Google Git
Sign in
android / platform / external / ncurses / 0dbb80ad4c2ff1fec2c1c67c2d1b8a2d539edf33 / . / doc / html / hackguide.html
blob: 71312a565f4c5c9d68cae44f7b9e70e95a59adba [file] [log] [blame]
<!--
$Id: hackguide.html,v 1.33 2020/02/02 23:34:34 tom Exp $
****************************************************************************
* Copyright 2019,2020 Thomas E. Dickey *
* Copyright 2000-2013,2017 Free Software Foundation, Inc. *
* *
* Permission is hereby granted, free of charge, to any person obtaining a *
* copy of this software and associated documentation files (the *
* "Software"), to deal in the Software without restriction, including *
* without limitation the rights to use, copy, modify, merge, publish, *
* distribute, distribute with modifications, sublicense, and/or sell *
* copies of the Software, and to permit persons to whom the Software is *
* furnished to do so, subject to the following conditions: *
* *
* The above copyright notice and this permission notice shall be included *
* in all copies or substantial portions of the Software. *
* *
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS *
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF *
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. *
* IN NO EVENT SHALL THE ABOVE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, *
* DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR *
* OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR *
* THE USE OR OTHER DEALINGS IN THE SOFTWARE. *
* *
* Except as contained in this notice, the name(s) of the above copyright *
* holders shall not be used in advertising or otherwise to promote the *
* sale, use or other dealings in this Software without prior written *
* authorization. *
****************************************************************************
-->
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN">
<html>
<head>
<meta name="generator" content=
"HTML Tidy for HTML5 for Linux version 5.2.0">
<title>A Hacker's Guide to Ncurses Internals</title>
<link rel="author" href="mailto:bugs-ncurses@gnu.org">
<meta http-equiv="Content-Type" content=
"text/html; charset=us-ascii"><!--
This document is self-contained, *except* that there is one relative link to
the ncurses-intro.html document, expected to be in the same directory with
this one.
-->
</head>
<body>
<h1>A Hacker's Guide to NCURSES</h1>
<h1>Contents</h1>
<ul>
<li><a href="#abstract">Abstract</a></li>
<li>
<a href="#objective">Objective of the Package</a>
<ul>
<li><a href="#whysvr4">Why System V Curses?</a></li>
<li><a href="#extensions">How to Design Extensions</a></li>
</ul>
</li>
<li><a href="#portability">Portability and Configuration</a></li>
<li><a href="#documentation">Documentation Conventions</a></li>
<li><a href="#bugtrack">How to Report Bugs</a></li>
<li>
<a href="#ncurslib">A Tour of the Ncurses Library</a>
<ul>
<li><a href="#loverview">Library Overview</a></li>
<li><a href="#engine">The Engine Room</a></li>
<li><a href="#input">Keyboard Input</a></li>
<li><a href="#mouse">Mouse Events</a></li>
<li><a href="#output">Output and Screen Updating</a></li>
</ul>
</li>
<li><a href="#fmnote">The Forms and Menu Libraries</a></li>
<li>
<a href="#tic">A Tour of the Terminfo Compiler</a>
<ul>
<li><a href="#nonuse">Translation of
Non-<strong>use</strong> Capabilities</a></li>
<li><a href="#uses">Use Capability Resolution</a></li>
<li><a href="#translation">Source-Form Translation</a></li>
</ul>
</li>
<li><a href="#utils">Other Utilities</a></li>
<li><a href="#style">Style Tips for Developers</a></li>
<li><a href="#port">Porting Hints</a></li>
</ul>
<h1><a name="abstract" id="abstract">Abstract</a></h1>
<p>This document is a hacker's tour of the
<strong>ncurses</strong> library and utilities. It discusses
design philosophy, implementation methods, and the conventions
used for coding and documentation. It is recommended reading for
anyone who is interested in porting, extending or improving the
package.</p>
<h1><a name="objective" id="objective">Objective of the
Package</a></h1>
<p>The objective of the <strong>ncurses</strong> package is to
provide a free software API for character-cell terminals and
terminal emulators with the following characteristics:</p>
<ul>
<li>Source-compatible with historical curses implementations
(including the original BSD curses and System V curses.</li>
<li>Conformant with the XSI Curses standard issued as part of
XPG4 by X/Open.</li>
<li>High-quality &mdash; stable and reliable code, wide
portability, good packaging, superior documentation.</li>
<li>Featureful &mdash; should eliminate as much of the drudgery
of C interface programming as possible, freeing programmers to
think at a higher level of design.</li>
</ul>
<p>These objectives are in priority order. So, for example,
source compatibility with older version must trump featurefulness
&mdash; we cannot add features if it means breaking the portion
of the API corresponding to historical curses versions.</p>
<h2><a name="whysvr4" id="whysvr4">Why System V Curses?</a></h2>
<p>We used System V curses as a model, reverse-engineering their
API, in order to fulfill the first two objectives.</p>
<p>System V curses implementations can support BSD curses
programs with just a recompilation, so by capturing the System V
API we also capture BSD's.</p>
<p>More importantly for the future, the XSI Curses standard
issued by X/Open is explicitly and closely modeled on System V.
