Age | Commit message (Collapse) | Author | Files | Lines |
|
still todo:
- 'fix' is missing
- 'del -R' does not work
- 'upgrade' does not do self-upgrade first
... and a lot of testing.
|
|
* fixup the help messages to align up properly
* refresh screen width on SIGWINCH
|
|
commit 4e72075fbab introduced a bug where package installation might happen
in wrong order (reminder for self to separate the package version deduction
to separate step from installation ordering). this restricts the earlier
commit to not mingle with the install order.
|
|
already
|
|
|
|
improves behaviour under certain corner case conditions.
|
|
The db parameter to apk_state_commit is not needed so we remove it.
|
|
Implement the logic for install_if lines. Update info applet to
also display the install_if related fields.
|
|
commit 4e72075fbab introduced late locking for top level packages,
but used the wrong package's top level flag for the check. this
fixes a problem that dependencies might not get pulled in.
|
|
* make it as wide as the screen
* make sure it's drawn after package change
* and draw it using ansi escapes in line buffered stderr
|
|
This allows us to get apk-tools dependencies get reset at proper
time in world. As a bonus, it reduces code amount.
|
|
This will fix certain scenarios where multiple packages are
installed with full package files specified on command line
and they depend on each other.
|
|
and make the progress bar disappear on regular runs too.
|
|
Just disable installation of packages using the new stuff. Also
flag lower case package info fields as non-critical and allow
installation even if that features is not supported.
|
|
When package is installed from commandline, we should always
install that specific instance of package (never favor repository
version if it has difference identity). Otherwise we might not
always end-up installing the .apk given on command line. The
dependency is now against specific checksum identity (marked
with >< dependency comparison). Fixes #492.
|
|
Also re-exec's apk-tools to perform rest of the upgrade using
the new apk-tools. This allows handling of new apk-tools features
properly. Fixes #140.
|
|
The version now needs to be initialized atom always, since it's
dereferenced in various places.
|
|
- implement a hash table for commonly shared fields such as
license, version and architecture
- use macroes to print blobs or pkgname-pkgver strings
- fix some old cruft
|
|
Installed package does not need to be checked for availability.
Account for packages missing if they get pruned out due to
installability check.
|
|
|
|
|
|
Instead of having a null pointer, use a dummy array which just
says the array is empty. This helps in multiple places of the code
which would otherwise need explicitly need to check first if the
array exists. This has been cause of multiple seg.faults in the
past as the array check is easily omitted.
This also removes (or fixes) all existing checks accordingly.
|
|
|
|
especially important if the package requested from command line
does not exist. otherwise we would not print an error at all.
|
|
Print more information why installation changeset calculation failed.
Fixes #187.
|
|
To print upgrading if package is actually being changed instead of
pure reinstall.
|
|
If version is equal, but package contents are different we should
show the package being upgraded.
|
|
In addition to autocleaning dependencies on deletion, we need to
autoclean the old package on upgrade too. This is to make sure that
obsoleted dependencies (existed previously, but not in new package)
are removed where appropriate.
|
|
they do not have any package associated. this is indicated by
package with zero installed_size.
|
|
even if we have only one package as candidate, we need to check
it's availability. otherwise we can endup with bad changeset
referring to unavailable package.
|
|
prints like:
(1/12) Installing...
(2/12) Installing...
etc...
|
|
This makes things a little bit nicer when installing from network with
slow lines.
|
|
When pressing only <enter> on the question "..continue [Y/n]?" then
lets take that as a "yes"
|
|
otherwise we would not be able to pick a package that exists
as a candidate for re-installation. fixes #138.
|
|
- make virtual packages work again
- make apk del (for non-empty packages) work again
|
|
allow also overriding old (possibly bad) world dependency.
hopefully it's more bullet proof now.
|
|
|
|
this makes the database package entry smaller, and we propbably
get more fields to installed_package later too. this cleans up
the way scripts are stored and is a preparation for supporting
triggers. some parsing for trigger meta-data. ref #45.
|
|
so do not skip it even if it's no longer available.
|
|
|
|
instead enforce world dependencies when the package name
is first referenced upon.
|
|
this helps boots sequence when network is not available.
|
|
|
|
|
|
|
|
so do not choke on it.
|
|
also makes it possibly to upgrade package without adding it to
top-level deps. fixes #69.
|
|
previously they might have been skipped on certain situations.
this also fixes some other reverse dependency enforcements and
implements new "pending" state for locked name.
|
|
and by default just update the world, and dump a lost of packages
that are not removed. fixes #47.
|
|
this way we never change cwd, and relative filenames are always
parsed consistently. this also helps filename construction in many
places. this patch also changes '--root' to override location of
all configuration to be in the new root. previously it depended
on the file which one was used.
|