GitLab

When creating an index apk warns if a dependency is missing a provider.
However when using a multi-arch repository, it's not an error that
a certain architecture is missing a dependency because that dependency
could be in an other architecture. Since apk index doesn't know about
this, add an argument to disable that warning.

Maintainer note: rebased for new option handling, and minor stylistic
adjustments.
Signed-off-by: Fredrik Gustafsson <fredrigu@axis.com>

This reduces the number of relocations on PIE binaries, and also
reduces the executable size. Parsing of the options is slightly
sped up as only the exact matching option group parser is called.

Also few minor fixups added.

This creates main help like:
--
usage: apk [<OPTIONS>...] COMMAND [<ARGUMENTS>...]

Package installation and removal:
  add        Add packages to WORLD and commit changes
  del        Remove packages from WORLD and commit changes

System maintenance:
  fix        Check WORLD against the system and ensure consistency
  update     Update repository indexes
  upgrade    Install upgrades available from repositories
  cache      Commands related to the management of an offline package cache

Querying package information:
  info       Give detailed information about packages or repositories
  list       List packages matching a pattern or other criteria
  dot        Generate graphviz graphs
  policy     Show repository policy for packages

Repository maintenance:
  index      Create repository index file from packages
  fetch      Download packages from global repositories to a local directory
  manifest   Show checksums of package contents
  verify     Verify package integrity and signature

Miscellaneous:
  audit      Audit directories for changes
  stats      Show statistics about repositories and installations
  version    Compare package versions or perform tests on version strings

This apk has coffee making abilities.
--

And applet specific help like:
--
usage: apk add [<OPTIONS>...] PACKAGES...

Description:
  apk add adds the requested packages to WORLD and installs (or upgrades)
  them if not already present, ensuring all dependencies are met.

Options:
  --initdb         Initialize a new package database
  -l, --latest     Disables normal heuristics for choosing which repository to install a
  -u, --upgrade    When adding packages which are already installed, upgrade them rather
  -t, --virtual NAME
                   Instead of adding the specified packages to WORLD, create a new
  --no-chown       Do not change file owner or group
--

If apk is run as a non-root user, it's not possible to chown files.

Maintainers note: minor wording changes on commit log and man page.
Signed-off-by: Fredrik Gustafsson <fredrigu@axis.com>

Some email addresses was underlined and some wasn't. Remove underline
from all email addresses to be consistent. The reason for chosing
no underline is that the email address isn't clickable and to be
consistent with the output from git log.

We want to include via <apk/$headerfile> and not via <$headerfile>,
so we want to add an include statement for the dir the apk folder
which contains the headers is contained in.

Fix comparing of the hostname portion that matches exactly.
The no_proxy matching is pretty rudimentary though and probably
could go through a bit of additional rework.

Fixes #10681

Most notably this happens after blank database is initialized with --initdb,
but can happen also in other conditions.

The error checking changes modified behaviour if the file does not exist.
Explicitly check for ENOENT and ignore it. But the behaviour is improved
from earlier as now e.g. EPERM will be detected and reported correctly.

Fixes #10679

Fixes: 6da3e8eb "istream, archive, db: convert db and tar function to use istream"

- split the code to a helper function
- do not set sockets to corked state when putting back to
  cache so socket state is always deterministic
- cork/uncork also when sending CONNECT to a proxy, this
  can reduce a little bit the latency how fast the packet
  gets sent out
- also pair corking with uncorking in http_request to make
  it more obvious pairing

The recent TCP_CORK change missed this bit of code.  This change
should improve performance a bit when making HTTP requests by calling
http_cmd only once instead of three times.

Some screen size are quite small. For example, the default phosh
terminal is less than 50 character wide on Pinephone. This lowers the minimum
loading bar size to 25 characters.

For comparison, 25 character wide is just as wide as "apk add firefox
linux-lts" without the quotes.

Here's a bad picture to illustrate the result
gitlab.alpine.org/uploads/48c20f746fbf685b62b6bd73585ecbf2/pinephone-phosh.png

fixes #10677

src/apk_defines.h:152:15: error: unknown type name 'uint32_t'
 static inline uint32_t get_unaligned32(const void *ptr)
               ^~~~~~~~

- always create .so and .a with the apk libary code
- create .pc file
- install headers and above mentioned files

Maintainers edit: merged commits, cleaned up sed script logic,
and few other minor fixes.

I'm sorry about that. :/

Unfortunately libfetch operates on raw sockets and is sending
each HTTP request line using separate syscall which causes the
HTTP request to be sent as multiple packets over the wire in most
configurations. This is not good for performance, but can also
cause subtle breakage if there's DPI firewall that does not get
the Host header.

Incidentally, it seems that on BSDs libfetch already sets
TCP_NOPUSH optimize the packetization. This commit adds same
logic for using TCP_CORK if available. When using TCP_CORK
there is no requirement to set TCP_NODELAY as uncorking will
also cause immediate send. Keep TCP_NODELAY in the fallback
codepaths.

Long term, it might make sense to replace or rewrite libfetch
to use application level buffering.

- do not do unaligned accesses on non-x86 hardware
- clean up the code a little bit

Related to MR !15

Instead, to make sure test mode produces same output, redefine
time() for the test mode binary.

Reverts parts of 0b82bcc5.

The help has been moved to man pages now.

That list is needed in multiple places, so it simplifies
things a bit.

Detailed docs have been removed from the apk binaries, in favor of
git-style short summaries of each command.