1 <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 3.0//EN">
3 $Id: hackguide.html,v 1.25 2000/03/25 18:45:21 tom Exp $
7 <TITLE>A Hacker's Guide to Ncurses Internals</TITLE>
8 <link rev="made" href="mailto:bugs-ncurses@gnu.org">
10 This document is self-contained, *except* that there is one relative link to
11 the ncurses-intro.html document, expected to be in the same directory with
17 <H1>A Hacker's Guide to NCURSES</H1>
21 <LI><A HREF="#abstract">Abstract</A>
22 <LI><A HREF="#objective">Objective of the Package</A>
24 <LI><A HREF="#whysvr4">Why System V Curses?</A>
25 <LI><A HREF="#extensions">How to Design Extensions</A>
27 <LI><A HREF="#portability">Portability and Configuration</A>
28 <LI><A HREF="#documentation">Documentation Conventions</A>
29 <LI><A HREF="#bugtrack">How to Report Bugs</A>
30 <LI><A HREF="#ncurslib">A Tour of the Ncurses Library</A>
32 <LI><A HREF="#loverview">Library Overview</A>
33 <LI><A HREF="#engine">The Engine Room</A>
34 <LI><A HREF="#input">Keyboard Input</A>
35 <LI><A HREF="#mouse">Mouse Events</A>
36 <LI><A HREF="#output">Output and Screen Updating</A>
38 <LI><A HREF="#fmnote">The Forms and Menu Libraries</A>
39 <LI><A HREF="#tic">A Tour of the Terminfo Compiler</A>
41 <LI><A HREF="#nonuse">Translation of Non-<STRONG>use</STRONG> Capabilities</A>
42 <LI><A HREF="#uses">Use Capability Resolution</A>
43 <LI><A HREF="#translation">Source-Form Translation</A>
45 <LI><A HREF="#utils">Other Utilities</A>
46 <LI><A HREF="#style">Style Tips for Developers</A>
47 <LI><A HREF="#port">Porting Hints</A>
50 <H1><A NAME="abstract">Abstract</A></H1>
52 This document is a hacker's tour of the <STRONG>ncurses</STRONG> library and utilities.
53 It discusses design philosophy, implementation methods, and the
54 conventions used for coding and documentation. It is recommended
55 reading for anyone who is interested in porting, extending or improving the
58 <H1><A NAME="objective">Objective of the Package</A></H1>
60 The objective of the <STRONG>ncurses</STRONG> package is to provide a free software API for
61 character-cell terminals and terminal emulators with the following
65 <LI>Source-compatible with historical curses implementations (including
66 the original BSD curses and System V curses.
67 <LI>Conformant with the XSI Curses standard issued as part of XPG4 by
69 <LI>High-quality -- stable and reliable code, wide portability, good
70 packaging, superior documentation.
71 <LI>Featureful -- should eliminate as much of the drudgery of C interface
72 programming as possible, freeing programmers to think at a higher
76 These objectives are in priority order. So, for example, source
77 compatibility with older version must trump featurefulness -- we cannot
78 add features if it means breaking the portion of the API corresponding
79 to historical curses versions.
81 <H2><A NAME="whysvr4">Why System V Curses?</A></H2>
83 We used System V curses as a model, reverse-engineering their API, in
84 order to fulfill the first two objectives. <P>
86 System V curses implementations can support BSD curses programs with
87 just a recompilation, so by capturing the System V API we also
90 More importantly for the future, the XSI Curses standard issued by X/Open
91 is explicitly and closely modeled on System V. So conformance with
92 System V took us most of the way to base-level XSI conformance.
94 <H2><A NAME="extensions">How to Design Extensions</A></H2>
96 The third objective (standards conformance) requires that it be easy to
97 condition source code using <STRONG>ncurses</STRONG> so that the absence of nonstandard
98 extensions does not break the code. <P>
100 Accordingly, we have a policy of associating with each nonstandard extension
101 a feature macro, so that ncurses client code can use this macro to condition
102 in or out the code that requires the <STRONG>ncurses</STRONG> extension. <P>
104 For example, there is a macro <CODE>NCURSES_MOUSE_VERSION</CODE> which XSI Curses
105 does not define, but which is defined in the <STRONG>ncurses</STRONG> library header.
