Since the termcap database is an ASCII representation of the data, people may believe that there is no need for a compiler. Syntax checks and unification however are a property of compilers. The termcap program checks for correct data types, outputs all entries in a unified order and recodes all strings with the same escape notation. This is needed in to compare two entries and it makes life easier.
When termcap compiles entries, it checks the entries for noncompliances. If any problems were found, it inserts comments in the form: "# NOTICE(" name ")" " some reason or "# BAD(" name ")" " some reason before the related termcap entry. These comments should be carefully checked. Frequent problems are illegally escaped colons \e: instead of \e072 that in most cases have been created by defective terminfo software.
-help Prints a short summary of the termcap (1) options and exists.
-version Prints the termcap (1) version number string and exists.
-dooctal When printing terminal capabilities, prefer octal output like \e003 before escape sequences like ^C .
-dumplist Dump the complete internal list of terminal capabilities.
-e Import the termcap description from the TERMCAP environment if present. By default, termcap (1) ignores the TERMCAP environment in case that it does not hold a pathname, but a preparsed termcap entry.
if= filename Read the file filename with many termcap entries and produce compiled output in standard order on stdout . This option is used in order to recompile a whole set of termcap entries. To make the recompiling as smooth as possible, -inorder is assumed by default. Use -noinorder if you like to add further processing by hand.
-inorder Output all capabilities in the standard order. First the boolean capabilities, then the numeric capabilities and at last the string capabilities. Without -inorder , the obsolete capabilities are printed last. The option -inorder is assumed by default if the if= name option has been specified.
-gnugoto Allow the GNU tgoto() format extensions %C and %m.
-nodisabled Do not output disabled termcap entries. A comment with a warning for this entry is still printed in the output.
-noinorder Output all capabilities in the standard order, but print obsoleted capabilities, unknown capabilities and disabled capabilities in a separate block before any tc= entry.
-nounknown Do not output unknown termcap entries. A comment with a warning for this entry is still printed in the output.
-nowarn Do not warn about incorrect entries that could be fixed by assuming that the problem is a result from not correctly converting escape sequences in terminfo syntax to termcap syntax. For this purpose, currently null characters ( \e0 ) are converted to \e200 . In addition, the escape sequences \ea , \ee , \e: , \e, , \es and \el are mapped to BEL, ESCAPE, ':', ',', ' ' and LINEFEED.
-oneline Output all termcap entries in a single line per entry. This helps to scan the whole termcap data base for terminals with specific properties.
-s Output commands to set and export TERM and TERMCAP . This can be done with " eval `tset -s ...` to bring the terminal information into the environment of shells that do not support termcap internally as e.g. sh (recent Bourne Shell versions) and bsh already do. Doing so makes termcap based programs start up faster. If the SHELL environment variable ends with csh , C shell commands are created, otherwise Bourne shell commands are created.
-tc Follow tc= entries and generate cumulative output. Amongst others, this allows to retrieve the cumulative size of the related termcap entry. Termcap normally does not interpret tc= entries in order to allow the entries to be reused the same way they are in the termcap file. When the -tc option is used together with if= filename and termcap (1) is in compile mode, the output will increase noticeably in size as all termcap entries for all terminals will be fully expanded. This also removes all hints to similar terminals in the output.
-v Increment verbose level. By default, termcap only prints a recoded termcap entry. With increasing verbose level, more information on the termcap entry is printed. With a higher verbose level, termcap prints the size of the termcap entry.
HOME To find the private files like ~/.termcap
TERM The type of terminal for which the output should be formatted.
TERMCAP This environment variable holds either a precompiled termcap entry or the pathname to be used to find a termcap database file. If it holds a precompiled entry that does not match the TERM environment, the termcap database is parsed as if the TERMCAP environment is not set.
TERMPATH If TERMCAP is empty, then the TERMPATH environment is scanned for pathnames of files that contain a termcap database. It holds a list of filenames separated by colons or spaces (i.e., ":" or " "). If the TERMPATH symbol is not set, the files $HOME/.termcap and /etc/termcap are scanned in that order.
/etc/termcap The system's terminal capabilities database.
~/.termcap The private terminal capabilities database. This file is checked first, in case that the TERMPATH environment is not defined.
The following messages are created by termcap (1):
"# BAD(name). Skipping blank entry: 'ent'" A colon (":") is followed by white space. White space is only permitted in a termcap entry if it appears at the beginning of a line. This error type is usually caused by a bad manual edit. Carefully check every such problem in the source and manually correct it.