So conformance with System V took us most of the way to
base-level XSI conformance.</p>
<h2><a name="extensions" id="extensions">How to Design
Extensions</a></h2>
<p>The third objective (standards conformance) requires that it
be easy to condition source code using <strong>ncurses</strong>
so that the absence of nonstandard extensions does not break the
code.</p>
<p>Accordingly, we have a policy of associating with each
nonstandard extension a feature macro, so that ncurses client
code can use this macro to condition in or out the code that
requires the <strong>ncurses</strong> extension.</p>
<p>For example, there is a macro
<code>NCURSES_MOUSE_VERSION</code> which XSI Curses does not
define, but which is defined in the <strong>ncurses</strong>
library header. You can use this to condition the calls to the
mouse API calls.</p>
<h1><a name="portability" id="portability">Portability and
Configuration</a></h1>
<p>Code written for <strong>ncurses</strong> may assume an
ANSI-standard C compiler and POSIX-compatible OS interface. It
may also assume the presence of a System-V-compatible
<em>select(2)</em> call.</p>
<p>We encourage (but do not require) developers to make the code
friendly to less-capable UNIX environments wherever possible.</p>
<p>We encourage developers to support OS-specific optimizations
and methods not available under POSIX/ANSI, provided only
that:</p>
<ul>
<li>All such code is properly conditioned so the build process
does not attempt to compile it under a plain ANSI/POSIX
environment.</li>
<li>Adding such implementation methods does not introduce
incompatibilities in the <strong>ncurses</strong> API between
platforms.</li>
</ul>
<p>We use GNU <code>autoconf(1)</code> as a tool to deal with
portability issues. The right way to leverage an OS-specific
feature is to modify the autoconf specification files
(configure.in and aclocal.m4) to set up a new feature macro,
which you then use to condition your code.</p>
<h1><a name="documentation" id="documentation">Documentation
Conventions</a></h1>
<p>There are three kinds of documentation associated with this
package. Each has a different preferred format:</p>
<ul>
<li>Package-internal files (README, INSTALL, TO-DO etc.)</li>
<li>Manual pages.</li>
<li>Everything else (i.e., narrative documentation).</li>
</ul>
<p>Our conventions are simple:</p>
<ol>
<li><strong>Maintain package-internal files in plain
text.</strong> The expected viewer for them <em>more(1)</em> or
an editor window; there is no point in elaborate mark-up.</li>
<li><strong>Mark up manual pages in the man macros.</strong>
These have to be viewable through traditional <em>man(1)</em>
programs.</li>
<li><strong>Write everything else in HTML.</strong>
</li>
</ol>
<p>When in doubt, HTMLize a master and use <em>lynx(1)</em> to
generate plain ASCII (as we do for the announcement
document).</p>
<p>The reason for choosing HTML is that it is (a) well-adapted
for on-line browsing through viewers that are everywhere; (b)
more easily readable as plain text than most other mark-ups, if
you do not have a viewer; and (c) carries enough information that
you can generate a nice-looking printed version from it. Also, of
course, it make exporting things like the announcement document
to WWW pretty trivial.</p>
<h1><a name="bugtrack" id="bugtrack">How to Report Bugs</a></h1>
<p>The <a name="bugreport" id="bugreport">reporting address for
bugs</a> is <a href=
"mailto:bug-ncurses@gnu.org">bug-ncurses@gnu.org</a>. This is a
majordomo list; to join, write to
<code>bug-ncurses-request@gnu.org</code> with a message
containing the line:</p>
<pre>
subscribe &lt;name&gt;@&lt;host.domain&gt;
</pre>
<p>The <code>ncurses</code> code is maintained by a small group
of volunteers. While we try our best to fix bugs promptly, we
simply do not have a lot of hours to spend on elementary
hand-holding. We rely on intelligent cooperation from our users.
