cscope toolman page
NAME
cscope - interactively examine a C program
SYNOPSIS
cscope [ options ] [ C source files ]
DESCRIPTION
Cscope is an interactive screen-oriented tool that helps
you:
Learn how a C program works without endless flip-
ping through a thick listing.
Locate the section of code to change to fix a bug
without having to learn the entire program.
Examine the effect of a proposed change such as
adding a value to an enum variable.
Verify that a change has been made in all source
files such as adding an argument to an existing
function.
Rename a global variable in all source files.
Change a constant to a preprocessor symbol in
selected lines of files.
It is designed to answer questions like:
Where is this symbol used?
Where is it defined?
Where did this variable get its value?
What is this global symbol's definition?
Where is this function in the source files?
What functions call this function?
What functions are called by this function?
Where does the message "out of space" come from?
Where is this source file in the directory structure?
What files include this header file?
Cscope answers these questions from a symbol database that
it builds the first time it is used on the source files.
On a subsequent call, cscope rebuilds the database only if
a source file has changed or the list of source files is
different. When the database is rebuilt the data for the
unchanged files is copied from the old database, which
makes rebuilding much faster than the initial build.
After the database is ready cscope will display these
input fields:
Find this C symbol:
Find this definition:
Find functions called by this function:
Find functions calling this function:
Find assignments to:
Change this grep pattern:
Find this egrep pattern:
Find this file:
Find files #including this file:
Press the TAB key repeatedly to move to the desired input
field, type the pattern to search for, and then press the
RETURN key. For the first 5 and last 2 input fields, the
pattern can be a regcmp(3X) regular expression. If the
search is successful, you can use these command charac-
ters:
1-9 Edit the file containing the displayed line.
space Display next lines.
+ Display next lines.
- Display previous lines.
^E Edit all lines.
> Write all lines to a file.
>> Append all lines to a file.
< Read lines from a file.
^ Filter all lines through a shell command.
| Pipe all lines to a shell command.
At any time you can use these command characters:
^P Move to the previous input field.
^A Search again with the last pattern typed.
^B Recall previous input field and search pattern.
^F Recall next input field and search pattern.
^C Toggle ignore/use letter case when searching.
^R Rebuild the symbol database.
! Start an interactive shell (type ^D to return to
cscope).
^L Redraw the screen.
? Display this list of commands.
^D Exit cscope.
Note: If the first character of the pattern you want to
search for matches one of the above commands, type a \
character first.
Changing Text
After typing the grep(1) pattern to be searched for, you
will be prompted for the replacement text, then the lines
containing the pattern will be displayed. You select the
lines to be changed with these command characters:
1-9 Mark or unmark the line to be changed.
* Mark or unmark all displayed lines to be changed.
space Display next lines.
+ Display next lines.
- Display previous lines.
a Mark or unmark all lines to be changed.
^D Change the marked lines and exit.
RETURN Exit without changing the marked lines.
! Start an interactive shell (type ^D to return to
cscope).
^L Redraw the screen.
? Display this list of commands.
Special Keys
If your keyboard has arrow keys that work in vi(1) you can
use them to move around the input fields. The up-arrow
key is useful to move to the previous input field instead
of using the TAB key repeatedly. If you have the CLEAR,
NEXT, or PREV keys they will act as the ^L, +, and - com-
mands respectively.
Mouse
If you have an AT&T 5620/630/730 terminal and use the
emacsterm(1), myx(1), or viterm(1) terminal programs; or
use the X Window System and the etermx(1) terminal pro-
gram; or SunView and the eterm(1) terminal program; you
can use your mouse with cscope. If you use the myx(1)
terminal program, set the MOUSE environment variable to
myx (see ENVIRONMENT VARIABLES).
Point with the mouse and click button 1 to move to the
desired input field, type the pattern to search for, and
then press the RETURN key. If the search is successful,
you can edit the file containing a displayed line by
pointing with the mouse and clicking button 1. When
changing text, point with the mouse and click button 1 to
mark or unmark the line to be changed. Press mouse button
2 to get a menu of commands.
