thac0/vim
1
0
Fork 0
forked from aniani/vim
Code Activity

updated for version 7.2a-00

Browse Source
This commit is contained in:
Bram Moolenaar
2008-06-25 20:13:35 +00:00
parent 7b6d4a8ce8
commit aba8857aa3
101 changed files with 490 additions and 12020 deletions
Download Patch File Download Diff File Expand all files Collapse all files

515
runtime/doc/ada.txt
View File

@@ -1,515 +0,0 @@
*ada.txt* For Vim version 7.1. Last change: 2007 May 08
ADA FILE TYPE PLUG-INS REFERENCE MANUAL~
ADA *ada.vim*
1. Syntax Highlighting |ft-ada-syntax|
2. Plug-in |ft-ada-plugin|
3. Omni Completion |ft-ada-omni|
3.1 Omni Completion with "gnat xref" |gnat-xref|
3.2 Omni Completion with "ctags" |ada-ctags|
4. Compiler Support |ada-compiler|
4.1 GNAT |compiler-gnat|
4.1 Dec Ada |compiler-decada|
5. References |ada-reference|
5.1 Options |ft-ada-options|
5.2 Functions |ft-ada-functions|
5.3 Commands |ft-ada-commands|
5.4 Variables |ft-ada-variables|
5.5 Constants |ft-ada-constants|
8. Extra Plug-ins |ada-extra-plugins|
==============================================================================
1. Syntax Highlighting ~
*ft-ada-syntax*
This mode is designed for the 2005 edition of Ada ("Ada 2005"), which includes
support for objected-programming, protected types, and so on. It handles code
written for the original Ada language ("Ada83", "Ada87", "Ada95") as well,
though code which uses Ada 2005-only keywords will be wrongly colored (such
code should be fixed anyway). For more information about Ada, see
http://www.adapower.com.
The Ada mode handles a number of situations cleanly.
For example, it knows that the "-" in "-5" is a number, but the same character
in "A-5" is an operator. Normally, a "with" or "use" clause referencing
another compilation unit is coloured the same way as C's "#include" is coloured.
If you have "Conditional" or "Repeat" groups coloured differently, then "end
if" and "end loop" will be coloured as part of those respective groups.
You can set these to different colours using vim's "highlight" command (e.g.,
to change how loops are displayed, enter the command ":hi Repeat" followed by
the colour specification; on simple terminals the colour specification
ctermfg=White often shows well).
There are several options you can select in this Ada mode. See|ft-ada-options|
for a complete list.
To enable them, assign a value to the option. For example, to turn one on:
>
> let g:ada_standard_types = 1
>
To disable them use ":unlet". Example:
>
> unlet g:ada_standard_types
You can just use ":" and type these into the command line to set these
temporarily before loading an Ada file. You can make these option settings
permanent by adding the "let" command(s), without a colon, to your "~/.vimrc"
file.
Even on a slow (90Mhz) PC this mode works quickly, but if you find the
performance unacceptable, turn on |g:ada_withuse_ordinary|.
Syntax folding instructions (|fold-syntax|) are added when |g:ada_folding| is
set.
==============================================================================
2. File type Plug-in ~
*ft-ada-indent* *ft-ada-plugin*
The Ada plug-in provides support for:
- auto indenting (|indent.txt|)
- insert completion (|i_CTRL-N|)
- user completion (|i_CTRL-X_CTRL-U|)
- tag searches (|tagsrch.txt|)
- Quick Fix (|quickfix.txt|)
- backspace handling (|'backspace'|)
- comment handling (|'comments'|, |'commentstring'|)
The plug-in only activates the features of the Ada mode whenever an Ada
files is opened and add adds Ada related entries to the main and pop-up menu.
==============================================================================
3. Omni Completion ~
*ft-ada-omni*
The Ada omni-completions (|i_CTRL-X_CTRL-O|) uses tags database created either
by "gnat xref -v" or the "exuberant Ctags (http://ctags.sourceforge.net). The
complete function will automatically detect which tool was used to create the
tags file.
