In case `git branch --remote` yields no result, fall back to `git branch`
without `--remote` since it is possible that the local branch has not
been pushed yet.
Also simplify extraction of the branch name by not passing `--verbose`
and utilizing variable substitutions instead of sed expressions.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Detect the presence of directories using Make $(wildcard ...) and emit
plain install commands depending on the outcome instead of relying on
shell conditionals which impose syntax constraints on code outside of
the scope of the LuCI repo.
Fixes: #5936
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Since english is the implementation language of LuCI, it makes no sense to
ship translation packages for it.
Ref: #5860
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Introduce three new default enabled variables `LUCI_MINIFY_JS`,
`LUCI_MINIFY_LUA` and `LUCI_MINIFY_CSS` that allow LuCI component
Makefiles to opt out of JavaScript, Lua and CSS compression respectively.
To e.g. inhibit CSSTidy postprocessing in a theme, the corresponding
Makefile can set `LUCI_MINIFY_CSS:=0` prior to including `luci.mk`.
Fixes: #5755
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Do not run the uci-defaults scripts of a package in the package-postinst, but leave this
for the default_postinst() of the OpenWrt base-files package.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
When including luci.mk in external repos it's sometimes usefull to not use
the default LuCI-submenu hierarchy.
This change defines the LUCI_SUBMENU_FORCED variable which completely overrides
the default submenu of the LuCI config-section. When LUCI_SUBMENU_FORCED is not
defined, the default submenu derrived from LUCI_TYPE or "Application" fallback
is used.
Defining LUCI_SUBMENU_FORCED in the package Makefile will just use this value.
Setting it to "none" will not define a submenu at all.
Together with LUCI_SECTION and LUCI_CATEGORY menu items can now created at any
place in the menu structure.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
Add the LUCI_URL and LUCI_MAINTAINER variables to pass them to the buildpackage
defines. Give them some sane defaults and allow overwritting by the individual
package.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
When the Dutch translation was added there was no name for the language
added, this created such a Kconfig entry:
config LUCI_LANG_nl
tristate " (nl)"
This entry created the following warning message when doing make
oldconfig in OpenWrt:
tmp/.config-package.in:24525:warning: leading whitespace ignored
This is fixed here by adding a name to the language.
Fixes: 1e301ca2df ("Translated using Weblate (Dutch)")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The luci.mk file has grown over the years and became more complex. This
change reorders some sections to make it more easy to read and understand
on the first view.
* move all constants into a common block
* put all standard package definitons together
* put the specific sections for the "luci-base" behind the common things
* have the defines for the translations at the very end
Signed-off-by: Sven Roederer <freifunk@it-solutions.geroedel.de>
There is some translation effort ongoing in weblate,
so add also Finnish to the list of supported languages.
Katsotaan riittääkö innostuneita kääntäjiä...
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
This includes `bn_BD` and `bn` via an alias.
Weblate currently does not support `bn_IN` (Indian Bangla), so I think
the alias is reasonable.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Also pass correct variable to builds to ensure that version.lua and cache
buster strings are properly interpolated.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
- Derive the version of packages from the last change of the sources
- Derive the version of i18n packages from the last change of the po dir
Fixes: #2257
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Send a SIGHUP to rpcd after installing LuCI packages to reload ACL rules.
Also remove redundant code while we're at it.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
If a postinst script is defined in the package Makefile, it will
not be included because it will be replaced by the default postinst
script. To solve the problem, we now check, if we have a postinst script.
And if so, it will be installed instead of the default postinst script.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Almost every uci-defaults script for ucitrack execute after the uci
commands the following line '/tmp/luci-indexcache'.
So that we don't always write the same thing, we now do this in
postinst.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
luasrcdiet is being moved to the packages feed as a separate package.
Adjust luci feed accordingly.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Fix the length of the abbreviated commit hash used in
the LuCI version string used in the package names.
Use the old standard of seven characters in hash.
New git versions vary the length of the abbreviated
commit hash based on the number of objects in repo. That
may lead into different hash lengths with e.g. having
full/shallow git history, having all/single branch or
having different git versions in different build hosts.
Different hash lengths in package name may lead to
sneaky opkg install errors, as opkg apparently preserves
info for removed packages and may prefer that.
Example:
$ opkg install http://downloads...luci-app-statistics_git-17.073.42825-b47a21f-1_all.ipk
Downloading http://downloads...luci-app-statistics_git-17.073.42825-b47a21f-1_all.ipk
Installing luci-app-statistics (git-17.073.42825-b47a21fbb-1) to root...
Collected errors:
* opkg_download_pkg: Package luci-app-statistics is not available from any configured src.
The error above is ultimately in opkg, but mitigate
the problem by fixing the hash length when building LuCI.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
LEDE's scripts/feeds script creates local branches named after the checked
out commit hash, confusing local branch detection.
Solely use remote branch detection to avoid misidentifying such cases.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>