Options
When invoking cscope you can use these options:
-a When using the "Find this egrep pattern:" search,
ask for a regular expression describing the sub-
set of files to search. If no expression is sup-
plied, all files known to cscope are searched.
This option is particularly useful for very large
software systems where thousands of source files
are used. In such systems the set of files to be
searched is a small subset of the total. This
option is also available as the 'a' command in
the command-line interface. The -l or -L options
turn off this option.
-b Build the database only.
-c Use only ASCII characters in the database file,
that is, do not compress the data.
-C Ignore letter case when searching.
-d Do not update the database.
-e Suppress the ^E command prompt between files.
-f file Use file as the database file name instead of the
default (cscope.out).
-F file Read symbol reference lines from file, just like
the < command.
-i file Read any -I, -p, -q, and -T options and the list
of source files from file instead of the default
(cscope.files).
-I dir Look in dir for #include files. If you will be
editing these files, use the -s option instead so
the rebuild (^R) command will notice they have
changed.
-k Kernel Mode turns off the use of the default
include dir (usually /usr/include) when building
the database, since kernel source trees generally
do not use it.
-l Line-oriented interface (see Calling Cscope with
a Line-Oriented Interface).
-L Do a single search with line-oriented output when
used with the -n pattern option.
-o Display OGS book and subsystem names.
-p n Display the last n file path components instead
of the default (1). Use 0 to not display the
file name at all.
-P path Prepend path to relative file names in a pre-
built database so you do not have to change to
the directory where the database was built. This
option is only valid with the -d option. Do not
use this option if a viewpath was used when
building the database (see Viewpath subsection of
EXAMPLES section).
-q Build an inverted index for quick symbol seach-
ing. If you use this option with the -f option,
you must use -f on every call to cscope, includ-
ing building the database, because it changes the
names of the inverted index files.
NOTE: This is not available on MSDOS or MS Win-
dows versions.
-r Require pressing the RETURN key after typing the
selection digit(s) for a displayed line to edit
the file containing it, which allows as many
lines as possible to be displayed.
-R Recurse subdirectories during search for source
files.
-s dir Look in dir for additional source files. This
option is ignored if source files are given on
the command line.
-t hours No activity timeout in hours (minimum of 4
hours). Removing a database will not release the
disk space if a cscope process has the file open,
so a project can use this option in its script to
call cscope on a large database so unattended
cscope processes will exit overnight.
-T Use only the first eight characters to match
against C symbols. A regular expression contain-
ing special characters other than a period (.)
will not match any symbol if its minimum length
is greater than eight characters.
-u Unconditionally build the database (assume that
all files have changed).
-U Do not check file time stamps (assume that some
files have changed).
-x ext Use the extended menu with option ext. When ext
is set to "c", cscope will emulate ccalls(1) and
will display the following input fields instead
of the normal menu:
Find this type definition:
Find this preprocessor definition:
Find parameters of this function:
Find functions calling this function:
Find this global definition:
Find this local definition:
Find this structure definition:
Find this file:
Find this preprocessor include:
When ext is set to "s", system calls are ignored in the
output. To combine extension options, use concatenation
e.g. "cs". See cscalls(1).
-n pattern
Go to input field n (counting from 0) and find pat-
tern. The input field number is the line number,
counting from 0, of the search type displayed on
the screen. That is, given the following table
+----------------------------------------------------+
|LINE SEARCH TYPE |
+----------------------------------------------------+
| 0 Find this C symbol: |
| 1 Find this definition: |
| 2 Find functions called by this function: |
| 3 Find functions calling this function: |
| 4 Find assignments to: |
| 5 Change this grep pattern: |
| 6 Find this egrep pattern: |
| 7 Find this file: |
| 8 Find files #including this file: |
+----------------------------------------------------+
| (Used by samuel and cdb only) |
| 9 Find all function/class definitions (samuel) |
| : Find all function/class definitions (cdb) |
+----------------------------------------------------+
| (Used by ccalls with the extended menu) |
| 0 Find this type definition: |
| 1 Find this preprocessor definition: |
| 2 Find parameters of this function: |
| 3 Find functions calling this function: |
| 4 Find this global definition: |
| 5 Find this local definition: |
| 6 Find this structure definition: |
| 7 Find this file: |
| 8 Find this preprocessor include: |
| 9 Find definition of this function: |
| : Find all function/class definitions: |
+----------------------------------------------------+
If you wanted to "Find assignments to", then you would use
"-4 pattern"
The -c, -k, -I, -p, -q, and -T options can also be in the
cscope.files file. See below for common uses of most of
these options.