If you think you have found a bug in <code>ncurses</code>, there
are some steps you can take before contacting us that will help
get the bug fixed quickly.</p>
<p>In order to use our bug-fixing time efficiently, we put people
who show us they have taken these steps at the head of our queue.
This means that if you do not, you will probably end up at the
tail end and have to wait a while.</p>
<ol>
<li>Develop a recipe to reproduce the bug.
<p>Bugs we can reproduce are likely to be fixed very quickly,
often within days. The most effective single thing you can do
to get a quick fix is develop a way we can duplicate the bad
behavior &mdash; ideally, by giving us source for a small,
portable test program that breaks the library. (Even better
is a keystroke recipe using one of the test programs provided
with the distribution.)</p>
</li>
<li>Try to reproduce the bug on a different terminal type.
<p>In our experience, most of the behaviors people report as
library bugs are actually due to subtle problems in terminal
descriptions. This is especially likely to be true if you are
using a traditional asynchronous terminal or PC-based
terminal emulator, rather than xterm or a UNIX console
entry.</p>
<p>It is therefore extremely helpful if you can tell us
whether or not your problem reproduces on other terminal
types. Usually you will have both a console type and xterm
available; please tell us whether or not your bug reproduces
on both.</p>
<p>If you have xterm available, it is also good to collect
xterm reports for different window sizes. This is especially
true if you normally use an unusual xterm window size &mdash;
a surprising number of the bugs we have seen are either
triggered or masked by these.</p>
</li>
<li>Generate and examine a trace file for the broken behavior.
<p>Recompile your program with the debugging versions of the
libraries. Insert a <code>trace()</code> call with the
argument set to <code>TRACE_UPDATE</code>. (See <a href=
"ncurses-intro.html#debugging">"Writing Programs with
NCURSES"</a> for details on trace levels.) Reproduce your
bug, then look at the trace file to see what the library was
actually doing.</p>
<p>Another frequent cause of apparent bugs is application
coding errors that cause the wrong things to be put on the
virtual screen. Looking at the virtual-screen dumps in the
trace file will tell you immediately if this is happening,
and save you from the possible embarrassment of being told
that the bug is in your code and is your problem rather than
ours.</p>
<p>If the virtual-screen dumps look correct but the bug
persists, it is possible to crank up the trace level to give
more and more information about the library's update actions
and the control sequences it issues to perform them. The test
directory of the distribution contains a tool for digesting
these logs to make them less tedious to wade through.</p>
<p>Often you will find terminfo problems at this stage by
noticing that the escape sequences put out for various
capabilities are wrong. If not, you are likely to learn
enough to be able to characterize any bug in the
screen-update logic quite exactly.</p>
</li>
<li>Report details and symptoms, not just interpretations.
<p>If you do the preceding two steps, it is very likely that
you will discover the nature of the problem yourself and be
able to send us a fix. This will create happy feelings all
around and earn you good karma for the first time you run
into a bug you really cannot characterize and fix
yourself.</p>
<p>If you are still stuck, at least you will know what to
tell us. Remember, we need details. If you guess about what
is safe to leave out, you are too likely to be wrong.</p>
<p>If your bug produces a bad update, include a trace file.
Try to make the trace at the <em>least</em> voluminous level
that pins down the bug. Logs that have been through
tracemunch are OK, it does not throw away any information
(actually they are better than un-munched ones because they
are easier to read).</p>
<p>If your bug produces a core-dump, please include a
symbolic stack trace generated by gdb(1) or your local
equivalent.</p>
<p>Tell us about every terminal on which you have reproduced
the bug &mdash; and every terminal on which you cannot.