------------------------------------------------------------------------------
3.1 Omni Completion with "gnat xref" ~
*gnat-xref*
GNAT XREF uses the compiler internal informations (ali-files) to produce the
tags file. This has the advantage to be 100% correct and the option of deep
nested analysis. However the code must compile, the generator is quite
slow and the created tags file contains only the basic Ctags informations for
each entry - not enough for some of the more advanced Vim code browser
plug-ins.
NOTE: "gnat xref -v" is very tricky to use as it has almost no diagnostic
output - If nothing is printed then usually the parameters are wrong.
Here some important tips:
1) You need to compile your code first and use the "-aO" option to point to
your .ali files.
2) "gnat xref -v ../Include/adacl.ads" won't work - use the "gnat xref -v
-aI../Include adacl.ads" instead.
3) "gnat xref -v -aI../Include *.ad?" won't work - use "cd ../Include" and
then "gnat xref -v *.ad?"
4) Project manager support is completely broken - don't even try "gnat xref
-Padacl.gpr".
5) VIM is faster when the tags file is sorted - use "sort --unique
--ignore-case --output=tags tags" .
6) Remember to insert "!_TAG_FILE_SORTED 2 %sort ui" as first line to mark
the file assorted.
------------------------------------------------------------------------------
3.2 Omni Completion with "ctags"~
*ada-ctags*
Exuberant Ctags uses its own multi-language code parser. The parser is quite
fast, produces a lot of extra informations (hence the name "Exuberant Ctags")
and can run on files which currently do not compile.
There are also lots of other Vim-tools which use exuberant Ctags.
You will need to install a version of the Exuberant Ctags which has Ada
support patched in. Such a version is available from the GNU Ada Project
(http://gnuada.sourceforge.net).
The Ada parser for Exuberant Ctags is fairly new - don't expect complete
support yet.
==============================================================================
4. Compiler Support ~
*ada-compiler*
The Ada mode supports more then one Ada compiler and will automatically load the
compiler set in|g:ada_default_compiler|whenever an Ada source is opened. The
provided compiler plug-ins are split into the actual compiler plug-in and a
collection of support functions and variables. This allows the easy
development of specialized compiler plug-ins fine tuned to your development
environment.
------------------------------------------------------------------------------
4.1 GNAT ~
*compiler-gnat*
GNAT is the only free (beer and speech) Ada compiler available. There are
several version available which differentiate in the licence terms used.
The GNAT compiler plug-in will perform a compile on pressing <F7> and then
immediately shows the result. You can set the project file to be used by
setting:
>
> call g:gnat.Set_Project_File ('my_project.gpr')
Setting a project file will also create a Vim session (|views-sessions|) so -
like with the GPS - opened files, window positions etc. will remembered
separately for all projects.
*gnat_members*
GNAT OBJECT ~
*g:gnat.Make()*
g:gnat.Make()
Calls|g:gnat.Make_Command|and displays the result inside a
|quickfix| window.
*g:gnat.Pretty()*
g:gnat.Pretty()
Calls|g:gnat.Pretty_Command|
*g:gnat.Find()*
g:gnat.Find()
Calls|g:gnat.Find_Command|
*g:gnat.Tags()*
g:gnat.Tags()
Calls|g:gnat.Tags_Command|
*g:gnat.Set_Project_File()*
g:gnat.Set_Project_File([{file}])
Set gnat project file and load associated session. An open
project will be closed and the session written. If called
without file name the file selector opens for selection of a
project file. If called with an empty string then the project
and associated session are closed.
*g:gnat.Project_File*
g:gnat.Project_File string
Current project file.
*g:gnat.Make_Command*
g:gnat.Make_Command string
External command used for|g:gnat.Make()| (|'makeprg'|).