EXAMPLES
Calling Cscope from the Shell
To use cscope on all the C, lex, yacc, breakpoint listing,
Ingres(TM), and SDL files in the current directory just
type
cscope
You can use cscope on particular files if you have more
than one program in the same directory. The above is
equivalent to
cscope *.[chlyCGHL] *.bp *.q[ch] *.pr *.sd
If you want to use cscope on all the source files in the
current and a few other directories, use the -s option to
give the other directory names, e.g.
cscope -s ../common,../../common
If you want to use cscope on all the source files in the
current and lower level directories, first create a list
of the files with
find . -name '*.[chlyCGHL]' -print >cscope.files
and then type
cscope
Cscope will look up the viewpath for the cscope.files
file, so you can put it under change management to prevent
accidental destruction.
If a file #include's files that are in directories other
than the current directory or the standard directory
(/usr/include), use the -I option just like when using the
cc(1) command
cscope -I../hdr -I../../hdr
If you will be editing these files, use the -s option
instead so the rebuild (^R) command will notice they have
changed.
If files may have the same base name, you will want to use
the -p option to display the last n file path components
instead of the default (1).
If you use a compiler that does not support flexnames,
that is, symbol name differences past eight characters are
ignored, you will want to use the -T option to have cscope
do the same.
You can put -c, -k, -I, -p, -q, and -T options in the file
containing the list of source files (see above). This
shortens the cscope command line and makes it easier to
call cscope from an editor.
If you use cscope on thousands of source files, use the -b
option to only build the database so you can do run it in
the background with a batch(1) or at(1) job.
echo 'cscope -b' | batch
Use the -d option on subsequent calls to cscope
cscope -d
to avoid the time spent checking if the database is out of
date, which can be significant.
If the source files change infrequently, you can get quick
symbol searching by building the database with the -q
option. For a very large database you will be able to
find a symbol in a few seconds vs. several minutes if
built without -q, at the expense of about twice as much
database disk space and build CPU time. Updating a -q
database takes about as half as long as building it. It
contains binary numbers, so it is portable only between
machines with the same byte order and scalar-type sizes
(such as int, long, short, char, and pointer). Note that
3B, Amdahl/IBM, and MC680x0-based machines like Suns all
have the same byte order.
The -q option makes it practical to have databases for
entire projects. If you try to build a project database
and get a "File too large" message, you need to get your
login's ulimit (file size limit) raised by your system
administrator. If you get the "No space left on device"
message you will have to use a file system with more
space. You can change the temporary file system by set-
ting the TMPDIR environment variable. If you have enough
space to build the database but not to rebuild it after
some files have changed, try removing the inverted index
cscope.in.out and cscope.po.out files. If you still don't
have enough space to rebuild, you will have to remove the
cscope.out file.
If you want to use cscope separately on several programs
in the same directory structure, but keep the databases in
the same (top) directory, use the -f and -i options to
rename the cscope.out and cscope.files file, respectively,
e.g.
find dir1 -name '*.[chlyCGHL]' -print >dir1.files
find dir2 -name '*.[chlyCGHL]' -print >dir2.files
cscope -b -f dir1.db -i dir1.files
cscope -b -f dir2.db -i dir2.files
and call cscope with
cscope -d -f dir2.db
Note that options used only when building the database,
such as -i, are not needed with the -d option. If other
people are going to use these databases, you will should
provide a shell script to call cscope with the correct
options. If a viewpath (see Viewpath below) was not used
when building the database, then use the -P option to give
the path to relative file names so the script does not
have to change to the directory where the database was
built, which can confuse users.
