2 A Hacker's Guide to NCURSES
7 * Objective of the Package
9 + How to Design Extensions
10 * Portability and Configuration
12 * Documentation Conventions
14 * A Tour of the Ncurses Library
19 + Output and Screen Updating
20 * The Forms and Menu Libraries
21 * A Tour of the Terminfo Compiler
22 + Translation of Non-use Capabilities
23 + Use Capability Resolution
24 + Source-Form Translation
26 * Style Tips for Developers
31 This document is a hacker's tour of the ncurses library and utilities.
32 It discusses design philosophy, implementation methods, and the
33 conventions used for coding and documentation. It is recommended
34 reading for anyone who is interested in porting, extending or
35 improving the package.
37 Objective of the Package
39 The objective of the ncurses package is to provide a freeware API for
40 character-cell terminals and terminal emulators with the following
43 * Source-compatible with historical curses implementations
44 (including the original BSD curses and System V curses.
45 * Conformant with the XSI Curses standard issued as part of XPG4 by
47 * High-quality -- stable and reliable code, wide portability, good
48 packaging, superior documentation.
49 * Featureful -- should eliminate as much of the drudgery of C
50 interface programming as possible, freeing programmers to think at
51 a higher level of design.
53 These objectives are in priority order. So, for example, source
54 compatibility with older version must trump featurefulness -- we
55 cannot add features if it means breaking the portion of the API
56 corresponding to historical curses versions.
60 We used System V curses as a model, reverse-engineering their API, in
61 order to fulfill the first two objectives.
63 System V curses implementations can support BSD curses programs with
64 just a recompilation, so by capturing the System V API we also capture
67 More importantly for the future, the XSI Curses standard issued by
68 X/Open is explicitly and closely modeled on System V. So conformance
69 with System V took us most of the way to base-level XSI conformance.
71 How to Design Extensions
73 The third objective (standards conformance) requires that it be easy
74 to condition source code using ncurses so that the absence of
75 nonstandard extensions does not break the code.
77 Accordingly, we have a policy of associating with each nonstandard
78 extension a feature macro, so that ncurses client code can use this
79 macro to condition in or out the code that requires the ncurses
82 For example, there is a macro NCURSES_MOUSE_VERSION which XSI Curses
83 does not define, but which is defined in the ncurses library header.
84 You can use this to condition the calls to the mouse API calls.
86 Portability and Configuration
88 Code written for ncurses may assume an ANSI-standard C compiler and
89 POSIX-compatible OS interface. It may also assume the presence of a
90 System-V-compatible select(2) call.
92 We encourage (but do not require) developers to make the code friendly
93 to less-capable UNIX environments wherever possible.
95 We encourage developers to support OS-specific optimizations and
96 methods not available under POSIX/ANSI, provided only that:
98 * All such code is properly conditioned so the build process does
99 not attempt to compile it under a plain ANSI/POSIX environment.
100 * Adding such implementation methods does not introduce
101 incompatibilities in the ncurses API between platforms.
103 We use GNU autoconf(1) as a tool to deal with portability issues. The
104 right way to leverage an OS-specific feature is to modify the autoconf
105 specification files (configure.in and aclocal.m4) to set up a new
106 feature macro, which you then use to condition your code.
110 The 'configure' script usually gets your system environment right
111 automatically. Here are some -D options you might need to compile with
115 if <unistd.h> is present
118 if the sigaction function is present
121 if the usleep function is present
124 if (e.g., svr4) you need _POSIX_SOURCE to have sigaction
127 if you have <termios.h>
130 if you have <termio.h>; otherwise it uses <sgtty.h>
132 -DBROKEN_TIOCGETWINSZ
133 on SVR4 and HPUX, the get window size ioctl is broken.
135 Documentation Conventions
137 There are three kinds of documentation associated with this package.
138 Each has a different preferred format:
140 * Package-internal files (README, INSTALL, TO-DO etc.)
142 * Everything else (i.e., narrative documentation).