Ideally, send us terminfo sources for all of these (yours
might differ from ours).</p>
<p>Include your ncurses version and your OS/machine type, of
course! You can find your ncurses version in the
<code>curses.h</code> file.</p>
</li>
</ol>
<p>If your problem smells like a logic error or in cursor
movement or scrolling or a bad capability, there are a couple of
tiny test frames for the library algorithms in the progs
directory that may help you isolate it. These are not part of the
normal build, but do have their own make productions.</p>
<p>The most important of these is <code>mvcur</code>, a test
frame for the cursor-movement optimization code. With this
program, you can see directly what control sequences will be
emitted for any given cursor movement or scroll/insert/delete
operations. If you think you have got a bad capability
identified, you can disable it and test again. The program is
command-driven and has on-line help.</p>
<p>If you think the vertical-scroll optimization is broken, or
just want to understand how it works better, build
<code>hashmap</code> and read the header comments of
<code>hardscroll.c</code> and <code>hashmap.c</code>; then try it
out. You can also test the hardware-scrolling optimization
separately with <code>hardscroll</code>.</p>
<h1><a name="ncurslib" id="ncurslib">A Tour of the Ncurses
Library</a></h1>
<h2><a name="loverview" id="loverview">Library Overview</a></h2>
<p>Most of the library is superstructure &mdash; fairly trivial
convenience interfaces to a small set of basic functions and data
structures used to manipulate the virtual screen (in particular,
none of this code does any I/O except through calls to more
fundamental modules described below). The files</p>
<blockquote>
<code>lib_addch.c lib_bkgd.c lib_box.c lib_chgat.c lib_clear.c
lib_clearok.c lib_clrbot.c lib_clreol.c lib_colorset.c
lib_data.c lib_delch.c lib_delwin.c lib_echo.c lib_erase.c
lib_gen.c lib_getstr.c lib_hline.c lib_immedok.c lib_inchstr.c
lib_insch.c lib_insdel.c lib_insstr.c lib_instr.c
lib_isendwin.c lib_keyname.c lib_leaveok.c lib_move.c
lib_mvwin.c lib_overlay.c lib_pad.c lib_printw.c lib_redrawln.c
lib_scanw.c lib_screen.c lib_scroll.c lib_scrollok.c
lib_scrreg.c lib_set_term.c lib_slk.c lib_slkatr_set.c
lib_slkatrof.c lib_slkatron.c lib_slkatrset.c lib_slkattr.c
lib_slkclear.c lib_slkcolor.c lib_slkinit.c lib_slklab.c
lib_slkrefr.c lib_slkset.c lib_slktouch.c lib_touch.c
lib_unctrl.c lib_vline.c lib_wattroff.c lib_wattron.c
lib_window.c</code>
</blockquote>
<p>are all in this category. They are very unlikely to need
change, barring bugs or some fundamental reorganization in the
underlying data structures.</p>
<p>These files are used only for debugging support:</p>
<blockquote>
<code>lib_trace.c lib_traceatr.c lib_tracebits.c lib_tracechr.c
lib_tracedmp.c lib_tracemse.c trace_buf.c</code>
</blockquote>
<p>It is rather unlikely you will ever need to change these,
unless you want to introduce a new debug trace level for some
reason.</p>
<p>There is another group of files that do direct I/O via
<em>tputs()</em>, computations on the terminal capabilities, or
queries to the OS environment, but nevertheless have only fairly
low complexity. These include:</p>
<blockquote>
<code>lib_acs.c lib_beep.c lib_color.c lib_endwin.c
lib_initscr.c lib_longname.c lib_newterm.c lib_options.c
lib_termcap.c lib_ti.c lib_tparm.c lib_tputs.c lib_vidattr.c
read_entry.c.</code>
</blockquote>
<p>They are likely to need revision only if ncurses is being
ported to an environment without an underlying terminfo
capability representation.</p>
<p>These files have serious hooks into the tty driver and signal
facilities:</p>
<blockquote>
<code>lib_kernel.c lib_baudrate.c lib_raw.c lib_tstp.c
lib_twait.c</code>
</blockquote>
<p>If you run into porting snafus moving the package to another
UNIX, the problem is likely to be in one of these files. The file
<code>lib_print.c</code> uses sleep(2) and also falls in this
category.</p>
<p>Almost all of the real work is done in the files</p>
<blockquote>
<code>hardscroll.c hashmap.c lib_addch.c lib_doupdate.c
lib_getch.c lib_mouse.c lib_mvcur.c lib_refresh.c lib_setup.c
lib_vidattr.c</code>
</blockquote>
<p>Most of the algorithmic complexity in the library lives in
these files. If there is a real bug in <strong>ncurses</strong>
itself, it is probably here. We will tour some of these files in
detail below (see <a href="#engine">The Engine Room</a>).</p>
<p>Finally, there is a group of files that is actually most of
the terminfo compiler. The reason this code lives in the
<strong>ncurses</strong> library is to support fallback to
/etc/termcap. These files include</p>
<blockquote>
<code>alloc_entry.c captoinfo.c comp_captab.c comp_error.c
comp_hash.c comp_parse.c comp_scan.c parse_entry.c
read_termcap.c write_entry.c</code>
</blockquote>
<p>We will discuss these in the compiler tour.</p>
<h2><a name="engine" id="engine">The Engine Room</a></h2>
<h3><a name="input" id="input">Keyboard Input</a></h3>
<p>All <code>ncurses</code> input funnels through the function
<code>wgetch()</code>, defined in <code>lib_getch.c</code>. This
function is tricky; it has to poll for keyboard and mouse events
and do a running match of incoming input against the set of
defined special keys.</p>
<p>The central data structure in this module is a FIFO queue,
used to match multiple-character input sequences against
special-key capabilities; also to implement pushback via
<code>ungetch()</code>.</p>
<p>The <code>wgetch()</code> code distinguishes between function
key sequences and the same sequences typed manually by doing a
timed wait after each input character that could lead a function
key sequence. If the entire sequence takes less than 1 second, it
is assumed to have been generated by a function key press.</p>
<p>Hackers bruised by previous encounters with variant
<code>select(2)</code> calls may find the code in
<code>lib_twait.c</code> interesting. It deals with the problem
that some BSD selects do not return a reliable time-left value.