Calling Cscope with Compressed Databases
To call cscope with a database which has been compressed
using gzip, bzip2 or i<I>p just type
cscope -d
The compressed database will be uncompressed into a cache
directory before being opened by cscope. If the database
already exists in the cache, then no uncompressing will be
performed and the existing database in the cache will be
used. The uncompressed database will be placed in a sub-
directory in the cache, which derives its name from the
last two directories of the database filename. This means
different databases will be stored separately in the
cache.
If a compressed database exists in the VPATH and in the
current directory, then following normal VPATH conven-
tions, the database in the current directory will be
uncompressed into the cache. If an uncompressed database
exists before a compressed database in the VPATH, then the
uncompressed database will be accessed as normal. How-
ever, if the cscope.out file has been removed in the cur-
rent directory but the inverted indexes have been not been
deleted, then be warned that cscope may try to open the
database in the cache and the inverted indexes in the cur-
rent directory. If this is not the desired behaviour,
then please remove the unwanted inverted indexes along
with the uncompressed database.
When building the database, if a compressed database
exists in the VPATH, then cscope will uncompress it in the
current directory before attempting to build the database.
This allows the database to be built and compressed and
when the database is updated, cscope will automatically
uncompress the database ready for update.
The compression of databases is not implemented by cscope.
Control of the cache is suggested to be performed by the
system administrator. The file /etc/cscopemaxcache can be
created, containing the maximum number of files in the
cache. If the cache size has been exceeded, then cscope
will attempt to remove files from the cache before uncom-
pressing a new database. If /etc/cscopemaxcache does not
exist, then no limit on the number of files in the cache
will be applied, except for the following check. If
TMPDIR is more than 50% full, then cscope will attempt to
remove 1 sub-directory before uncompressing a new
database.
Calling Cscope from an Editor
To call cscope from within vi(1) type :!cscope followed by
any options or file arguments. From emacs(1) type
M-!cscope followed by any file arguments. There is an
emacs(1) macro for the two most common calls to cscope.
Type M-X to call cscope with no file arguments, and type
^UM-X to call cscope with the -d option to not update the
database. If nothing happens when you type M-X, add this
line to your $HOME/.emacs_init and/or $HOME/.ve_init file:
^X^L~EMACS/macros/basic
Stacking Cscope and Editor Calls
When cscope puts you in the editor to display one symbol
reference and you see another symbol of interest, don't
leave the editor, just call cscope again from within the
editor. In this way you can stack cscope and editor calls
so you can go off on a tangent, and then back up to where
you were by exiting the tangential cscope and editor
calls.
For example, when learning a program you will probably
have cscope find the main function and display it with
your editor. You can visualize this as an inverse stack
of calls:
cscope: find and display main()
editor: display first lines of main()
You may then see that main calls an initialize function.
Rather than exiting the editor to go back to cscope, call
cscope again from within the editor to find the initialize
function and display it with another editor call. Your
stack of calls now looks like this:
cscope: find and display main()
editor: display first lines of main()
cscope: find and display initialize()
editor: display first lines of initialize()
When you are through looking at this function, you can
exit this second editor call, exit the second cscope call,
and you are back in the first editor call that is display-
ing the main function. You are now back to:
cscope: find and display main()
editor: display first lines of main()
and can continue reading the main function.
Boolean Searching
You can do the equivalent of Boolean or logical AND, OR,
and NOT operations on searches with the ^, >, >>, and <
commands. For example, if you wanted all references to
the symbol stderr in file invlib.c, search for the symbol
and then use the ^ command with shell command
grep invlib.c
If you wanted all the references not in this file, use the
grep(1) -v option. If you wanted all the lines where sym-
bols stderr or stdout are used, search for the first sym-
bol, write the references to a file with the > command,
search for the second symbol, append the references to the
same file with the >> command, read the file with the <
command, and get rid of duplicate lines with the ^ command
using the shell command
sort | uniq
Note that all <global> references will no longer be first
because the references will be sorted by file name.