144 Our conventions are simple:
146 1. Maintain package-internal files in plain text. The expected viewer
147 for them more(1) or an editor window; there's no point in
149 2. Mark up manual pages in the man macros. These have to be viewable
150 through traditional man(1) programs.
151 3. Write everything else in HTML.
153 When in doubt, HTMLize a master and use lynx(1) to generate plain
154 ASCII (as we do for the announcement document).
156 The reason for choosing HTML is that it's (a) well-adapted for on-line
157 browsing through viewers that are everywhere; (b) more easily readable
158 as plain text than most other mark-ups, if you don't have a viewer;
159 and (c) carries enough information that you can generate a
160 nice-looking printed version from it. Also, of course, it make
161 exporting things like the announcement document to WWW pretty trivial.
165 The reporting address for bugs is ncurses@bsdi.com. This is a
166 majordomo list; to join, write to ncurses-request@mailgate.bsdi.com
167 with a message containing the line:
169 subscribe <name>@<host.domain>
171 The ncurses code is maintained by a small group of volunteers. While
172 we try our best to fix bugs promptly, we simply don't have a lot of
173 hours to spend on elementary hand-holding. We rely on intelligent
174 cooperation from our users. If you think you have found a bug in
175 ncurses, there are some steps you can take before contacting us that
176 will help get the bug fixed quickly.
178 In order to use our bug-fixing time efficiently, we put people who
179 show us they've taken these steps at the head of our queue. This means
180 that if you don't, you'll probably end up at the tail end and have to
183 1. Develop a recipe to reproduce the bug.
184 Bugs we can reproduce are likely to be fixed very quickly, often
185 within days. The most effective single thing you can do to get a
186 quick fix is develop a way we can duplicate the bad behavior --
187 ideally, by giving us source for a small, portable test program
188 that breaks the library. (Even better is a keystroke recipe using
189 one of the test programs provided with the distribution.)
190 2. Try to reproduce the bug on a different terminal type.
191 In our experience, most of the behaviors people report as library
192 bugs are actually due to subtle problems in terminal descriptions.
193 This is especially likely to be true if you're using a traditional
194 asynchronous terminal or PC-based terminal emulator, rather than
195 xterm or a UNIX console entry.
196 It's therefore extremely helpful if you can tell us whether or not
197 your problem reproduces on other terminal types. Usually you'll
198 have both a console type and xterm available; please tell us
199 whether or not your bug reproduces on both.
200 If you have xterm available, it is also good to collect xterm
201 reports for different window sizes. This is especially true if you
202 normally use an unusual xterm window size -- a surprising number
203 of the bugs we've seen are either triggered or masked by these.
204 3. Generate and examine a trace file for the broken behavior.
205 Recompile your program with the debugging versions of the
206 libraries. Insert a trace() call with the argument set to
207 TRACE_UPDATE. (See "Writing Programs with NCURSES" for details on
208 trace levels.) Reproduce your bug, then look at the trace file to
209 see what the library was actually doing.
210 Another frequent cause of apparent bugs is application coding
211 errors that cause the wrong things to be put on the virtual
212 screen. Looking at the virtual-screen dumps in the trace file will
213 tell you immediately if this is happening, and save you from the
214 possible embarrassment of being told that the bug is in your code
215 and is your problem rather than ours.
216 If the virtual-screen dumps look correct but the bug persists,
217 it's possible to crank up the trace level to give more and more
218 information about the library's update actions and the control
219 sequences it issues to perform them. The test directory of the
220 distribution contains a tool for digesting these logs to make them
221 less tedious to wade through.
222 Often you'll find terminfo problems at this stage by noticing that
223 the escape sequences put out for various capabilities are wrong.
224 If not, you're likely to learn enough to be able to characterize
225 any bug in the screen-update logic quite exactly.
226 4. Report details and symptoms, not just interpretations.
227 If you do the preceding two steps, it is very likely that you'll
228 discover the nature of the problem yourself and be able to send us
229 a fix. This will create happy feelings all around and earn you
230 good karma for the first time you run into a bug you really can't
231 characterize and fix yourself.
