---
 logread.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/logread.c b/logread.c
index a4382b3..d996e36 100644
--- a/logread.c
+++ b/logread.c
@@ -20,7 +20,7 @@ usage(char *name)
 	fprintf(stderr, "Arguments:\n");
 	fprintf(stderr, "  -f          Follow the live log output\n");
 	fprintf(stderr, "  -F          Same as -f but also show the old messages\n");
-	fprintf(stderr, "  -n limit    Limit the amount of old messages to fetch\n");
+	fprintf(stderr, "  -n limit    Limit the amount of old messages to fetch (specify '0' for all messages)\n");
[message trimmed]

This refactors the get_recommends function that was originally used for
UI packages to support pmb_recommends for any package (and subpackage).

Extending pmb_recommends will, for example, help us create better
generic device packages [1] and can be used to improve packaging for UIs
with shared pmb_recommends[2].

1. https://gitlab.com/postmarketOS/pmaports/-/merge_requests/4673
2. https://gitlab.com/postmarketOS/pmaports/-/merge_requests/3700
---
 pmb/config/__init__.py                        |  6 +--
 pmb/install/_install.py                       | 49 ++++++++++++++++++-
 pmb/install/ui.py                             | 30 ------------
 test/test_install.py                          | 36 ++++++--------
[message trimmed]

---
 pmb/parse/_apkbuild.py | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/pmb/parse/_apkbuild.py b/pmb/parse/_apkbuild.py
index 6e9d004f..a76dedc0 100644
--- a/pmb/parse/_apkbuild.py
+++ b/pmb/parse/_apkbuild.py
@@ -271,6 +271,12 @@ def _parse_subpackage(path, lines, apkbuild, subpackages, subpkg):
     # Copy variables
     apkbuild = apkbuild.copy()
     apkbuild["subpkgname"] = subpkgname
+    # Don't inherit pmb_recommends from the top-level package.
+    # There are two reasons for this:
[message trimmed]

I zap chroots a lot, since I've found that it often "fixes" a lot of
weird issues that come about if you have stale chroots laying around.
So a common pattern I do is "pmb zap && pmb install ...". Having an
option to pmb install let's me simplify this.
---
 pmb/install/_install.py | 3 +++
 pmb/parse/arguments.py  | 2 ++
 2 files changed, 5 insertions(+)

diff --git a/pmb/install/_install.py b/pmb/install/_install.py
index 1adfe789..2f8e6e6a 100644
--- a/pmb/install/_install.py
+++ b/pmb/install/_install.py
@@ -1170,6 +1170,9 @@ def install(args):
[message trimmed]

Without this, the falls through to the next case and throws an
error.

All credit goes to Caleb Connelly for spotting this!
---
 main.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/main.c b/main.c
index b617bfc..9335c17 100644
--- a/main.c
+++ b/main.c
@@ -738,6 +738,7 @@ main(int argc, char *argv[])
 					exit(1);
[message trimmed]

---
 pmb/config/init.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/pmb/config/init.py b/pmb/config/init.py
index e86209df..f3ee980b 100644
--- a/pmb/config/init.py
+++ b/pmb/config/init.py
@@ -709,6 +709,7 @@ def frontend(args):
 
     cfg["pmbootstrap"]["user"] = ask_for_username(args)
     ask_for_provider_select_pkg(args, "postmarketos-base", cfg["providers"])
+    ask_for_provider_select_pkg(args, "postmarketos-base-ui", cfg["providers"])
 
[message trimmed]

If a package had _pmb_select set and an appropriate provider configured
in pmb's cfg, it was not being used when providers were being resolved.
With this change, those are now being considered. The order is also
important, we want a selected provider to be chosen before it tries to
make a decision based on provider_priority later in step #6.
---
 pmb/parse/depends.py | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/pmb/parse/depends.py b/pmb/parse/depends.py
index 0920202a..cb487ca1 100644
--- a/pmb/parse/depends.py
+++ b/pmb/parse/depends.py
@@ -72,13 +72,20 @@ def package_provider(args, pkgname, pkgnames_install, suffix="native"):
[message trimmed]

This is in support of
https://gitlab.com/postmarketOS/pmaports/-/issues/1379, specifically
to enable selecting a wifi backend in pmos-base-ui.

There is something quirky with this though...
pmb/chroot/apk.py:install_run_apk, due to 6a741091, causes some extra
work to be done. Basically what happens is if a package depends on
something that can be selected with _pmb_select, pmb will install the
option with the highest priority first, then later it will install the
selected option.

From what I can tell by looking at install_run_apk, to_add installs
pmos-ui-*, which pulls in pmos-base-ui. That in turn depends on
pmos-wifi-backend (which is also in _pmb_select). The select preference

On Sun, 19 Nov 2023 19:22:50 +0100 Oliver Smith <ollieparanoid@postmarketos.org> wrote:
> Rename the argument, because any block device can be passed to the
> argument. Use "disk", because the other short word "device" usually
> means the target device/phone to install.
> 
> Keep --sdcard as alias for compatibility with existing scripts and
> muscle memory.

Reviewed-by: Clayton Craft <clayton@craftyguy.net>

This fixes an issue when pmb is run on an Alpine/pmOS host and apk.static sees
that /etc/apk/interactive is set on the host. It's really annoying to have the
build pause so apk can prompt when building chroots, so this uses a flag to apk
to disable interactive mode. I can't think of any situations where we would
actually want to prompt users when building chroots, by then all user
configuration should have been completed via pmb init/config.
---

V2: also patched chroot.apk.

For clarification, IMHO pmb should never hang up at a prompt like this:

    (003851) [10:46:39] (rootfs_valve-jupiter) install postmarketos-mkinitfs-hook-debug-shell
    (003851) [10:46:39] % doas rm -f /home/clayton/.local/var/pmbootstrap/chroot_rootfs_valve-jupiter/tmp/apk_progress_fifo
[message trimmed]