*g:gnat.Pretty_Program*
g:gnat.Pretty_Program string
External command used for|g:gnat.Pretty()|
*g:gnat.Find_Program*
g:gnat.Find_Program string
External command used for|g:gnat.Find()|
*g:gnat.Tags_Command*
g:gnat.Tags_Command string
External command used for|g:gnat.Tags()|
*g:gnat.Error_Format*
g:gnat.Error_Format string
Error format (|'errorformat'|)
------------------------------------------------------------------------------
4.2 Dec Ada ~
*compiler-hpada* *compiler-decada*
*compiler-vaxada* *compiler-compaqada*
Dec Ada (also known by - in chronological order - VAX Ada, Dec Ada, Compaq Ada
and HP Ada) is a fairly dated Ada 83 compiler. Support is basic: <F7> will
compile the current unit.
The Dec Ada compiler expects the package name and not the file name to be
passed a parameter. The compiler plug-in supports the usual file name
convention to convert the file into a unit name. For separates both '-' and
'__' are allowed.
*decada_members*
DEC ADA OBJECT ~
*g:decada.Make()*
g:decada.Make() function
Calls|g:decada.Make_Command|and displays the result inside a
|quickfix| window.
*g:decada.Unit_Name()*
g:decada.Unit_Name() function
Get the Unit name for the current file.
*g:decada.Make_Command*
g:decada.Make_Command string
External command used for|g:decadat.Make()| (|'makeprg'|).
*g:decada.Error_Format*
g:decada.Error_Format| string
Error format (|'errorformat'|).
==============================================================================
5. References ~
*ada-reference*
------------------------------------------------------------------------------
5.1 Options ~
*ft-ada-options*
*g:ada_standard_types*
g:ada_standard_types bool (true when exists)
Highlight types in package Standard (e.g., "Float")
*g:ada_space_errors*
*g:ada_no_trail_space_error*
*g:ada_no_tab_space_error*
*g:ada_all_tab_usage*
g:ada_space_errors bool (true when exists)
Highlight extraneous errors in spaces ...
g:ada_no_trail_space_error
- but ignore trailing spaces at the end of a line
g:ada_no_tab_space_error
- but ignore tabs after spaces
g:ada_all_tab_usage
- highlight all tab use
*g:ada_line_errors*
g:ada_line_errors bool (true when exists)
Highlight lines which are to long. Note: This highlighting
option is quite CPU intensive.
*g:ada_rainbow_color*
g:ada_rainbow_color bool (true when exists)
Use rainbow colours for '(' and ')'. You need the
rainbow_parenthesis for this to work
*g:ada_folding*
g:ada_folding set ('sigpft')
Use folding for Ada sources.
's': activate syntax folding on load
'p': fold packages
'f': fold functions and procedures
't': fold types
'c': fold conditionals
'g': activate gnat pretty print folding on load
'i': lone 'is' folded with line above
'b': lone 'begin' folded with line above
'p': lone 'private' folded with line above
'x': lone 'exception' folded with line above
'i': activate indent folding on load
Note: Syntax folding is in an early (unusable) stage and
indent or gnat pretty folding is suggested.
For gnat pretty folding to work the following settings are
suggested: -cl3 -M79 -c2 -c3 -c4 -A1 -A2 -A3 -A4 -A5
For indent folding to work the following settings are
suggested: shiftwidth=3 softtabstop=3
*g:ada_abbrev*
g:ada_abbrev bool (true when exists)
Add some abbreviations. This feature more or less superseded
by the various completion methods.
*g:ada_withuse_ordinary*
g:ada_withuse_ordinary bool (true when exists)
Show "with" and "use" as ordinary keywords (when used to
reference other compilation units they're normally highlighted
specially).
*g:ada_begin_preproc*
g:ada_begin_preproc bool (true when exists)
Show all begin-like keywords using the colouring of C
preprocessor commands.
*g:ada_omni_with_keywords*
g:ada_omni_with_keywords
Add Keywords, Pragmas, Attributes to omni-completions
(|compl-omni|). Note: You can always complete then with user
completion (|i_CTRL-X_CTRL-U|).
*g:ada_extended_tagging*
g:ada_extended_tagging enum ('jump', 'list')
use extended tagging, two options are available
'jump': use tjump to jump.
'list': add tags quick fix list.