"# BAD(name). Illegal entry (3rd char 'c' for ':ccc'): 'ent'" A valid termcap entry is made of two characters followed by a colon (":") in case of a boolean entry, a hash mark ("#") in case of an integer entry, an equal sign ("=") for a string entry, or the at sign ("@") for disabled entries. This error type is usually caused by a bad conversion from a terminfo entry that is caused by illegal quoting. A bad quoting may cause the termcap parser to be out of sync. A quoted colon ("\e:") is permitted in terminfo but not in termcap ( termcap only permits "\e072" to include a colon in a string). This error may also be caused by a bad manual edit, e.g. a backslash ("\e") at the end of a termcap entry that causes the next terminal name to be interpreted as capability, a colon (":") instead of a bar ("|") in the list of alternative terminal names or superfluous colon (":") in the list of capabilities. Carefully check every such problem in the source and manually correct it.
"# BAD(name). Type mismatch 'ent' in 'full-ent' ... The type for the current entry does not match the type from the termcap documentation. This error type is usually caused by a bad manual edit, e.g. from typing ":", "=" or "#" instead of the correct character "=" or "#" that is expected as the third character for the entry. Carefully check every such problem in the source and manually correct it.
"# BAD(name). Bad format '%c' in 'ent=value'" An invalid format character was found in a parameterized string (a string that is passed to tgoto() ). This error type is usually caused by badly converted terminfo entry. Carefully check every such problem in the source and manually correct it.
"# BAD(name). Buffer overflow in 'ent=value'" A string that is passed to tgoto() is longer than 80 characters. tgoto() is not granted to work correctly in case that the parameterized string is longer than 80 characters. This error type is usually caused by badly converted terminfo entry. Carefully check every such problem in the source and manually correct it.
"# NOTICE(name). Short entry (':c:') removed" All termcap capability names must have two characters. A single char entry was found and removed.
"# NOTICE(name). Disabled entry: '.ent'" An entry that starts with a dot (".") was found. This is usually caused by an intentional manual edit.
"# NOTICE(name). Unknown entry ('ent'): 'full-ent'" An entry was found that is not part of the official termcap documentation. It is unlikely that such an entry is supported by any program, so these entries may be removed from the the termcap database. This may be done automatically by specifying the -nounknown option.
"# NOTICE(name). Canceled entry 'ent@' followed by 'full-ent'..." A cancelled entry in the form "xx@" was followed by "xx=string" or "xx#number" in the same termcap entry. The cancel feature in termcap entries is intended to cancel entries from capabilities imported via the tc= name feature. Carefully check every such problem in the source whether it was introduced intentionally (e.g. in order to temporarily disable a feature) or whether it was introduced by accident.
"# NOTICE(name). NULL char (fixed) in entry ('ent') at..." A "\e0" string was found in a termcap entry. This is not permitted by the termcap documentation. This error type is usually caused by a bad conversion from a terminfo entry. This typically happens if bad software converted a "\eE" string to "\e" or in case that the string "\e0" that is valid in a terminfo entry was not converted to the correct termcap equivalent: "\e200". termcap (1) assumes the latter and automatically converts "\e0" to "\e200". A NULL character may be used as a delay in an output sequence, but it is not expected in function key input sequence. Carefully check every such problem in the source and manually correct it.
"# NOTICE(name). Nonoctal char 'c' in entry ('ent') at position..." A character outside the range '0'..'7' was found in a '\ennn' octal sequence. Carefully check every such problem in the source and manually correct it.
"# NOTICE(name). Badly quoted char '\ec' (fixed) in ('ent') at ..." A character quoted with "\e" was found but the character is not in the list of quotes documented by termcap . As the character was from the list "\ea", "\ee", "\e:", "\e,", "\es", "\el" it is assumed that the problem was caused by a bad conversion from a terminfo entry. Carefully check every such problem in the source and manually correct it if needed.
"# NOTICE(name). Badly quoted char '\ec' in ('ent') at ..." A character quoted with "\e" was found but the character is not in the list of quotes documented by termcap . As the character was neither from the list "\ea", "\ee", "\e:", "\e,", "\es", "\el" it is assumed that the problem was caused by a bad manual edit. Carefully check every such problem in the source and manually correct it.
Termcap databases that have been created from a terminfo database often are full of noncompliances that prevent successful parsing by termcap-compliant software. The fact that terminfo permits "\e:" is the cause for the majority of all problems but fortunately can be automatically resolved.
There is a non-compliance between older and newer termcap entries. The capability named "ma" in former times (around 1980) was an arrow key map for the rogue (1) "Dungeons of Doom game. In newer times, the "ma" entry is used for the maximum number of attributes a terminal can handle and thus is of type integer.
As the functions tgetstr () and tgetnum () both skip entries with unappropriate type, this is not a problem. For this reason, we do not warn of a possible string/number type conflict with the "ma" capability. It is even possible to have a ma= maplist and ma# num-attr entry at the same time for the same terminal entry.