bsd.pkg.mk. buildlink.mk files never need to add rpaths to LDFLAGS unless
a package installs libraries in some place other than ${PREFIX}/lib, e.g.
mysql-client, qt2-libs.
when you encounter libncurses, turn it into libcurses, and likewise when
you encounter ncurses.h, turn it into curses.h.
Makes mtr work on a 1.5.2 shark.
BUILDLINK_PREFIX.<pkgname>. This allows buildlink to find X11BASE packages
regardless of whether they were installed before or after xpkgwedge was
installed. Idea by Alistair Crooks <agc@pkgsrc.org>.
name as real ncurses headers. This is to prevent accidentally resolving
something like <include/term.h> into ${LOCALBASE}/include/term.h if the
compiler is passed "-I${BUILDLINK_DIR}/include -I${LOCALBASE}/include".
FOO_REQD=1.0 being converted to foo>=1.0, one can now directly specify
the dependency pattern as FOO_DEPENDS=foo>=1.0. This allows things like
JPEG_DEPENDS=jpeg-6b, or fancier expressions like for postgresql-lib.
Change existing FOO_REQD definitions in Makefiles to FOO_DEPENDS.
ncurses libraries and headers into ${BUILDLINK_DIR}, or if a sufficiently
advanced curses is available in-system, then it links the curses libraries
and headers into ${BUILDLINK_DIR} with ncurses names. This allows us to
consistently reference "ncurses" capability via <ncurses.h> and -lncurses.
into LDFLAGS. This makes building of ncurses on a system already having
ncurses installed possible. This is just another chapter of the continuing
buildlink saga.
needed for the xerces-p package, but ended up not being necessary. By
removing this hack, we reduce the number of places in pkgsrc where we depend
upon libtool internals.
out of date - it was based on a.out OBJECT_FMT, and added entries in the
generated PLISTs to reflect the symlinks that ELF packages uses. It also
tried to be clever, and removed and recreated any symbolic links that were
created, which has resulted in some fun, especially with packages which
use dlopen(3) to load modules. Some recent changes to our ld.so to bring
it more into line with other Operating Systems also exposed some cracks.
+ Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain
the ELF symlinks.
+ Don't mess about with file system entries when handling shared objects in
bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will
have got it right, and have a much better idea than we do.
+ Modify PLISTs to contain "ELF symlinks"
+ On a.out platforms, delete any "ELF symlinks" from the generated PLISTs
+ On ELF platforms, no extra processing needs to be done in bsd.pkg.mk
+ Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on
a.out platforms
+ Update the documentation in Packages.txt
With many thanks to Thomas Klausner for keeping me honest with this.
but mostly comprises patches to ease building on NetBSD pkgsrc. Tom Dickey
has been kind enough to accept our patches back into the main source tree.
Many thanks to him.
Add a new USE_LIBTOOL definition that uses the libtool package instead of
pkglibtool which is now considered outdated.
USE_PKGLIBTOOL is available for backwards compatibility with old packages
but is deprecated for new packages.
We decided to release ncurses as a new whole number release (5.0)
because it incorporates several interface changes, including some that
would invalidate existing shared libraries. These are the highlights
from the change-log since ncurses 4.2 release.
Interface changes:
* The principal source of changes to the interface comes from the
release of X/Open Curses in 1997. Earlier versions of ncurses (4.0
and before) were based on a draft version of the specification.
The release version adds parameters to some functions to support
the evolving internationalization of curses. These summarize the
impact:
+ modified several prototypes to correspond with 1997 version
of X/Open Curses (affects ABI since developers have used
attr_get).
+ corrected prototypes for slk_* functions, using chtype rather
than attr_t.
+ the slk_attr_{set,off,on} functions need an additional void*
parameter according to XSI.
+ correct macros for wattr_set, wattr_get, separate wattrset
macro from these to preserve behavior that allows attributes
to be combined with color pair numbers.
+ reviewed/updated curses.h, term.h against X/Open Curses Issue
4 Version 2. This includes making some parameters
NCURSES_CONST rather than const, e.g., in termcap.h.
+ reviewed/corrected macros in curses.h as per XSI document.
+ add set_a_attributes and set_pglen_inch to terminfo
structure, as per XSI and Solaris 2.5.
* The newest version of the X/Open Curses is implemented on Solaris
and other vendor's systems. It adds new features to the terminfo
descriptions:
+ implement tparm %l format.