Viewpath
As described above, cscope searches for source files in
the current directory by default. When the VPATH environ-
ment variable is set, cscope searches for source files in
directories that comprise your viewpath. A viewpath is an
ordered list of directories, each of which has the same
directory structure below it.
For example, suppose you are part of a software project.
There is an official set of source files in directories
below /fs1/ofc and each user has some matching directories
under their home directory ($HOME). If you make changes
to the software system, you may have copies of just those
files you are changing in $HOME/src/cmd/prog1. The offi-
cial versions of the entire program can be found in
/fs1/ofc/src/cmd/prog1.
Suppose you use cscope to browse through the three files
that comprise prog1, namely f1.c, f2.c, and f3.c. You
would set VPATH to $HOME and /fs1/ofc, and export it:
VPATH=$HOME:/fs1/ofc
export VPATH
You would then make your current directory
$HOME/src/cmd/prog1 and invoke cscope:
cd $HOME/src/cmd/prog1
cscope
Cscope will locate all files in the viewpath. In case
duplicates are found, cscope uses the file whose parent
directory appears earlier in VPATH. Thus if f2.c is in
your directory (and all three files are in the official
directory), cscope will use f2.c from your directory and
f1.c and f3.c from the official directory.
The first directory in VPATH must be a prefix (usually
$HOME) of the directory you will be working in. Each
colon-separated directory in VPATH must be a full path,
that is, it must start with a slash (/).
Calling Cscope with a Line-Oriented Interface
The -l option lets you use cscope where a screen-oriented
interface would not be useful, e.g. from another screen-
oriented program. Cscope will prompt with ">> " when it
is ready for an input line starting with the field number
(counting from 0) immediately followed by the search pat-
tern, e.g. "1main" finds the definition of the main func-
tion. Field number 9 followed by any pattern, e.g. "9a",
finds all function and C++ class definitions. If you just
want a single search, instead of the -l option use the -L
and -n pattern options, and you won't get the ">> "
prompt. For -l, cscope outputs the number of reference
lines
cscope: 2 lines
For each reference found, cscope outputs a line consisting
of the file name, function name, line number, and line
text, separated by spaces, e.g.
main.c main 161 main(argc, argv)
Note that as of this time, the file name, and the function
name must not have embedded spaces (such as what happens
on Microsoft boxes).
Also note that the editor is not called to display a sin-
gle reference, unlike the screen-oriented interface. You
can use the 'c' command to toggle ignoring/using letter
case when searching, the 'r' command to rebuild the
database, and the 'P' command to print the path to rela-
tive file names (either the -P option path or the full
current directory path). Cscope will quit when it detects
end-of-file, or when the first character of an input line
is ^D or 'q'.
This option allows you to write a program like samuel(1)
that uses cscope to look up symbol information. The pro-
gram can send search commands to cscope and read the
results. The xcscope macro package available for emacs(1)
uses this option to look up C symbols. This option may
also be useful for terminals not supported by terminfo(4)
or for sight-impaired users of printing terminals.
The following table lists the commands available in line-
oriented mode (-l). If a string is specified, there must
be no whitespace between the command letter and the
string. The string is shown here as '<str>' for readabil-
ity but the '<>' chars are not used.
+----------------------------------------------------------+
| CMD ACTION |
+----------------------------------------------------------+
|0<str> Find this C symbol: |
|1<str> Find this definition: |
|2<str> Find functions called by this function: |
|3<str> Find functions calling this function: |
|4<str> Find assignments to: |
|5<str> Change this grep pattern: |
|6<str> Find this egrep pattern: |
|7<str> Find this file: |
|8<str> Find files #including this file: |
+----------------------------------------------------------+
|9<str> Find all function/class definitions (samuel) |
|:<str> Find all function/class definitions (cdb) |
+----------------------------------------------------------+
|a<str> Set the directory search pattern for 'Find this |
| egrep pattern'. This is the equivalent of the |
| command line option of -a |
|^<str> Set the directory filter pattern for all queries |
|c Toggle Caseless mode |
|C Clear file names |
|F Add file name |
|P Print the path to the files |
|r rebuild the database cscope style |
|R rebuild the database samuel style |
+----------------------------------------------------------+
|q Quit |
+----------------------------------------------------------+
| (used by ccalls with the extended menu) |
|0<str> Find this type definition: |
|1<str> Find this preprocessor definition: |
|2<str> Find parameters of this function: |
|3<str> Find functions calling this function: |
|4<str> Find this global definition: |
|5<str> Find this local definition: |
|6<str> Find this structure definition: |
|7<str> Find this file: |
|8<str> Find this preprocessor include: |
|9<str> Find definition of this function: |
|:<str> Find all function/class definitions: |
+----------------------------------------------------------+
If you wanted to "Find assignments to", then you would use
"-4pattern"
The -c, -k, -I, -p, -q, and -T options can also be in the
cscope.files file. See below for common uses of most of
these options.