106 You can use this to condition the calls to the mouse API calls.
108 <H1><A NAME="portability">Portability and Configuration</A></H1>
110 Code written for <STRONG>ncurses</STRONG> may assume an ANSI-standard C compiler and
111 POSIX-compatible OS interface. It may also assume the presence of a
112 System-V-compatible <EM>select(2)</EM> call. <P>
114 We encourage (but do not require) developers to make the code friendly
115 to less-capable UNIX environments wherever possible. <P>
117 We encourage developers to support OS-specific optimizations and methods
118 not available under POSIX/ANSI, provided only that:
121 <LI>All such code is properly conditioned so the build process does not
122 attempt to compile it under a plain ANSI/POSIX environment.
123 <LI>Adding such implementation methods does not introduce incompatibilities
124 in the <STRONG>ncurses</STRONG> API between platforms.
127 We use GNU <CODE>autoconf(1)</CODE> as a tool to deal with portability issues.
128 The right way to leverage an OS-specific feature is to modify the autoconf
129 specification files (configure.in and aclocal.m4) to set up a new feature
130 macro, which you then use to condition your code.
132 <H1><A NAME="documentation">Documentation Conventions</A></H1>
134 There are three kinds of documentation associated with this package. Each
135 has a different preferred format:
138 <LI>Package-internal files (README, INSTALL, TO-DO etc.)
140 <LI>Everything else (i.e., narrative documentation).
143 Our conventions are simple:
145 <LI><STRONG>Maintain package-internal files in plain text.</STRONG>
146 The expected viewer for them <EM>more(1)</EM> or an editor window; there's
147 no point in elaborate mark-up.
149 <LI><STRONG>Mark up manual pages in the man macros.</STRONG> These have to be viewable
150 through traditional <EM>man(1)</EM> programs.
152 <LI><STRONG>Write everything else in HTML.</STRONG>
155 When in doubt, HTMLize a master and use <EM>lynx(1)</EM> to generate
156 plain ASCII (as we do for the announcement document). <P>
158 The reason for choosing HTML is that it's (a) well-adapted for on-line
159 browsing through viewers that are everywhere; (b) more easily readable
160 as plain text than most other mark-ups, if you don't have a viewer; and (c)
161 carries enough information that you can generate a nice-looking printed
162 version from it. Also, of course, it make exporting things like the
163 announcement document to WWW pretty trivial.
165 <H1><A NAME="bugtrack">How to Report Bugs</A></H1>
167 The <A NAME="bugreport">reporting address for bugs</A> is
168 <A HREF="mailto:bug-ncurses@gnu.org">bug-ncurses@gnu.org</A>.
169 This is a majordomo list; to join, write
170 to <CODE>bug-ncurses-request@gnu.org</CODE> with a message containing the line:
172 subscribe <name>@<host.domain>
175 The <CODE>ncurses</CODE> code is maintained by a small group of
176 volunteers. While we try our best to fix bugs promptly, we simply
177 don't have a lot of hours to spend on elementary hand-holding. We rely
178 on intelligent cooperation from our users. If you think you have
179 found a bug in <CODE>ncurses</CODE>, there are some steps you can take
180 before contacting us that will help get the bug fixed quickly. <P>
182 In order to use our bug-fixing time efficiently, we put people who
183 show us they've taken these steps at the head of our queue. This
184 means that if you don't, you'll probably end up at the tail end and
185 have to wait a while.
188 <LI>Develop a recipe to reproduce the bug.
190 Bugs we can reproduce are likely to be fixed very quickly, often
191 within days. The most effective single thing you can do to get a
192 quick fix is develop a way we can duplicate the bad behavior --
193 ideally, by giving us source for a small, portable test program that
194 breaks the library. (Even better is a keystroke recipe using one of
195 the test programs provided with the distribution.)
197 <LI>Try to reproduce the bug on a different terminal type. <P>
199 In our experience, most of the behaviors people report as library bugs
200 are actually due to subtle problems in terminal descriptions. This is
201 especially likely to be true if you're using a traditional
202 asynchronous terminal or PC-based terminal emulator, rather than xterm
203 or a UNIX console entry. <P>
205 It's therefore extremely helpful if you can tell us whether or not your
206 problem reproduces on other terminal types. Usually you'll have both
207 a console type and xterm available; please tell us whether or not your
208 bug reproduces on both. <P>
210 If you have xterm available, it is also good to collect xterm reports for
211 different window sizes. This is especially true if you normally use an
212 unusual xterm window size -- a surprising number of the bugs we've seen
213 are either triggered or masked by these.