Normal tagging does not support function or operator
overloading as these features are not available in C and
tagging was originally developed for C.
*g:ada_extended_completion*
g:ada_extended_completion
Uses extended completion for <C-N> and <C-R> completions
(|i_CTRL-N|). In this mode the '.' is used as part of the
identifier so that 'Object.Method' or 'Package.Procedure' are
completed together.
*g:ada_gnat_extensions*
g:ada_gnat_extensions bool (true when exists)
Support GNAT extensions.
*g:ada_with_gnat_project_files*
g:ada_with_gnat_project_files bool (true when exists)
Add gnat project file keywords and Attributes.
*g:ada_default_compiler*
g:ada_default_compiler string
set default compiler. Currently supported is 'gnat' and
'decada'.
An "exists" type is a boolean is considered true when the variable is defined
and false when the variable is undefined. The value which the variable is
set makes no difference.
------------------------------------------------------------------------------
5.3 Commands ~
*ft-ada-commands*
:AdaRainbow *:AdaRainbow*
Toggles rainbow colour (|g:ada_rainbow_color|) mode for
'(' and ')'
:AdaLines *:AdaLines*
Toggles line error (|g:ada_line_errors|) display
:AdaSpaces *:AdaSpaces*
Toggles space error (|g:ada_space_errors|) display.
:AdaTagDir *:AdaTagDir*
Creates tags file for the directory of the current file.
:AdaTagFile *:AdaTagFile*
Creates tags file for the current file.
:AdaTypes *:AdaTypes*
Toggles standard types (|g:ada_standard_types|) colour.
:GnatFind *:GnatFind*
Calls |g:gnat.Find()|
:GnatPretty *:GnatPretty*
Calls |g:gnat.Pretty()|
:GnatTags *:GnatTags*
Calls |g:gnat.Tags()|
------------------------------------------------------------------------------
5.3 Variables ~
*ft-ada-variables*
*g:gnat*
g:gnat object
Control object which manages GNAT compiles. The object
is created when the first Ada source code is loaded provided
that |g:ada_default_compiler|is set to 'gnat'. See|gnat_members|
for details.
*g:decada*
g:decada object
Control object which manages Dec Ada compiles. The object
is created when the first Ada source code is loaded provided
that |g:ada_default_compiler|is set to 'decada'. See
|decada_members|for details.
------------------------------------------------------------------------------
5.4 Constants ~
*ft-ada-constants*
All constants are locked. See |:lockvar| for details.
*g:ada#WordRegex*
g:ada#WordRegex string
Regular expression to search for Ada words
*g:ada#DotWordRegex*
g:ada#DotWordRegex string
Regular expression to search for Ada words separated by dots.
*g:ada#Comment*
g:ada#Comment string
Regular expression to search for Ada comments
*g:ada#Keywords*
g:ada#Keywords list of dictionaries
List of keywords, attributes etc. pp. in the format used by
omni completion. See |complete-items| for details.
*g:ada#Ctags_Kinds*
g:ada#Ctags_Kinds dictionary of lists
Dictionary of the various kinds of items which the Ada support
for Ctags generates.
------------------------------------------------------------------------------
5.2 Functions ~
*ft-ada-functions*
ada#Word([{line}, {col}]) *ada#Word()*
Return full name of Ada entity under the cursor (or at given
line/column), stripping white space/newlines as necessary.
ada#List_Tag([{line}, {col}]) *ada#Listtags()*
List all occurrences of the Ada entity under the cursor (or at
given line/column) inside the quick-fix window
ada#Jump_Tag ({ident}, {mode}) *ada#Jump_Tag()*
List all occurrences of the Ada entity under the cursor (or at
given line/column) in the tag jump list. Mode can either be
'tjump' or 'stjump'.
ada#Create_Tags ({option}) *ada#Create_Tags()*
Creates tag file using Ctags. The option can either be 'file'
for the current file, 'dir' for the directory of the current
file or a file name.
gnat#Insert_Tags_Header() *gnat#Insert_Tags_Header()*
Adds the tag file header (!_TAG_) informations to the current
file which are missing from the GNAT XREF output.
ada#Switch_Syntax_Option ({option}) *ada#Switch_Syntax_Option()*
Toggles highlighting options on or off. Used for the Ada menu.