SYMBOL DATABASE FORMAT
The symbol database (cscope.out) file format is
<header>
<first file symbol data>
...
<last file symbol data>
<trailer>
The header is a single line
cscope <format version> <current dir> [-c] [-q <symbols>] [-T] <trailer offset>
The format version is first number in the cscope version
that wrote the database, e.g. the format version is 9 for
cscope version 9.14. When the format version number in
the cscope program is greater than that in the database,
the entire database will be rebuilt when any part of it is
out of date. The current directory is either a full path
or prefixed by $HOME, allowing the user's login to be
moved to a different file system without rebuilding the
database. The trailer offset is the fseek(3) offset of
the trailer.
The header is followed by the symbol data for each file in
alphabetical order. This allows fast updating of the
database when only a few files have changed, because the
old database is copied to the new database up to the point
of the changed file. Then the changed file's symbol data
is added to the new database, the old database is copied
up to the next changed file, and so on.
Two data compression techniques are used on the symbol
data: (1) keywords and trailing syntax are converted to
control characters, and (2) common digraphs (character
pairs) are compressed to meta-characters (characters with
the eight bit set). Since this makes it hard to read or
postprocess the database, you may want to use the -c
option to turn off data compression when the database is
built.
The symbol data for each file starts with
<file mark><file path>
<empty line>
and for each source line containing a symbol
<line number><blank><non-symbol text>
<optional mark><symbol>
<non-symbol text>
repeat above 2 lines as necessary
<empty line>
Leading and trailing white space in the source line is
removed. Tabs are changed to blanks, and multiple blanks
are squeezed to a single blank, even in character and
string constants. The symbol data for the last file ends
with
<file mark>
A mark is a tab followed by one of these characters:
Char Meaning
@ file
$ function definition
` function call
} function end
# #define
) #define end
~ #include
= direct assignment, increment, or decrement
; enum/struct/union definition end
c class definition
e enum definition
g other global definition
l function/block local definition
m global enum/struct/union member definition
p function parameter definition
s struct definition
t typedef definition
u union definition
A #include mark is followed by '<' or '"' to determine if
the current directory should be searched. An untagged
enum/struct/union definition has a mark without a symbol
name so its beginning can be found.
Historical note: The first three mark characters were cho-
sen because they are not used by the C language, but since
they can appear in a string constant, the tab prefix was
added. They were never changed to more meaningful charac-
ters because other programs had been written to read the
database file.
The trailer contains lists of source directories, include
directories, and source files; its format is
<number of viewpath nodes>
<first directory in viewpath, if any>
...
<last directory in viewpath>
<number of source directories>
<first source directory path (always .)>
...
<last source directory path>
<number of include directories>
<first include directory path>
...
<last include directory path (normally /usr/include)>
<number of source and included files>
<length of string space needed>
<first file path>
...
<last file path>
Low level access to the symbol database was traditionally
provided by ccalls. Cscope now emulates this functional-
ity with the -x option so that everyone can get access to
the extra information stored in the database. One of the
main differences in the output is that a new column is
added, entitled "Ccalls", which contains the symbol name
from the database. This information was previously only
extracted as part of the code fragment associated with
each set of results. As such, the new output format when
using the -x option consists of the file name, function
name, line number, ccalls information and line text, sepa-
rated by spaces.