The function <code>timed_wait()</code> effectively simulates a
System V select.</p>
<h3><a name="mouse" id="mouse">Mouse Events</a></h3>
<p>If the mouse interface is active, <code>wgetch()</code> polls
for mouse events each call, before it goes to the keyboard for
input. It is up to <code>lib_mouse.c</code> how the polling is
accomplished; it may vary for different devices.</p>
<p>Under xterm, however, mouse event notifications come in via
the keyboard input stream. They are recognized by having the
<strong>kmous</strong> capability as a prefix. This is kind of
klugey, but trying to wire in recognition of a mouse key prefix
without going through the function-key machinery would be just
too painful, and this turns out to imply having the prefix
somewhere in the function-key capabilities at terminal-type
initialization.</p>
<p>This kluge only works because <strong>kmous</strong> is not
actually used by any historic terminal type or curses
implementation we know of. Best guess is it is a relic of some
forgotten experiment in-house at Bell Labs that did not leave any
traces in the publicly-distributed System V terminfo files. If
System V or XPG4 ever gets serious about using it again, this
kluge may have to change.</p>
<p>Here are some more details about mouse event handling:</p>
<p>The <code>lib_mouse()</code> code is logically split into a
lower level that accepts event reports in a device-dependent
format and an upper level that parses mouse gestures and filters
events. The mediating data structure is a circular queue of event
structures.</p>
<p>Functionally, the lower level's job is to pick up primitive
events and put them on the circular queue. This can happen in one
of two ways: either (a) <code>_nc_mouse_event()</code> detects a
series of incoming mouse reports and queues them, or (b) code in
<code>lib_getch.c</code> detects the <strong>kmous</strong>
prefix in the keyboard input stream and calls _nc_mouse_inline to
queue up a series of adjacent mouse reports.</p>
<p>In either case, <code>_nc_mouse_parse()</code> should be
called after the series is accepted to parse the digested mouse
reports (low-level events) into a gesture (a high-level or
composite event).</p>
<h3><a name="output" id="output">Output and Screen Updating</a></h3>
<p>With the single exception of character echoes during a
<code>wgetnstr()</code> call (which simulates cooked-mode line
editing in an ncurses window), the library normally does all its
output at refresh time.</p>
<p>The main job is to go from the current state of the screen (as
represented in the <code>curscr</code> window structure) to the
desired new state (as represented in the <code>newscr</code>
window structure), while doing as little I/O as possible.</p>
<p>The brains of this operation are the modules
<code>hashmap.c</code>, <code>hardscroll.c</code> and
<code>lib_doupdate.c</code>; the latter two use
<code>lib_mvcur.c</code>. Essentially, what happens looks like
this:</p>
<ul>
<li>
<p>The <code>hashmap.c</code> module tries to detect vertical
motion changes between the real and virtual screens. This
information is represented by the oldindex members in the
newscr structure. These are modified by vertical-motion and
clear operations, and both are re-initialized after each
update. To this change-journalling information, the hashmap
code adds deductions made using a modified Heckel algorithm
on hash values generated from the line contents.</p>
</li>
<li>
<p>The <code>hardscroll.c</code> module computes an optimum
set of scroll, insertion, and deletion operations to make the
indices match. It calls <code>_nc_mvcur_scrolln()</code> in
<code>lib_mvcur.c</code> to do those motions.</p>
</li>
<li>
<p>Then <code>lib_doupdate.c</code> goes to work. Its job is
to do line-by-line transformations of <code>curscr</code>
lines to <code>newscr</code> lines. Its main tool is the
routine <code>mvcur()</code> in <code>lib_mvcur.c</code>.
