dependancy tree is usually correct although I have
noticed times where it doesnt detect all the
dependancies that it should. I have only noticed this
when there are circular dependancies so i think this
might be the cause. It's not a big deal currently
because makepkg -i installed repo deps for us which
handles the repo deps for us and will get the correct
ones. So yay might not list all the dependancies. but
they will get installed so I consider this a visual bug.
I have yet to see any circular dependancies in the AUR
so I can not say what will happend but I#m guessing that
it will break.
Versioned packages/dependencies
Targets and dependancies with version constriants such
as 'linux>=4.1' will not be checked on the aur side of
things but will be checked on the repo side.
Ignorepkg/Ignoregroup
Currently I do not handle this in any way but it
shouldn't be too hard to implement.
Conflict checking
This is not currently implemented either
Split Paclages
Split packages are not Handles properly. If we only
specify one package so install from a split package
makepkg -i ends up installing them all anyway. If we
specify more than one (n) package it will actually build the
package base n times and reinstall every split package
n times.
Makepkg
To get things working I decided to keep using the
makepkg -i method. I plan to eventually replace this
with a pacman -U based method. This should allow passing
args such as --dbpath and --config to aur packages
aswell as help solve some problems such as the split
packages.
Clean build
I plan to improve the clean build choice to be a little
more smart and instead of check if the directory exists,
check if the package is already build and if so skip the
build all together.
This bug was caused by me not thinking when passing flags to aurInstall.
Currently a bunch of functions take an array of flags but we don't really
use them any more after the argument parsing update. These should be
refactored out eventually but I'm holding off until I'm more sure about
how these functions should look.
Argument parsing now works mostly as expected for repo packages.
AUR packages are a little tricky becauce makepkg cant handle args such
as '--dbpath'.
Also out alpm handle does not read the commandline options so any
arguments relient on alpm will be ignored.
For now though it seems yay has gained back the functionality it once
had. While also having improved argument handling which should also be
expandable and make it easier to handle anything new that might have
been missed.
This reimplemens all operations yay previously supported:
'-S' 'Syu' 'Si' ect.
Currently the argument objects are not fully implemented with the code.
Theres alot of funky conversion from
argument object -> pkg, flags -> argument object
This is to just get back to the functionally we had before (almost).
I have not looked into it yet but alot of the time pacman flags get
passed to makepkg. this cases an error for most commands now because the
new system Passes all flags:
`yay -Syu` -> flags = '-S' '-y' '-u'
while the old system would have done:
`yay -Syu` -> op = '-Suy', flags = ''
So extra flags are no longer passed at all currently.
This means:
'yay -S aic94xx-firmware --noconfirm -b /tmp/pacutilesu2q6hw/tmp-pacman -d'
will no longer error and 'aic94xx-firmware' will be installed but the
database path change will not apply and the dep checking will not be
skipped.
passToPacman now takes and argParser as a paramater. And is implemented
for the simple cases in cmd.go. Although passToPacman is now left non
working in places which still try to usr the old call format and will
need to be reimplemented.