215 <LI>Generate and examine a trace file for the broken behavior. <P>
217 Recompile your program with the debugging versions of the libraries.
218 Insert a <CODE>trace()</CODE> call with the argument set to <CODE>TRACE_UPDATE</CODE>.
219 (See <A HREF="ncurses-intro.html#debugging">"Writing Programs with
220 NCURSES"</A> for details on trace levels.)
221 Reproduce your bug, then look at the trace file to see what the library
222 was actually doing. <P>
224 Another frequent cause of apparent bugs is application coding errors
225 that cause the wrong things to be put on the virtual screen. Looking
226 at the virtual-screen dumps in the trace file will tell you immediately if
227 this is happening, and save you from the possible embarrassment of being
228 told that the bug is in your code and is your problem rather than ours. <P>
230 If the virtual-screen dumps look correct but the bug persists, it's
231 possible to crank up the trace level to give more and more information
232 about the library's update actions and the control sequences it issues
233 to perform them. The test directory of the distribution contains a
234 tool for digesting these logs to make them less tedious to wade
237 Often you'll find terminfo problems at this stage by noticing that the
238 escape sequences put out for various capabilities are wrong. If not,
239 you're likely to learn enough to be able to characterize any bug in
240 the screen-update logic quite exactly.
242 <LI>Report details and symptoms, not just interpretations. <P>
244 If you do the preceding two steps, it is very likely that you'll discover
245 the nature of the problem yourself and be able to send us a fix. This
246 will create happy feelings all around and earn you good karma for the first
247 time you run into a bug you really can't characterize and fix yourself. <P>
249 If you're still stuck, at least you'll know what to tell us. Remember, we
250 need details. If you guess about what is safe to leave out, you are too
251 likely to be wrong. <P>
253 If your bug produces a bad update, include a trace file. Try to make
254 the trace at the <EM>least</EM> voluminous level that pins down the
255 bug. Logs that have been through tracemunch are OK, it doesn't throw
256 away any information (actually they're better than un-munched ones because
257 they're easier to read). <P>
259 If your bug produces a core-dump, please include a symbolic stack trace
260 generated by gdb(1) or your local equivalent. <P>
262 Tell us about every terminal on which you've reproduced the bug -- and
263 every terminal on which you can't. Ideally, sent us terminfo sources
264 for all of these (yours might differ from ours). <P>
266 Include your ncurses version and your OS/machine type, of course! You can
267 find your ncurses version in the <CODE>curses.h</CODE> file.
270 If your problem smells like a logic error or in cursor movement or
271 scrolling or a bad capability, there are a couple of tiny test frames
272 for the library algorithms in the progs directory that may help you
273 isolate it. These are not part of the normal build, but do have their
274 own make productions. <P>
276 The most important of these is <CODE>mvcur</CODE>, a test frame for the
277 cursor-movement optimization code. With this program, you can see
278 directly what control sequences will be emitted for any given cursor
279 movement or scroll/insert/delete operations. If you think you've got
280 a bad capability identified, you can disable it and test again. The
281 program is command-driven and has on-line help. <P>
283 If you think the vertical-scroll optimization is broken, or just want to
284 understand how it works better, build <CODE>hashmap</CODE> and read the
285 header comments of <CODE>hardscroll.c</CODE> and <CODE>hashmap.c</CODE>; then try
286 it out. You can also test the hardware-scrolling optimization separately
287 with <CODE>hardscroll</CODE>. <P>
289 There's one other interactive tester, <CODE>tctest</CODE>, that exercises
290 translation between termcap and terminfo formats. If you have a serious
291 need to run this, you probably belong on our development team!
293 <H1><A NAME="ncurslib">A Tour of the Ncurses Library</A></H1>
295 <H2><A NAME="loverview">Library Overview</A></H2>
297 Most of the library is superstructure -- fairly trivial convenience
298 interfaces to a small set of basic functions and data structures used
299 to manipulate the virtual screen (in particular, none of this code
300 does any I/O except through calls to more fundamental modules
301 described below). The files
363 are all in this category. They are very
364 unlikely to need change, barring bugs or some fundamental
365 reorganization in the underlying data structures. <P>
367 These files are used only for debugging support:
379 It is rather unlikely you will ever need to change these, unless
380 you want to introduce a new debug trace level for some reasoon.<P>
382 There is another group of files that do direct I/O via <EM>tputs()</EM>,
383 computations on the terminal capabilities, or queries to the OS
384 environment, but nevertheless have only fairly low complexity. These
404 They are likely to need revision only if
405 ncurses is being ported to an environment without an underlying
406 terminfo capability representation. <P>
409 have serious hooks into
410 the tty driver and signal facilities:
420 If you run into porting snafus
421 moving the package to another UNIX, the problem is likely to be in one
423 The file <CODE>lib_print.c</CODE> uses sleep(2) and also
424 falls in this category.<P>
426 Almost all of the real work is done in the files
441 Most of the algorithmic complexity in the
442 library lives in these files.