232 If you're still stuck, at least you'll know what to tell us.
233 Remember, we need details. If you guess about what is safe to
234 leave out, you are too likely to be wrong.
235 If your bug produces a bad update, include a trace file. Try to
236 make the trace at the least voluminous level that pins down the
237 bug. Logs that have been through tracemunch are OK, it doesn't
238 throw away any information (actually they're better than
239 un-munched ones because they're easier to read).
240 If your bug produces a core-dump, please include a symbolic stack
241 trace generated by gdb(1) or your local equivalent.
242 Tell us about every terminal on which you've reproduced the bug --
243 and every terminal on which you can't. Ideally, sent us terminfo
244 sources for all of these (yours might differ from ours).
245 Include your ncurses version and your OS/machine type, of course!
246 You can find your ncurses version in the curses.h file.
248 If your problem smells like a logic error or in cursor movement or
249 scrolling or a bad capability, there are a couple of tiny test frames
250 for the library algorithms in the progs directory that may help you
251 isolate it. These are not part of the normal build, but do have their
252 own make productions.
254 The most important of these is mvcur, a test frame for the
255 cursor-movement optimization code. With this program, you can see
256 directly what control sequences will be emitted for any given cursor
257 movement or scroll/insert/delete operations. If you think you've got a
258 bad capability identified, you can disable it and test again. The
259 program is command-driven and has on-line help.
261 If you think the vertical-scroll optimization is broken, or just want
262 to understand how it works better, build hashmap and read the header
263 comments of hardscroll.c and hashmap.c; then try it out. You can also
264 test the hardware-scrolling optimization separately with hardscroll.
266 There's one other interactive tester, tctest, that exercises
267 translation between termcap and terminfo formats. If you have a
268 serious need to run this, you probably belong on our development team!
270 A Tour of the Ncurses Library
274 Most of the library is superstructure -- fairly trivial convenience
275 interfaces to a small set of basic functions and data structures used
276 to manipulate the virtual screen (in particular, none of this code
277 does any I/O except through calls to more fundamental modules
278 described below). The files lib_addch.c, lib_bkgnd.c, lib_box.c,
279 lib_clear.c, lib_clrbot.c, lib_clreol.c, lib_data.c, lib_delch.c,
280 lib_delwin.c, lib_erase.c, lib_getstr.c, lib_inchstr.c, lib_insch.c,
281 lib_insdel.c, lib_insstr.c, lib_instr.c, lib_isendwin.c,
282 lib_keyname.c, lib_move.c, lib_mvwin.c, lib_overlay.c, lib_pad.c,
283 lib_printw.c, lib_scanw.c, lib_screen.c, lib_scroll.c, lib_scrreg.c,
284 lib_set_term.c, lib_slk.c, lib_touch.c, lib_unctrl.c, and lib_window.c
285 are all in this category. They are very unlikely to need change,
286 barring bugs or some fundamental reorganization in the underlying data
289 The lib_trace.c, lib_traceatr.c, and lib_tracechr.c file are used only
290 for debugging support. It is rather unlikely you will ever need to
291 change these, unless you want to introduce a new debug trace level for
294 There is another group of files that do direct I/O via tputs(),
295 computations on the terminal capabilities, or queries to the OS
296 environment, but nevertheless have only fairly low complexity. These
297 include: lib_acs.c, lib_beep.c, lib_color.c, lib_endwin.c,
298 lib_initscr.c, lib_longname.c, lib_newterm.c, lib_options.c,
299 lib_termcap.c, lib_ti.c, lib_tparm.c, lib_tputs.c, lib_vidattr.c, and
300 read_entry.c. These are likely to need revision only if ncurses is
301 being ported to an environment without an underlying terminfo
302 capability representation.
304 The files lib_kernel.c, lib_baudrate.c, lib_raw.c, lib_tstp.c, and
305 lib_twait.c have serious hooks into the tty driver and signal
306 facilities. If you run into porting snafus moving the package to
307 another UNIX, the problem is likely to be in one of these files. The
308 file lib_print.c uses sleep(2) and also falls in this category.
