Commit Graph

6 Commits

Author SHA1 Message Date
Leonid Pliushch
c352216c63 packages: standardize values of several control variables for build.sh
Variables

 TERMUX_PKG_PLATFORM_INDEPENDENT
 TERMUX_DEBUG
 TERMUX_PKG_HAS_DEBUG
 TERMUX_PKG_ESSENTIAL
 TERMUX_SUBPKG_ESSENTIAL
 TERMUX_PKG_NO_STATICSPLIT
 TERMUX_PKG_BUILD_IN_SRC
 TERMUX_PKG_FORCE_CMAKE
 TERMUX_PKG_HOSTBUILD

should not accept arbitrary values for marking them "enabled". Instead
they should accept boolean values which makes them easier to handle and
also makes their meaning clear.

build-package.sh should make decision based on variable's value but not on
whether it is set or empty.

%ci:no-build
2019-08-12 18:49:13 +03:00
Leonid Pliushch
361d8c519f [incomplete] set licenses for packages 2019-01-20 22:41:50 +01:00
Fredrik Fornwall
51c7760090 Drop need of setting TERMUX_PKG_FOLDERNAME
Instead figure out the top level source folder ourselves.
2017-09-20 00:48:04 +02:00
ElTopo
35aecce3a3 add new package: redir (#772)
* add bvi package

* add new package: cgdb

* add gdb in cgdb's TERMUX_PKG_DEPENDS

* add new package: redir

* bvi package revision 1: fix column number issue with clang (gcc does not seem to have this issue)
bvi (compiled with clang) incorrectly displayed only 12 bytes for each line on a 80-column terminal (instead of 16 bytes normally). command :set cm=## also had similar problems.
I found the issue was caused by the following statements:
       sprintf(addr_form,  "%%0%dllX  ", AnzAdd);
       AnzAdd = sprintf(tmp, addr_form, block_begin);

	   ('block_begin''s data type is off_t)
	   gcc automatically typecasts 'block_begin' from off_t to loff_t,
	   on the otherhand, clang takes address of 'block_begin' and uses
	   it as loff_t thus writes some garbage value in 'tmp' string.
2017-02-20 22:14:52 +01:00
Fredrik Fornwall
afaa9e8b84 bvi: Add source checksum 2017-01-25 01:39:15 +01:00
ElTopo
fad0fde667 add bvi package (#656) 2017-01-25 01:36:06 +01:00