+ implement tparm printf-style width and precision for %s, %d,
%x, %o as per XSI.
* We made additional changes to reduce impact by future interface
changes:
+ rename key_names[] array to _nc_key_names since it is not
part of the curses interface.
+ move macro winch to a function, to hide details of struct
ldat
* modify configure script to embed ABI in shared libraries for HP-UX
10.x (detailed request by Tim Mooney).
* modify configuration of shared libraries on Digital Unix so that
versioning is embedded in the library, rather than implied by
links (patch by Tim Mooney).
New features:
* enable sigwinch handler by default.
* turn on hashmap scrolling code by default
* improved support for termcap applications
+ modify tput to accept termcap names as an alternative to
terminfo names.
+ provide support for termcap PC variable by copying it from
terminfo data and using it as the padding character in tputs.
+ provide support for termcap ospeed variable by copying it
from the internal cur_term member, and using ospeed as the
baudrate reference for the delay_output and tputs functions.
+ change name-comparisons in lib_termcap to compare no more
than 2 characters.
+ add configure option --enable-tcap-names, which essentially
allows users to define new capabilities as in termcap.
* add mouse support to ncurses menus.
* add mouse and dll support for OS/2 EMX
* modify terminfo parsing to accept octal and hexadecimal constants
* add configure option --enable-no-padding, to allow environment
variable $NCURSES_NO_PADDING to eliminate non-mandatory padding,
thereby making terminal emulators (e.g., for vt100) a little more
efficient.
* modify lib_color.c to eliminate dependency on orig_colors and
orig_pair, since SVr4 curses does not require these either, but
uses them when they are available.
* add -f option to infocmp and tic, which formats the terminfo
if/then/else/endif so that they are readable (with newlines and
tabs).
* modify tic to compile into %'char' form in preference to
%{number}, since that is a little more efficient.
Major bug fixes:
* modify lib_tstp.c to block SIGTTOU when handling SIGTSTP, fixes a
problem where ncurses applications which were run via a shell
script would hang when given a ^Z. Also, check if the terminal's
process group is consistent, i.e., a shell has not taken ownership
of it, before deciding to save the current terminal settings in
the SIGTSTP handler.
* suppress sc/rc capabilities from terminal description if they
appear in smcup/rmcup. This affects only scrolling optimization,
to fix a problem reported by several people with xterm's alternate
screen, though the problem is more general.
* modify relative_move and tputs to avoid an interaction with the
BSD-style padding. The relative_move function could produce a
string to replace on the screen which began with a numeric
character, which was then interpreted by tputs as padding.
* modify setupterm so that cancelled strings are treated the same as
absent strings, cancelled and absent booleans false (does not
affect tic, infocmp).
* modify lib_vidattr.c to allow for terminal types (e.g.,
xterm-color) which may reset all attributes in the 'op'
capability, so that colors are set before turning on bold and
other attributes, but still after turning attributes off.
* use 'access()' to check if ncurses library should be permitted to
open or modify files with fopen/open/link/unlink/remove calls, in
case the calling application is running in setuid mode.
* correction to doupdate, for case where terminal does not support
insert/delete character. The logic did not check that there was a
difference in alignment of changes to old/new screens before
repainting the whole non-blank portion of the line. Modified to
fall through into logic that reduces by the portion which does not
differ.
library. This was a side-effect of trying to deal with --with-cxx and
--without-cxx. We accept the default --with-cxx. This should work
regardless of whether libg++ exists.
ago. With this switch on, configure DTRT WRT the "bool" data type, resulting
in bizarre compile errors when using ncurses with c++ later:
/usr/pkg/include/ncurses.h:96: abstract declarator used as declaration
(caused by a "typedef bool;")
libncurses.la is listed first. This allows a previous installation
of ncurses to exist and have the build of the current ncurses still
find the correct libs.
># XXX The following comment is somewhat incorrect. The package will build
># XXX correctly on alpha, and even build shared libs, but not quite correctly:
># XXX The shlibs have no SONAME, so references to them from programs
># XXX will end up being to libfoo.so. This may cause upgrade problems in the
># XXX future. However, an ncurses binary package appeared e.g. with 1.4,
># XXX seemingly built from these bits, so it's probably best to just enable
># XXX it for now. -- cgd 19990613
Note that that's a generic ELF problem, and not limited to the Alpha, as well.