310 Almost all of the real work is done in the files hashmap.c,
311 hardscroll.c, lib_addch.c, lib_doupdate.c, lib_mvcur.c, lib_getch.c,
312 lib_mouse.c, lib_refresh.c, and lib_setup.c. Most of the algorithmic
313 complexity in the library lives in these files. If there is a real bug
314 in ncurses itself, it's probably here. We'll tour some of these files
315 in detail below (see The Engine Room).
317 Finally, there is a group of files that is actually most of the
318 terminfo compiler. The reason this code lives in the ncurses library
319 is to support fallback to /etc/termcap. These files include
320 alloc_entry.c, captoinfo.c, comp_captab.c, comp_error.c, comp_hash.c,
321 comp_parse.c, comp_scan.c, and parse_entry.c, read_termcap.c, and
322 write_entry.c. We'll discuss these in the compiler tour.
328 All ncurses input funnels through the function wgetch(), defined in
329 lib_getch.c. This function is tricky; it has to poll for keyboard and
330 mouse events and do a running match of incoming input against the set
331 of defined special keys.
333 The central data structure in this module is a FIFO queue, used to
334 match multiple-character input sequences against special-key
335 capabilities; also to implement pushback via ungetch().
337 The wgetch() code distinguishes between function key sequences and the
338 same sequences typed manually by doing a timed wait after each input
339 character that could lead a function key sequence. If the entire
340 sequence takes less than 1 second, it is assumed to have been
341 generated by a function key press.
343 Hackers bruised by previous encounters with variant select(2) calls
344 may find the code in lib_twait.c interesting. It deals with the
345 problem that some BSD selects don't return a reliable time-left value.
346 The function timed_wait() effectively simulates a System V select.
350 If the mouse interface is active, wgetch() polls for mouse events each
351 call, before it goes to the keyboard for input. It is up to
352 lib_mouse.c how the polling is accomplished; it may vary for different
355 Under xterm, however, mouse event notifications come in via the
356 keyboard input stream. They are recognized by having the kmous
357 capability as a prefix. This is kind of klugey, but trying to wire in
358 recognition of a mouse key prefix without going through the
359 function-key machinery would be just too painful, and this turns out
360 to imply having the prefix somewhere in the function-key capabilities
361 at terminal-type initialization.
363 This kluge only works because kmous isn't actually used by any
364 historic terminal type or curses implementation we know of. Best guess
365 is it's a relic of some forgotten experiment in-house at Bell Labs
366 that didn't leave any traces in the publicly-distributed System V
367 terminfo files. If System V or XPG4 ever gets serious about using it
368 again, this kluge may have to change.
370 Here are some more details about mouse event handling:
372 The lib_mouse()code is logically split into a lower level that accepts
373 event reports in a device-dependent format and an upper level that
374 parses mouse gestures and filters events. The mediating data structure
375 is a circular queue of event structures.
377 Functionally, the lower level's job is to pick up primitive events and
378 put them on the circular queue. This can happen in one of two ways:
379 either (a) _nc_mouse_event() detects a series of incoming mouse
380 reports and queues them, or (b) code in lib_getch.c detects the kmous
381 prefix in the keyboard input stream and calls _nc_mouse_inline to
382 queue up a series of adjacent mouse reports.
384 In either case, _nc_mouse_parse() should be called after the series is
385 accepted to parse the digested mouse reports (low-level events) into a
386 gesture (a high-level or composite event).
388 Output and Screen Updating
390 With the single exception of character echoes during a wgetnstr() call
391 (which simulates cooked-mode line editing in an ncurses window), the
392 library normally does all its output at refresh time.
394 The main job is to go from the current state of the screen (as
395 represented in the curscr window structure) to the desired new state
396 (as represented in the newscr window structure), while doing as little
399 The brains of this operation are the modules hashmap.c, hardscroll.c
400 and lib_doupdate.c; the latter two use lib_mvcur.c. Essentially, what
401 happens looks like this:
403 The hashmap.c module tries to detect vertical motion changes between
404 the real and virtual screens. This information is represented by the
405 oldindex members in the newscr structure. These are modified by
406 vertical-motion and clear operations, and both are re-initialized
407 after each update. To this change-journalling information, the hashmap
408 code adds deductions made using a modified Heckel algorithm on hash
409 values generated from the line contents.