443 If there is a real bug in <STRONG>ncurses</STRONG> itself, it's probably here.
444 We'll tour some of these files in detail
445 below (see <A HREF="#engine">The Engine Room</A>). <P>
447 Finally, there is a group of files that is actually most of the
448 terminfo compiler. The reason this code lives in the <STRONG>ncurses</STRONG>
449 library is to support fallback to /etc/termcap. These files include
464 We'll discuss these in the compiler tour.
466 <H2><A NAME="engine">The Engine Room</A></H2>
468 <H3><A NAME="input">Keyboard Input</A></H3>
470 All <CODE>ncurses</CODE> input funnels through the function
471 <CODE>wgetch()</CODE>, defined in <CODE>lib_getch.c</CODE>. This function is
472 tricky; it has to poll for keyboard and mouse events and do a running
473 match of incoming input against the set of defined special keys. <P>
475 The central data structure in this module is a FIFO queue, used to
476 match multiple-character input sequences against special-key
477 capabilities; also to implement pushback via <CODE>ungetch()</CODE>. <P>
479 The <CODE>wgetch()</CODE> code distinguishes between function key
480 sequences and the same sequences typed manually by doing a timed wait
481 after each input character that could lead a function key sequence.
482 If the entire sequence takes less than 1 second, it is assumed to have
483 been generated by a function key press. <P>
485 Hackers bruised by previous encounters with variant <CODE>select(2)</CODE>
486 calls may find the code in <CODE>lib_twait.c</CODE> interesting. It deals
487 with the problem that some BSD selects don't return a reliable
488 time-left value. The function <CODE>timed_wait()</CODE> effectively
489 simulates a System V select.
491 <H3><A NAME="mouse">Mouse Events</A></H3>
493 If the mouse interface is active, <CODE>wgetch()</CODE> polls for mouse
494 events each call, before it goes to the keyboard for input. It is
495 up to <CODE>lib_mouse.c</CODE> how the polling is accomplished; it may vary
496 for different devices. <P>
498 Under xterm, however, mouse event notifications come in via the keyboard
499 input stream. They are recognized by having the <STRONG>kmous</STRONG> capability
500 as a prefix. This is kind of klugey, but trying to wire in recognition of
501 a mouse key prefix without going through the function-key machinery would
502 be just too painful, and this turns out to imply having the prefix somewhere
503 in the function-key capabilities at terminal-type initialization. <P>
505 This kluge only works because <STRONG>kmous</STRONG> isn't actually used by any
506 historic terminal type or curses implementation we know of. Best
507 guess is it's a relic of some forgotten experiment in-house at Bell
508 Labs that didn't leave any traces in the publicly-distributed System V
509 terminfo files. If System V or XPG4 ever gets serious about using it
510 again, this kluge may have to change. <P>
512 Here are some more details about mouse event handling: <P>
514 The <CODE>lib_mouse()</CODE>code is logically split into a lower level that
515 accepts event reports in a device-dependent format and an upper level that
516 parses mouse gestures and filters events. The mediating data structure is a
517 circular queue of event structures. <P>
519 Functionally, the lower level's job is to pick up primitive events and
520 put them on the circular queue. This can happen in one of two ways:
521 either (a) <CODE>_nc_mouse_event()</CODE> detects a series of incoming
522 mouse reports and queues them, or (b) code in <CODE>lib_getch.c</CODE> detects the
523 <STRONG>kmous</STRONG> prefix in the keyboard input stream and calls _nc_mouse_inline
524 to queue up a series of adjacent mouse reports. <P>
526 In either case, <CODE>_nc_mouse_parse()</CODE> should be called after the
527 series is accepted to parse the digested mouse reports (low-level
528 events) into a gesture (a high-level or composite event).