This routine does cursor-movement optimization, attempting to
get from given screen location A to given location B in the
fewest output characters possible.</p>
</li>
</ul>
<p>If you want to work on screen optimizations, you should use
the fact that (in the trace-enabled version of the library)
enabling the <code>TRACE_TIMES</code> trace level causes a report
to be emitted after each screen update giving the elapsed time
and a count of characters emitted during the update. You can use
this to tell when an update optimization improves efficiency.</p>
<p>In the trace-enabled version of the library, it is also
possible to disable and re-enable various optimizations at
runtime by tweaking the variable
<code>_nc_optimize_enable</code>. See the file
<code>include/curses.h.in</code> for mask values, near the
end.</p>
<h1><a name="fmnote" id="fmnote">The Forms and Menu Libraries</a></h1>
<p>The forms and menu libraries should work reliably in any
environment you can port ncurses to. The only portability issue
anywhere in them is what flavor of regular expressions the
built-in form field type TYPE_REGEXP will recognize.</p>
<p>The configuration code prefers the POSIX regex facility,
modeled on System V's, but will settle for BSD regexps if the
former is not available.</p>
<p>Historical note: the panels code was written primarily to
assist in porting u386mon 2.0 (comp.sources.misc v14i001-4) to
systems lacking panels support; u386mon 2.10 and beyond use it.
This version has been slightly cleaned up for
<code>ncurses</code>.</p>
<h1><a name="tic" id="tic">A Tour of the Terminfo Compiler</a></h1>
<p>The <strong>ncurses</strong> implementation of
<strong>tic</strong> is rather complex internally; it has to do a
trying combination of missions. This starts with the fact that,
in addition to its normal duty of compiling terminfo sources into
loadable terminfo binaries, it has to be able to handle termcap
syntax and compile that too into terminfo entries.</p>
<p>The implementation therefore starts with a table-driven,
dual-mode lexical analyzer (in <code>comp_scan.c</code>). The
lexer chooses its mode (termcap or terminfo) based on the first
&ldquo;,&rdquo; or &ldquo;:&rdquo; it finds in each entry. The
lexer does all the work of recognizing capability names and
values; the grammar above it is trivial, just "parse entries till
you run out of file".</p>
<h2><a name="nonuse" id="nonuse">Translation of
Non-<strong>use</strong> Capabilities</a></h2>
<p>Translation of most things besides <strong>use</strong>
capabilities is pretty straightforward. The lexical analyzer's
tokenizer hands each capability name to a hash function, which
drives a table lookup. The table entry yields an index which is
used to look up the token type in another table, and controls
interpretation of the value.</p>
<p>One possibly interesting aspect of the implementation is the
way the compiler tables are initialized. All the tables are
generated by various awk/sed/sh scripts from a master table
<code>include/Caps</code>; these scripts actually write C
initializers which are linked to the compiler. Furthermore, the
hash table is generated in the same way, so it doesn't have to be
generated at compiler startup time (another benefit of this
organization is that the hash table can be in shareable text
space).</p>
<p>Thus, adding a new capability is usually pretty trivial, just
a matter of adding one line to the <code>include/Caps</code>
file. We will have more to say about this in the section on
<a href="#translation">Source-Form Translation</a>.</p>
<h2><a name="uses" id="uses">Use Capability Resolution</a></h2>
<p>The background problem that makes <strong>tic</strong> tricky
is not the capability translation itself, it is the resolution of
<strong>use</strong> capabilities. Older versions would not
handle forward <strong>use</strong> references for this reason
(that is, a using terminal always had to follow its use target in
the source file). By doing this, they got away with a simple
implementation tactic; compile everything as it blows by, then
resolve uses from compiled entries.</p>
<p>This will not do for <strong>ncurses</strong>. The problem is
that that the whole compilation process has to be embeddable in
the <strong>ncurses</strong> library so that it can be called by
the startup code to translate termcap entries on the fly. The
embedded version cannot go promiscuously writing everything it
translates out to disk &mdash; for one thing, it will typically
be running with non-root permissions.</p>
<p>So our <strong>tic</strong> is designed to parse an entire