411 The hardscroll.c module computes an optimum set of scroll, insertion,
412 and deletion operations to make the indices match. It calls
413 _nc_mvcur_scrolln() in lib_mvcur.c to do those motions.
415 Then lib_doupdate.c goes to work. Its job is to do line-by-line
416 transformations of curscr lines to newscr lines. Its main tool is the
417 routine mvcur() in lib_mvcur.c. This routine does cursor-movement
418 optimization, attempting to get from given screen location A to given
419 location B in the fewest output characters posible.
421 If you want to work on screen optimizations, you should use the fact
422 that (in the trace-enabled version of the library) enabling the
423 TRACE_TIMES trace level causes a report to be emitted after each
424 screen update giving the elapsed time and a count of characters
425 emitted during the update. You can use this to tell when an update
426 optimization improves efficiency.
428 In the trace-enabled version of the library, it is also possible to
429 disable and re-enable various optimizations at runtime by tweaking the
430 variable _nc_optimize_enable. See the file include/curses.h.in for
431 mask values, near the end.
433 The Forms and Menu Libraries
435 The forms and menu libraries should work reliably in any environment
436 you can port ncurses to. The only portability issue anywhere in them
437 is what flavor of regular expressions the built-in form field type
438 TYPE_REGEXP will recognize.
440 The configuration code prefers the POSIX regex facility, modeled on
441 System V's, but will settle for BSD regexps if the former isn't
444 Historical note: the panels code was written primarily to assist in
445 porting u386mon 2.0 (comp.sources.misc v14i001-4) to systems lacking
446 panels support; u386mon 2.10 and beyond use it. This version has been
447 slightly cleaned up for ncurses.
449 A Tour of the Terminfo Compiler
451 The ncurses implementation of tic is rather complex internally; it has
452 to do a trying combination of missions. This starts with the fact
453 that, in addition to its normal duty of compiling terminfo sources
454 into loadable terminfo binaries, it has to be able to handle termcap
455 syntax and compile that too into terminfo entries.
457 The implementation therefore starts with a table-driven, dual-mode
458 lexical analyzer (in comp_scan.c). The lexer chooses its mode (termcap
459 or terminfo) based on the first `,' or `:' it finds in each entry. The
460 lexer does all the work of recognizing capability names and values;
461 the grammar above it is trivial, just "parse entries till you run out
464 Translation of Non-use Capabilities
466 Translation of most things besides use capabilities is pretty
467 straightforward. The lexical analyzer's tokenizer hands each
468 capability name to a hash function, which drives a table lookup. The
469 table entry yields an index which is used to look up the token type in
470 another table, and controls interpretation of the value.
472 One possibly interesting aspect of the implementation is the way the
473 compiler tables are initialized. All the tables are generated by
474 various awk/sed/sh scripts from a master table include/Caps; these
475 scripts actually write C initializers which are linked to the
476 compiler. Furthermore, the hash table is generated in the same way, so
477 it doesn't have to be generated at compiler startup time (another
478 benefit of this organization is that the hash table can be in
479 shareable text space).
481 Thus, adding a new capability is usually pretty trivial, just a matter
482 of adding one line to the include/Caps file. We'll have more to say
483 about this in the section on Source-Form Translation.
485 Use Capability Resolution
487 The background problem that makes tic tricky isn't the capability
488 translation itself, it's the resolution of use capabilities. Older
489 versions would not handle forward use references for this reason (that
490 is, a using terminal always had to follow its use target in the source
491 file). By doing this, they got away with a simple implementation
492 tactic; compile everything as it blows by, then resolve uses from
495 This won't do for ncurses. The problem is that that the whole
496 compilation process has to be embeddable in the ncurses library so
497 that it can be called by the startup code to translate termcap entries
498 on the fly. The embedded version can't go promiscuously writing
499 everything it translates out to disk -- for one thing, it will
500 typically be running with non-root permissions.