*gnat#New()*
gnat#New ()
Create a new gnat object. See |g:gnat| for details.
==============================================================================
8. Extra Plugins ~
*ada-extra-plugins*
You can optionally install the following extra plug-in. They work well with Ada
and enhance the ability of the Ada mode.:
backup.vim
http://www.vim.org/scripts/script.php?script_id=1537
Keeps as many backups as you like so you don't have to.
rainbow_parenthsis.vim
http://www.vim.org/scripts/script.php?script_id=1561
Very helpful since Ada uses only '(' and ')'.
nerd_comments.vim
http://www.vim.org/scripts/script.php?script_id=1218
Excellent commenting and uncommenting support for almost any
programming language.
matchit.vim
http://www.vim.org/scripts/script.php?script_id=39
'%' jumping for any language. The normal '%' jump only works for '{}'
style languages. The Ada mode will set the needed search patters.
taglist.vim
http://www.vim.org/scripts/script.php?script_id=273
Source code explorer sidebar. There is a patch for Ada available.
The GNU Ada Project distribution (http://gnuada.sourceforge.net) of Vim
contains all of the above.
==============================================================================
vim: textwidth=78 nowrap tabstop=8 shiftwidth=4 softtabstop=4 noexpandtab
vim: filetype=help encoding=latin1

763
runtime/doc/sql.txt
View File

@@ -1,763 +0,0 @@
*sql.txt* For Vim version 7.1. Last change: Wed Apr 26 2006 3:05:33 PM
by David Fishburn
This is a filetype plugin to work with SQL files.
The Structured Query Language (SQL) is a standard which specifies statements
that allow a user to interact with a relational database. Vim includes
features for navigation, indentation and syntax highlighting.
1. Navigation |sql-navigation|
1.1 Matchit |sql-matchit|
1.2 Text Object Motions |sql-object-motions|
1.3 Predefined Object Motions |sql-predefined-objects|
1.4 Macros |sql-macros|
2. SQL Dialects |sql-dialects|
2.1 SQLSetType |SQLSetType|
2.2 SQL Dialect Default |sql-type-default|
3. Adding new SQL Dialects |sql-adding-dialects|
4. OMNI SQL Completion |sql-completion|
4.1 Static mode |sql-completion-static|
4.2 Dynamic mode |sql-completion-dynamic|
4.3 Tutorial |sql-completion-tutorial|
4.3.1 Complete Tables |sql-completion-tables|
4.3.2 Complete Columns |sql-completion-columns|
4.3.3 Complete Procedures |sql-completion-procedures|
4.3.4 Complete Views |sql-completion-views|
4.4 Completion Customization |sql-completion-customization|
4.5 SQL Maps |sql-completion-maps|
4.6 Using with other filetypes |sql-completion-filetypes|
==============================================================================
1. Navigation *sql-navigation*
The SQL ftplugin provides a number of options to assist with file
navigation.
1.1 Matchit *sql-matchit*
-----------
The matchit plugin (http://www.vim.org/scripts/script.php?script_id=39)
provides many additional features and can be customized for different
languages. The matchit plugin is configured by defining a local
buffer variable, b:match_words. Pressing the % key while on various
keywords will move the cursor to its match. For example, if the cursor
is on an "if", pressing % will cycle between the "else", "elseif" and
"end if" keywords.
The following keywords are supported: >
if
elseif | elsif
else [if]
end if
[while condition] loop
leave
break
continue
exit
end loop
for
leave
break
continue
exit
end loop
do
statements
doend
case
when
when
default
end case
merge
when not matched
when matched
create[ or replace] procedure|function|event
returns
1.2 Text Object Motions *sql-object-motions*
-----------------------
Vim has a number of predefined keys for working with text |object-motions|.
This filetype plugin attempts to translate these keys to maps which make sense
for the SQL language.