530 <H3><A NAME="output">Output and Screen Updating</A></H3>
532 With the single exception of character echoes during a <CODE>wgetnstr()</CODE>
533 call (which simulates cooked-mode line editing in an ncurses window),
534 the library normally does all its output at refresh time. <P>
536 The main job is to go from the current state of the screen (as represented
537 in the <CODE>curscr</CODE> window structure) to the desired new state (as
538 represented in the <CODE>newscr</CODE> window structure), while doing as
539 little I/O as possible. <P>
541 The brains of this operation are the modules <CODE>hashmap.c</CODE>,
542 <CODE>hardscroll.c</CODE> and <CODE>lib_doupdate.c</CODE>; the latter two use
543 <CODE>lib_mvcur.c</CODE>. Essentially, what happens looks like this: <P>
545 The <CODE>hashmap.c</CODE> module tries to detect vertical motion
546 changes between the real and virtual screens. This information
547 is represented by the oldindex members in the newscr structure.
548 These are modified by vertical-motion and clear operations, and both are
549 re-initialized after each update. To this change-journalling
550 information, the hashmap code adds deductions made using a modified Heckel
551 algorithm on hash values generated from the line contents. <P>
553 The <CODE>hardscroll.c</CODE> module computes an optimum set of scroll,
554 insertion, and deletion operations to make the indices match. It calls
555 <CODE>_nc_mvcur_scrolln()</CODE> in <CODE>lib_mvcur.c</CODE> to do those motions. <P>
557 Then <CODE>lib_doupdate.c</CODE> goes to work. Its job is to do line-by-line
558 transformations of <CODE>curscr</CODE> lines to <CODE>newscr</CODE> lines. Its main
559 tool is the routine <CODE>mvcur()</CODE> in <CODE>lib_mvcur.c</CODE>. This routine
560 does cursor-movement optimization, attempting to get from given screen
561 location A to given location B in the fewest output characters posible. <P>
563 If you want to work on screen optimizations, you should use the fact
564 that (in the trace-enabled version of the library) enabling the
565 <CODE>TRACE_TIMES</CODE> trace level causes a report to be emitted after
566 each screen update giving the elapsed time and a count of characters
567 emitted during the update. You can use this to tell when an update
568 optimization improves efficiency. <P>
570 In the trace-enabled version of the library, it is also possible to disable
571 and re-enable various optimizations at runtime by tweaking the variable
572 <CODE>_nc_optimize_enable</CODE>. See the file <CODE>include/curses.h.in</CODE>
573 for mask values, near the end.
575 <H1><A NAME="fmnote">The Forms and Menu Libraries</A></H1>
577 The forms and menu libraries should work reliably in any environment you
578 can port ncurses to. The only portability issue anywhere in them is what
579 flavor of regular expressions the built-in form field type TYPE_REGEXP
582 The configuration code prefers the POSIX regex facility, modeled on
583 System V's, but will settle for BSD regexps if the former isn't available. <P>
585 Historical note: the panels code was written primarily to assist in
586 porting u386mon 2.0 (comp.sources.misc v14i001-4) to systems lacking
587 panels support; u386mon 2.10 and beyond use it. This version has been
588 slightly cleaned up for <CODE>ncurses</CODE>.
590 <H1><A NAME="tic">A Tour of the Terminfo Compiler</A></H1>
592 The <STRONG>ncurses</STRONG> implementation of <STRONG>tic</STRONG> is rather complex
593 internally; it has to do a trying combination of missions. This starts
594 with the fact that, in addition to its normal duty of compiling
595 terminfo sources into loadable terminfo binaries, it has to be able to
596 handle termcap syntax and compile that too into terminfo entries. <P>
598 The implementation therefore starts with a table-driven, dual-mode
599 lexical analyzer (in <CODE>comp_scan.c</CODE>). The lexer chooses its
600 mode (termcap or terminfo) based on the first `,' or `:' it finds in
601 each entry. The lexer does all the work of recognizing capability
602 names and values; the grammar above it is trivial, just "parse entries
603 till you run out of file".