502 So our tic is designed to parse an entire terminfo file into a
503 doubly-linked circular list of entry structures in-core, and then do
504 use resolution in-memory before writing everything out. This design
505 has other advantages: it makes forward and back use-references equally
506 easy (so we get the latter for free), and it makes checking for name
507 collisions before they're written out easy to do.
509 And this is exactly how the embedded version works. But the
510 stand-alone user-accessible version of tic partly reverts to the
511 historical strategy; it writes to disk (not keeping in core) any entry
512 with no use references.
514 This is strictly a core-economy kluge, implemented because the
515 terminfo master file is large enough that some core-poor systems swap
516 like crazy when you compile it all in memory...there have been reports
517 of this process taking three hours, rather than the twenty seconds or
518 less typical on the author's development box.
520 So. The executable tic passes the entry-parser a hook that immediately
521 writes out the referenced entry if it has no use capabilities. The
522 compiler main loop refrains from adding the entry to the in-core list
523 when this hook fires. If some other entry later needs to reference an
524 entry that got written immediately, that's OK; the resolution code
525 will fetch it off disk when it can't find it in core.
527 Name collisions will still be detected, just not as cleanly. The
528 write_entry() code complains before overwriting an entry that
529 postdates the time of tic's first call to write_entry(), Thus it will
530 complain about overwriting entries newly made during the tic run, but
531 not about overwriting ones that predate it.
533 Source-Form Translation
535 Another use of tic is to do source translation between various termcap
536 and terminfo formats. There are more variants out there than you might
537 think; the ones we know about are described in the captoinfo(1) manual
540 The translation output code (dump_entry() in ncurses/dump_entry.c) is
541 shared with the infocmp(1) utility. It takes the same internal
542 representation used to generate the binary form and dumps it to
543 standard output in a specified format.
545 The include/Caps file has a header comment describing ways you can
546 specify source translations for nonstandard capabilities just by
547 altering the master table. It's possible to set up capability aliasing
548 or tell the compiler to plain ignore a given capability without
549 writing any C code at all.
551 For circumstances where you need to do algorithmic translation, there
552 are functions in parse_entry.c called after the parse of each entry
553 that are specifically intended to encapsulate such translations. This,
554 for example, is where the AIX box1 capability get translated to an
559 The infocmp utility is just a wrapper around the same entry-dumping
560 code used by tic for source translation. Perhaps the one interesting
561 aspect of the code is the use of a predicate function passed in to
562 dump_entry() to control which capabilities are dumped. This is
563 necessary in order to handle both the ordinary De-compilation case and
564 entry difference reporting.
566 The tput and clear utilities just do an entry load followed by a
567 tputs() of a selected capability.
569 Style Tips for Developers
571 See the TO-DO file in the top-level directory of the source
572 distribution for additions that would be particularly useful.
574 The prefix _nc_ should be used on library public functions that are
575 not part of the curses API in order to prevent pollution of the
576 application namespace. If you have to add to or modify the function
577 prototypes in curses.h.in, read ncurses/MKlib_gen.sh first so you can
578 avoid breaking XSI conformance. Please join the ncurses mailing list.
579 See the INSTALL file in the top level of the distribution for details
582 Look for the string FIXME in source files to tag minor bugs and
583 potential problems that could use fixing.
585 Don't try to auto-detect OS features in the main body of the C code.
586 That's the job of the configuration system.
588 To hold down complexity, do make your code data-driven. Especially, if
589 you can drive logic from a table filtered out of include/Caps, do it.
590 If you find you need to augment the data in that file in order to
591 generate the proper table, that's still preferable to ad-hoc code --
592 that's why the fifth field (flags) is there.