The following |Normal| mode and |Visual| mode maps exist (when you edit a SQL
file): >
]] move forward to the next 'begin'
[[ move backwards to the previous 'begin'
][ move forward to the next 'end'
[] move backwards to the previous 'end'
1.3 Predefined Object Motions *sql-predefined-objects*
-----------------------------
Most relational databases support various standard features, tables, indices,
triggers and stored procedures. Each vendor also has a variety of proprietary
objects. The next set of maps have been created to help move between these
objects. Depends on which database vendor you are using, the list of objects
must be configurable. The filetype plugin attempts to define many of the
standard objects, plus many additional ones. In order to make this as
flexible as possible, you can override the list of objects from within your
|vimrc| with the following: >
let g:ftplugin_sql_objects = 'function,procedure,event,table,trigger' .
\ ',schema,service,publication,database,datatype,domain' .
\ ',index,subscription,synchronization,view,variable'
The following |Normal| mode and |Visual| mode maps have been created which use
the above list: >
]} move forward to the next 'create <object name>'
[{ move backward to the previous 'create <object name>'
Repeatedly pressing ]} will cycle through each of these create statements: >
create table t1 (
...
);
create procedure p1
begin
...
end;
create index i1 on t1 (c1);
The default setting for g:ftplugin_sql_objects is: >
let g:ftplugin_sql_objects = 'function,procedure,event,' .
\ '\\(existing\\\\|global\\s\\+temporary\\s\\+\\)\\\{,1}' .
\ 'table,trigger' .
\ ',schema,service,publication,database,datatype,domain' .
\ ',index,subscription,synchronization,view,variable'
The above will also handle these cases: >
create table t1 (
...
);
create existing table t2 (
...
);
create global temporary table t3 (
...
);
By default, the ftplugin only searches for CREATE statements. You can also
override this via your |vimrc| with the following: >
let g:ftplugin_sql_statements = 'create,alter'
The filetype plugin defines three types of comments: >
1. --
2. //
3. /*
*
*/
The following |Normal| mode and |Visual| mode maps have been created to work
with comments: >
]" move forward to the beginning of a comment
[" move forward to the end of a comment
1.4 Macros *sql-macros*
----------
Vim's feature to find macro definitions, |'define'|, is supported using this
regular expression: >
\c\<\(VARIABLE\|DECLARE\|IN\|OUT\|INOUT\)\>
This addresses the following code: >
CREATE VARIABLE myVar1 INTEGER;
CREATE PROCEDURE sp_test(
IN myVar2 INTEGER,
OUT myVar3 CHAR(30),
INOUT myVar4 NUMERIC(20,0)
)
BEGIN
DECLARE myVar5 INTEGER;
SELECT c1, c2, c3
INTO myVar2, myVar3, myVar4
FROM T1
WHERE c4 = myVar1;
END;
Place your cursor on "myVar1" on this line: >
WHERE c4 = myVar1;
^
Press any of the following keys: >
[d
[D
[CTRL-D
==============================================================================
2. SQL Dialects *sql-dialects* *sql-types*
*sybase* *TSQL* *Transact-SQL*
*sqlanywhere*
*oracle* *plsql* *sqlj*
*sqlserver*
*mysql* *postgres* *psql*
*informix*
All relational databases support SQL. There is a portion of SQL that is
portable across vendors (ex. CREATE TABLE, CREATE INDEX), but there is a
great deal of vendor specific extensions to SQL. Oracle supports the
"CREATE OR REPLACE" syntax, column defaults specified in the CREATE TABLE
statement and the procedural language (for stored procedures and triggers).
The default Vim distribution ships with syntax highlighting based on Oracle's
PL/SQL. The default SQL indent script works for Oracle and SQL Anywhere.
The default filetype plugin works for all vendors and should remain vendor
neutral, but extendable.
Vim currently has support for a variety of different vendors, currently this
is via syntax scripts. Unfortunately, to flip between different syntax rules
you must either create:
1. New filetypes
2. Custom autocmds
3. Manual steps / commands
The majority of people work with only one vendor's database product, it would
be nice to specify a default in your |vimrc|.