ENVIRONMENT VARIABLES
CSCOPEOPTIONS
Cscope options.
EDITOR Your editor, which defaults to vi(1).
HOME Your home directory, which is automatically
set when you login.
MOUSE Set to myx if that is your 5620 terminal pro-
gram.
MOUSEMENU Set to none if you do not want cscope to
switch to its mouse menu so you can have your
own customized Cscope submenu on the shell
menu.
SHELL Your shell, which defaults to sh(1).
TERM Your terminal type, which must be a screen
terminal.
TERMINFO Terminal information directory full path
name. If your terminal is not in the stan-
dard terminfo directory, see curses(3X) and
terminfo(4) for how to make your own terminal
description.
TMPDIR Temporary file directory, which defaults to
/usr/tmp.
VIEWER Your file display program, which overrides
EDITOR (see above).
VPATH Viewpath (see Viewpath subsection of EXAMPLES
section).
To set a variable such as EDITOR to a value such as emacs,
put this in your .profile:
export EDITOR
EDITOR=emacs
FILES
cscope.files Default file containing -c, -k, -I, -p, -q,
and -T options and the list of source files
(overridden by the -i option).
cscope.out Default symbol database file (overridden by
the -f option). It is put in your home
directory if it cannot be created in the
current directory.
cscope.in.out
cscope.po.out Default files containing the inverted index
for quick symbol searching (-q option). If
you use the -f option to rename the
cscope.out file name, .in and .po will be
suffixed to its base name to create these
file names.
ncscope*.out Temporary files containing the new database
before it replaces the old database. If you
use the -f option to rename the cscope.out
file name, an n will be prefixed to its base
name to create these file names.
/usr/include Standard directory for #include files.
SEE ALSO
cscalls(1), grep(1), regcmp(3X), slimmer(1)
WARNINGS
Spaces in the displayed source line text for a reference
may not exactly match the actual source line because of
the data compression used to store them in the database.
Leading spaces and tab characters are removed, and
sequences of them are replaced with a single space, even
inside a quoted string. A space may also be added after a
keyword, e.g. "(void)" will be displayed as "(void )".
The Function column of the search output for the "Find
functions called by this function:" input field will only
display the first function called in the line, that is,
for this function
e()
{
return(f() + g());
}
the display would be
Functions called by this function: e
File Function Line
a.c f 3 return(f() + g());
BUGS
C Parsing Bugs
Occasionally, a function definition or call may not be
recognized because of braces inside #if statements. Simi-
larly, the use of a variable may be incorrectly recognized
as a definition.
A typedef name preceding a preprocessor statement will be
incorrectly recognized as a global definition, e.g.
LDFILE *
#if AR16WR
Preprocessor statements can also prevent the recognition
of a global definition, e.g.
char flag
#ifdef ALLOCATE_STORAGE
= -1
#endif
;
A function definition will not be recognized if the '('
after the function name is on the next line, e.g.
f
()
{
C++ Parsing Bugs
Cscope recognizes C++ classes by looking for the class
keyword, but doesn't recognize that a struct is also a
class, so it doesn't recognize inline member function def-
initions in a structure. It also doesn't expect the class
keyword in a typedef, so it incorrectly recognizes X as a
definition in
typedef class X * Y;
It also doesn't recognize operator function definitions
Bool Feature::operator==(const Feature & other) {
and function definitions with a function pointer argument
ParseTable::Recognizes(int startState, char *pattern,
int finishState, void (*FinalAction)(char *)) {
AUTHOR
Joe Steffen
Man(1) output converted with man2html
'리눅스 > 스크립트/유틸' 카테고리의 다른 글
| [llinux] meld 설치 시 문제 해결 방법 (0) | 2014.12.20 |
|---|---|
| cscope + quickfix (0) | 2014.08.29 |
| samba log 관리 (0) | 2014.08.28 |
| iso-8859-1 포맷 파일에 들어간 한글이 깨져나올때 (0) | 2014.08.12 |
| 현재 설정된 java 위치 알아내기 (0) | 2014.07.14 |