605 <H2><A NAME="nonuse">Translation of Non-<STRONG>use</STRONG> Capabilities</A></H2>
607 Translation of most things besides <STRONG>use</STRONG> capabilities is pretty
608 straightforward. The lexical analyzer's tokenizer hands each capability
609 name to a hash function, which drives a table lookup. The table entry
610 yields an index which is used to look up the token type in another table,
611 and controls interpretation of the value. <P>
613 One possibly interesting aspect of the implementation is the way the
614 compiler tables are initialized. All the tables are generated by various
615 awk/sed/sh scripts from a master table <CODE>include/Caps</CODE>; these
616 scripts actually write C initializers which are linked to the compiler.
617 Furthermore, the hash table is generated in the same way, so it doesn't
618 have to be generated at compiler startup time (another benefit of this
619 organization is that the hash table can be in shareable text space). <P>
621 Thus, adding a new capability is usually pretty trivial, just a matter
622 of adding one line to the <CODE>include/Caps</CODE> file. We'll have more
623 to say about this in the section on <A HREF="#translation">Source-Form
626 <H2><A NAME="uses">Use Capability Resolution</A></H2>
628 The background problem that makes <STRONG>tic</STRONG> tricky isn't the capability
629 translation itself, it's the resolution of <STRONG>use</STRONG> capabilities. Older
630 versions would not handle forward <STRONG>use</STRONG> references for this reason
631 (that is, a using terminal always had to follow its use target in the
632 source file). By doing this, they got away with a simple implementation
633 tactic; compile everything as it blows by, then resolve uses from compiled
636 This won't do for <STRONG>ncurses</STRONG>. The problem is that that the whole
637 compilation process has to be embeddable in the <STRONG>ncurses</STRONG> library
638 so that it can be called by the startup code to translate termcap
639 entries on the fly. The embedded version can't go promiscuously writing
640 everything it translates out to disk -- for one thing, it will typically
641 be running with non-root permissions. <P>
643 So our <STRONG>tic</STRONG> is designed to parse an entire terminfo file into a
644 doubly-linked circular list of entry structures in-core, and then do
645 <STRONG>use</STRONG> resolution in-memory before writing everything out. This
646 design has other advantages: it makes forward and back use-references
647 equally easy (so we get the latter for free), and it makes checking for
648 name collisions before they're written out easy to do. <P>
650 And this is exactly how the embedded version works. But the stand-alone
651 user-accessible version of <STRONG>tic</STRONG> partly reverts to the historical
652 strategy; it writes to disk (not keeping in core) any entry with no
653 <STRONG>use</STRONG> references. <P>
655 This is strictly a core-economy kluge, implemented because the
656 terminfo master file is large enough that some core-poor systems swap
657 like crazy when you compile it all in memory...there have been reports of
658 this process taking <STRONG>three hours</STRONG>, rather than the twenty seconds
659 or less typical on the author's development box. <P>
661 So. The executable <STRONG>tic</STRONG> passes the entry-parser a hook that
662 <EM>immediately</EM> writes out the referenced entry if it has no use
663 capabilities. The compiler main loop refrains from adding the entry
664 to the in-core list when this hook fires. If some other entry later
665 needs to reference an entry that got written immediately, that's OK;
666 the resolution code will fetch it off disk when it can't find it in
669 Name collisions will still be detected, just not as cleanly. The
670 <CODE>write_entry()</CODE> code complains before overwriting an entry that
671 postdates the time of <STRONG>tic</STRONG>'s first call to
672 <CODE>write_entry()</CODE>, Thus it will complain about overwriting
673 entries newly made during the <STRONG>tic</STRONG> run, but not about
674 overwriting ones that predate it.
676 <H2><A NAME="translation">Source-Form Translation</A></H2>
678 Another use of <STRONG>tic</STRONG> is to do source translation between various termcap
679 and terminfo formats. There are more variants out there than you might
680 think; the ones we know about are described in the <STRONG>captoinfo(1)</STRONG>
683 The translation output code (<CODE>dump_entry()</CODE> in
684 <CODE>ncurses/dump_entry.c</CODE>) is shared with the <STRONG>infocmp(1)</STRONG>
685 utility. It takes the same internal representation used to generate
686 the binary form and dumps it to standard output in a specified
689 The <CODE>include/Caps</CODE> file has a header comment describing ways you
690 can specify source translations for nonstandard capabilities just by
691 altering the master table. It's possible to set up capability aliasing
692 or tell the compiler to plain ignore a given capability without writing
693 any C code at all. <P>
695 For circumstances where you need to do algorithmic translation, there
696 are functions in <CODE>parse_entry.c</CODE> called after the parse of each
697 entry that are specifically intended to encapsulate such
698 translations. This, for example, is where the AIX <STRONG>box1</STRONG> capability
699 get translated to an <STRONG>acsc</STRONG> string.