terminfo file into a doubly-linked circular list of entry
structures in-core, and then do <strong>use</strong> resolution
in-memory before writing everything out. This design has other
advantages: it makes forward and back use-references equally easy
(so we get the latter for free), and it makes checking for name
collisions before they are written out easy to do.</p>
<p>And this is exactly how the embedded version works. But the
stand-alone user-accessible version of <strong>tic</strong>
partly reverts to the historical strategy; it writes to disk (not
keeping in core) any entry with no <strong>use</strong>
references.</p>
<p>This is strictly a core-economy kluge, implemented because the
terminfo master file is large enough that some core-poor systems
swap like crazy when you compile it all in memory...there have
been reports of this process taking <strong>three hours</strong>,
rather than the twenty seconds or less typical on the author's
development box.</p>
<p>So. The executable <strong>tic</strong> passes the
entry-parser a hook that <em>immediately</em> writes out the
referenced entry if it has no use capabilities. The compiler main
loop refrains from adding the entry to the in-core list when this
hook fires. If some other entry later needs to reference an entry
that got written immediately, that is OK; the resolution code
will fetch it off disk when it cannot find it in core.</p>
<p>Name collisions will still be detected, just not as cleanly.
The <code>write_entry()</code> code complains before overwriting
an entry that postdates the time of <strong>tic</strong>'s first
call to <code>write_entry()</code>, Thus it will complain about
overwriting entries newly made during the <strong>tic</strong>
run, but not about overwriting ones that predate it.</p>
<h2><a name="translation" id="translation">Source-Form
Translation</a></h2>
<p>Another use of <strong>tic</strong> is to do source
translation between various termcap and terminfo formats. There
are more variants out there than you might think; the ones we
know about are described in the <strong>captoinfo(1)</strong>
manual page.</p>
<p>The translation output code (<code>dump_entry()</code> in
<code>ncurses/dump_entry.c</code>) is shared with the
<strong>infocmp(1)</strong> utility. It takes the same internal
representation used to generate the binary form and dumps it to
standard output in a specified format.</p>
<p>The <code>include/Caps</code> file has a header comment
describing ways you can specify source translations for
nonstandard capabilities just by altering the master table. It is
possible to set up capability aliasing or tell the compiler to
plain ignore a given capability without writing any C code at
all.</p>
<p>For circumstances where you need to do algorithmic
translation, there are functions in <code>parse_entry.c</code>
called after the parse of each entry that are specifically
intended to encapsulate such translations. This, for example, is
where the AIX <strong>box1</strong> capability get translated to
an <strong>acsc</strong> string.</p>
<h1><a name="utils" id="utils">Other Utilities</a></h1>
<p>The <strong>infocmp</strong> utility is just a wrapper around
the same entry-dumping code used by <strong>tic</strong> for
source translation. Perhaps the one interesting aspect of the
code is the use of a predicate function passed in to
<code>dump_entry()</code> to control which capabilities are
dumped. This is necessary in order to handle both the ordinary
De-compilation case and entry difference reporting.</p>
<p>The <strong>tput</strong> and <strong>clear</strong> utilities
just do an entry load followed by a <code>tputs()</code> of a
selected capability.</p>
<h1><a name="style" id="style">Style Tips for Developers</a></h1>
<p>See the TO-DO file in the top-level directory of the source
distribution for additions that would be particularly useful.</p>
<p>The prefix <code>_nc_</code> should be used on library public
functions that are not part of the curses API in order to prevent
pollution of the application namespace. If you have to add to or
modify the function prototypes in curses.h.in, read
ncurses/MKlib_gen.sh first so you can avoid breaking XSI
conformance. Please join the ncurses mailing list. See the
INSTALL file in the top level of the distribution for details on
the list.</p>
<p>Look for the string <code>FIXME</code> in source files to tag
minor bugs and potential problems that could use fixing.</p>
<p>Do not try to auto-detect OS features in the main body of the
C code. That is the job of the configuration system.</p>
<p>To hold down complexity, do make your code data-driven.