2.1 SQLSetType *sqlsettype* *SQLSetType*
--------------
For the people that work with many different databases, it would be nice to be
able to flip between the various vendors rules (indent, syntax) on a per
buffer basis, at any time. The ftplugin/sql.vim file defines this function: >
SQLSetType
Executing this function without any parameters will set the indent and syntax
scripts back to their defaults, see |sql-type-default|. If you have turned
off Vi's compatibility mode, |'compatible'|, you can use the <Tab> key to
complete the optional parameter.
After typing the function name and a space, you can use the completion to
supply a parameter. The function takes the name of the Vim script you want to
source. Using the |cmdline-completion| feature, the SQLSetType function will
search the |'runtimepath'| for all Vim scripts with a name containing 'sql'.
This takes the guess work out of the spelling of the names. The following are
examples: >
:SQLSetType
:SQLSetType sqloracle
:SQLSetType sqlanywhere
:SQLSetType sqlinformix
:SQLSetType mysql
The easiest approach is to the use <Tab> character which will first complete
the command name (SQLSetType), after a space and another <Tab>, display a list
of available Vim script names: >
:SQL<Tab><space><Tab>
2.2 SQL Dialect Default *sql-type-default*
-----------------------
As mentioned earlier, the default syntax rules for Vim is based on Oracle
(PL/SQL). You can override this default by placing one of the following in
your |vimrc|: >
let g:sql_type_default = 'sqlanywhere'
let g:sql_type_default = 'sqlinformix'
let g:sql_type_default = 'mysql'
If you added the following to your |vimrc|: >
let g:sql_type_default = 'sqlinformix'
The next time edit a SQL file the following scripts will be automatically
loaded by Vim: >
ftplugin/sql.vim
syntax/sqlinformix.vim
indent/sql.vim
>
Notice indent/sqlinformix.sql was not loaded. There is no indent file
for Informix, Vim loads the default files if the specified files does not
exist.
==============================================================================
3. Adding new SQL Dialects *sql-adding-dialects*
If you begin working with a SQL dialect which does not have any customizations
available with the default Vim distribution you can check http://www.vim.org
to see if any customization currently exist. If not, you can begin by cloning
an existing script. Read |filetype-plugins| for more details.
To help identify these scripts, try to create the files with a "sql" prefix.
If you decide you wish to create customizations for the SQLite database, you
can create any of the following: >
Unix
~/.vim/syntax/sqlite.vim
~/.vim/indent/sqlite.vim
Windows
$VIM/vimfiles/syntax/sqlite.vim
$VIM/vimfiles/indent/sqlite.vim
No changes are necessary to the SQLSetType function. It will automatically
pickup the new SQL files and load them when you issue the SQLSetType command.
==============================================================================
4. OMNI SQL Completion *sql-completion*
*omni-sql-completion*
Vim 7 includes a code completion interface and functions which allows plugin
developers to build in code completion for any language. Vim 7 includes
code completion for the SQL language.
There are two modes to the SQL completion plugin, static and dynamic. The
static mode populates the popups with the data generated from current syntax
highlight rules. The dynamic mode populates the popups with data retrieved
directly from a database. This includes, table lists, column lists,
procedures names and more.
4.1 Static Mode *sql-completion-static*
---------------
The static popups created contain items defined by the active syntax rules
while editing a file with a filetype of SQL. The plugin defines (by default)
various maps to help the user refine the list of items to be displayed.