598 The following notes are intended to be a first step towards DOS and
599 Macintosh ports of the ncurses libraries.
601 The following library modules are `pure curses'; they operate only on
602 the curses internal structures, do all output through other curses
603 calls (not including tputs() and putp()) and do not call any other
604 UNIX routines such as signal(2) or the stdio library. Thus, they
605 should not need to be modified for single-terminal ports.
607 lib_addch.c lib_addstr.c lib_bkgd.c lib_box.c lib_clear.c lib_clrbot.c
608 lib_clreol.c lib_delch.c lib_delwin.c lib_erase.c lib_inchstr.c
609 lib_insch.c lib_insdel.c lib_insstr.c lib_keyname.c lib_move.c
610 lib_mvwin.c lib_newwin.c lib_overlay.c lib_pad.c lib_printw.c
611 lib_refresh.c lib_scanw.c lib_scroll.c lib_scrreg.c lib_set_term.c
612 lib_touch.c lib_tparm.c lib_tputs.c lib_unctrl.c lib_window.c panel.c
614 This module is pure curses, but calls outstr():
618 These modules are pure curses, except that they use tputs() and
621 lib_beep.c lib_endwin.c lib_color.c lib_options.c lib_slk.c
624 This modules assist in POSIX emulation on non-POSIX systems:
629 The following source files will not be needed for a
630 single-terminal-type port.
632 captoinfo.c clear.c comp_captab.c comp_error.c comp_hash.c comp_main.c
633 comp_parse.c comp_scan.c alloc_entry.c dump_entry.c parse_entry.c
634 read_entry.c write_entry.c infocmp.c tput.c
636 The following modules will use open()/read()/write()/close()/lseek()
637 on files, but no other OS calls.
640 used to read/write screen dumps
643 used to write trace data to the logfile
645 Modules that would have to be modified for a port start here:
647 The following modules are `pure curses' but contain assumptions
648 inappropriate for a memory-mapped port.
650 lib_longname.c -- assumes there may be multiple terminals
651 longname() -- return long name of terminal
652 lib_acs.c -- assumes acs_map as a double indirection
653 init_acs() -- initialize acs map
654 lib_mvcur.c -- assumes cursor moves have variable cost
655 mvcur_init() -- initialize
656 mvcur() -- do physical cursor move
658 scrolln() -- do physical scrolling
659 lib_termcap.c -- assumes there may be multiple terminals
660 tgetent() -- load entry
661 tgetflag() -- get boolean capability
662 tgetnum() -- get numeric capability
663 tgetstr() -- get string capability
664 lib_ti.c -- assumes there may be multiple terminals
665 tigetent() -- load entry
666 tigetflag() -- get boolean capability
667 tigetnum() -- get numeric capability
668 tigetstr() -- get string capability
670 The following modules use UNIX-specific calls:
672 lib_doupdate.c -- input checking
673 doupdate() -- repaint real screen to match virtual
674 _nc_outch() -- put out a single character
675 lib_getch.c -- read()
676 wgetch() -- get single character
677 wungetch() -- push back single character
678 lib_initscr.c -- getenv()
679 initscr() -- initialize curses functions
681 newterm() -- set up new terminal screen
683 baudrate() -- return the baudrate
684 lib_kernel.c -- various tty-manipulation and system calls
685 reset_prog_mode() -- reset ccurses-raw mode
686 reset_shell_mode() -- reset cooked mode
687 erasechar() -- return the erase char
688 killchar() -- return the kill character
689 flushinp() -- flush pending input
690 savetty() -- save tty state
691 resetty() -- reset tty to state at last savetty()
692 lib_raw.c -- various tty-manipulation calls
703 lib_setup.c -- various tty-manipulation calls
706 lib_restart.c -- various tty-manipulation calls
711 lib_tstp.c -- signal-manipulation calls
712 _nc_signal_handler() -- enable/disable window-mode signal catching
713 lib_twait.c -- gettimeofday(), select().
714 usleep() -- microsecond sleep
715 _nc_timed_wait() -- timed wait for input
717 The package kernel could be made smaller.
718 _________________________________________________________________
721 Eric S. Raymond <esr@snark.thyrsus.com>
723 (Note: This is not the bug address!)