Especially, if you can drive logic from a table filtered out of
<code>include/Caps</code>, do it. If you find you need to augment
the data in that file in order to generate the proper table, that
is still preferable to ad-hoc code &mdash; that is why the fifth
field (flags) is there.</p>
<p>Have fun!</p>
<h1><a name="port" id="port">Porting Hints</a></h1>
<p>The following notes are intended to be a first step towards
DOS and Macintosh ports of the ncurses libraries.</p>
<p>The following library modules are &ldquo;pure curses&rdquo;;
they operate only on the curses internal structures, do all
output through other curses calls (not including
<code>tputs()</code> and <code>putp()</code>) and do not call any
other UNIX routines such as signal(2) or the stdio library. Thus,
they should not need to be modified for single-terminal
ports.</p>
<blockquote>
<code>lib_addch.c lib_addstr.c lib_bkgd.c lib_box.c lib_clear.c
lib_clrbot.c lib_clreol.c lib_delch.c lib_delwin.c lib_erase.c
lib_inchstr.c lib_insch.c lib_insdel.c lib_insstr.c
lib_keyname.c lib_move.c lib_mvwin.c lib_newwin.c lib_overlay.c
lib_pad.c lib_printw.c lib_refresh.c lib_scanw.c lib_scroll.c
lib_scrreg.c lib_set_term.c lib_touch.c lib_tparm.c lib_tputs.c
lib_unctrl.c lib_window.c panel.c</code>
</blockquote>
<p>This module is pure curses, but calls outstr():</p>
<blockquote>
<code>lib_getstr.c</code>
</blockquote>
<p>These modules are pure curses, except that they use
<code>tputs()</code> and <code>putp()</code>:</p>
<blockquote>
<code>lib_beep.c lib_color.c lib_endwin.c lib_options.c
lib_slk.c lib_vidattr.c</code>
</blockquote>
<p>This modules assist in POSIX emulation on non-POSIX
systems:</p>
<dl>
<dt>sigaction.c</dt>
<dd>signal calls</dd>
</dl>
<p>The following source files will not be needed for a
single-terminal-type port.</p>
<blockquote>
<code>alloc_entry.c captoinfo.c clear.c comp_captab.c
comp_error.c comp_hash.c comp_main.c comp_parse.c comp_scan.c
dump_entry.c infocmp.c parse_entry.c read_entry.c tput.c
write_entry.c</code>
</blockquote>
<p>The following modules will use
open()/read()/write()/close()/lseek() on files, but no other OS
calls.</p>
<dl>
<dt>lib_screen.c</dt>
<dd>used to read/write screen dumps</dd>
<dt>lib_trace.c</dt>
<dd>used to write trace data to the logfile</dd>
</dl>
<p>Modules that would have to be modified for a port start
here:</p>
<p>The following modules are &ldquo;pure curses&rdquo; but
contain assumptions inappropriate for a memory-mapped port.</p>
<dl>
<dt>lib_longname.c</dt>
<dd>assumes there may be multiple terminals</dd>
<dt>lib_acs.c</dt>
<dd>assumes acs_map as a double indirection</dd>
<dt>lib_mvcur.c</dt>
<dd>assumes cursor moves have variable cost</dd>
<dt>lib_termcap.c</dt>
<dd>assumes there may be multiple terminals</dd>
<dt>lib_ti.c</dt>
<dd>assumes there may be multiple terminals</dd>
</dl>
<p>The following modules use UNIX-specific calls:</p>
<dl>
<dt>lib_doupdate.c</dt>
<dd>input checking</dd>
<dt>lib_getch.c</dt>
<dd>read()</dd>
<dt>lib_initscr.c</dt>
<dd>getenv()</dd>
<dt>lib_newterm.c</dt>
<dt>lib_baudrate.c</dt>
<dt>lib_kernel.c</dt>
<dd>various tty-manipulation and system calls</dd>
<dt>lib_raw.c</dt>
<dd>various tty-manipulation calls</dd>
<dt>lib_setup.c</dt>
<dd>various tty-manipulation calls</dd>
<dt>lib_restart.c</dt>
<dd>various tty-manipulation calls</dd>
<dt>lib_tstp.c</dt>
<dd>signal-manipulation calls</dd>
<dt>lib_twait.c</dt>
<dd>gettimeofday(), select().</dd>
</dl>
<hr>
<address>
Eric S. Raymond &lt;esr@snark.thyrsus.com&gt;
</address>
(Note: This is <em>not</em> the <a href="#bugtrack">bug
address</a>!)
</body>
</html>