701 <H1><A NAME="utils">Other Utilities</A></H1>
703 The <STRONG>infocmp</STRONG> utility is just a wrapper around the same
704 entry-dumping code used by <STRONG>tic</STRONG> for source translation. Perhaps
705 the one interesting aspect of the code is the use of a predicate
706 function passed in to <CODE>dump_entry()</CODE> to control which
707 capabilities are dumped. This is necessary in order to handle both
708 the ordinary De-compilation case and entry difference reporting. <P>
710 The <STRONG>tput</STRONG> and <STRONG>clear</STRONG> utilities just do an entry load
711 followed by a <CODE>tputs()</CODE> of a selected capability.
713 <H1><A NAME="style">Style Tips for Developers</A></H1>
715 See the TO-DO file in the top-level directory of the source distribution
716 for additions that would be particularly useful. <P>
718 The prefix <CODE>_nc_</CODE> should be used on library public functions that are
719 not part of the curses API in order to prevent pollution of the
720 application namespace.
722 If you have to add to or modify the function prototypes in curses.h.in,
723 read ncurses/MKlib_gen.sh first so you can avoid breaking XSI conformance.
725 Please join the ncurses mailing list. See the INSTALL file in the
726 top level of the distribution for details on the list. <P>
728 Look for the string <CODE>FIXME</CODE> in source files to tag minor bugs
729 and potential problems that could use fixing. <P>
731 Don't try to auto-detect OS features in the main body of the C code.
732 That's the job of the configuration system. <P>
734 To hold down complexity, do make your code data-driven. Especially,
735 if you can drive logic from a table filtered out of
736 <CODE>include/Caps</CODE>, do it. If you find you need to augment the
737 data in that file in order to generate the proper table, that's still
738 preferable to ad-hoc code -- that's why the fifth field (flags) is
743 <H1><A NAME="port">Porting Hints</A></H1>
745 The following notes are intended to be a first step towards DOS and Macintosh
746 ports of the ncurses libraries. <P>
748 The following library modules are `pure curses'; they operate only on
749 the curses internal structures, do all output through other curses
750 calls (not including <CODE>tputs()</CODE> and <CODE>putp()</CODE>) and do not
751 call any other UNIX routines such as signal(2) or the stdio library.
752 Thus, they should not need to be modified for single-terminal
793 This module is pure curses, but calls outstr():
802 These modules are pure curses, except that they use <CODE>tputs()</CODE>
803 and <CODE>putp()</CODE>:
817 This modules assist in POSIX emulation on non-POSIX systems:
823 The following source files will not be needed for a
824 single-terminal-type port.
847 The following modules will use open()/read()/write()/close()/lseek() on files,
848 but no other OS calls.
852 <DD>used to read/write screen dumps
854 <DD>used to write trace data to the logfile
857 Modules that would have to be modified for a port start here: <P>
859 The following modules are `pure curses' but contain assumptions inappropriate
860 for a memory-mapped port.
863 <dt>lib_longname.c<dd>assumes there may be multiple terminals
864 <dt>lib_acs.c<dd>assumes acs_map as a double indirection
865 <dt>lib_mvcur.c<dd>assumes cursor moves have variable cost
866 <dt>lib_termcap.c<dd>assumes there may be multiple terminals
867 <dt>lib_ti.c<dd>assumes there may be multiple terminals
870 The following modules use UNIX-specific calls:
873 <dt>lib_doupdate.c<dd>input checking
874 <dt>lib_getch.c<dd>read()
875 <dt>lib_initscr.c<dd>getenv()
878 <dt>lib_kernel.c<dd>various tty-manipulation and system calls
879 <dt>lib_raw.c<dd>various tty-manipulation calls
880 <dt>lib_setup.c<dd>various tty-manipulation calls
881 <dt>lib_restart.c<dd>various tty-manipulation calls
882 <dt>lib_tstp.c<dd>signal-manipulation calls
883 <dt>lib_twait.c<dd>gettimeofday(), select().
887 <ADDRESS>Eric S. Raymond <esr@snark.thyrsus.com></ADDRESS>
888 (Note: This is <EM>not</EM> the <A HREF="#bugtrack">bug address</A>!)