The defaults static maps are: >
imap <buffer> <C-C>a <C-\><C-O>:call sqlcomplete#Map('syntax')<CR><C-X><C-O>
imap <buffer> <C-C>k <C-\><C-O>:call sqlcomplete#Map('sqlKeyword')<CR><C-X><C-O>
imap <buffer> <C-C>f <C-\><C-O>:call sqlcomplete#Map('sqlFunction')<CR><C-X><C-O>
imap <buffer> <C-C>o <C-\><C-O>:call sqlcomplete#Map('sqlOption')<CR><C-X><C-O>
imap <buffer> <C-C>T <C-\><C-O>:call sqlcomplete#Map('sqlType')<CR><C-X><C-O>
imap <buffer> <C-C>s <C-\><C-O>:call sqlcomplete#Map('sqlStatement')<CR><C-X><C-O>
The static maps (which are based on the syntax highlight groups) follow this
format: >
imap <buffer> <C-C>k <C-\><C-O>:call sqlcomplete#Map('sqlKeyword')<CR><C-X><C-O>
This command breaks down as: >
imap - Create an insert map
<buffer> - Only for this buffer
<C-C>k - Your choice of key map
<C-\><C-O> - Execute one command, return to Insert mode
:call sqlcomplete#Map( - Allows the SQL completion plugin to perform some
housekeeping functions to allow it to be used in
conjunction with other completion plugins.
Indicate which item you want the SQL completion
plugin to complete.
In this case we are asking the plugin to display
items from the syntax highlight group
'sqlKeyword'.
You can view a list of highlight group names to
choose from by executing the
:syntax list
command while editing a SQL file.
'sqlKeyword' - Display the items for the sqlKeyword highlight
group
)<CR> - Execute the :let command
<C-X><C-O> - Trigger the standard omni completion key stroke.
Passing in 'sqlKeyword' instructs the SQL
completion plugin to populate the popup with
items from the sqlKeyword highlight group. The
plugin will also cache this result until Vim is
restarted. The syntax list is retrieved using
the syntaxcomplete plugin.
Using the 'syntax' keyword is a special case. This instructs the
syntaxcomplete plugin to retrieve all syntax items. So this will effectively
work for any of Vim's SQL syntax files. At the time of writing this includes
10 different syntax files for the different dialects of SQL (see section 3
above, |sql-dialects|).
Here are some examples of the entries which are pulled from the syntax files: >
All
- Contains the contents of all syntax highlight groups
Statements
- Select, Insert, Update, Delete, Create, Alter, ...
Functions
- Min, Max, Trim, Round, Date, ...
Keywords
- Index, Database, Having, Group, With
Options
- Isolation_level, On_error, Qualify_owners, Fire_triggers, ...
Types
- Integer, Char, Varchar, Date, DateTime, Timestamp, ...
4.2 Dynamic Mode *sql-completion-dynamic*
----------------
Dynamic mode populates the popups with data directly from a database. In
order for the dynamic feature to be enabled you must have the dbext.vim
plugin installed, (http://vim.sourceforge.net/script.php?script_id=356).
Dynamic mode is used by several features of the SQL completion plugin.
After installing the dbext plugin see the dbext-tutorial for additional
configuration and usage. The dbext plugin allows the SQL completion plugin
to display a list of tables, procedures, views and columns. >
Table List
- All tables for all schema owners
Procedure List
- All stored procedures for all schema owners
View List
- All stored procedures for all schema owners
Column List
- For the selected table, the columns that are part of the table
To enable the popup, while in INSERT mode, use the following key combinations
for each group (where <C-C> means hold the CTRL key down while pressing
the space bar):
Table List - <C-C>t
- <C-X><C-O> (the default map assumes tables)
Stored Procedure List - <C-C>p
View List - <C-C>v
Column List - <C-C>c
Windows platform only - When viewing a popup window displaying the list
of tables, you can press <C-Right>, this will
replace the table currently highlighted with
the column list for that table.
- When viewing a popup window displaying the list
of columns, you can press <C-Left>, this will
replace the column list with the list of tables.
- This allows you to quickly drill down into a
table to view it's columns and back again.
The SQL completion plugin caches various lists that are displayed in
the popup window. This makes the re-displaying of these lists very
fast. If new tables or columns are added to the database it may become
necessary to clear the plugins cache. The default map for this is: >
imap <buffer> <C-C>R <C-\><C-O>:call sqlcomplete#Map('ResetCache')<CR><C-X><C-O>
4.3 SQL Tutorial *sql-completion-tutorial*
----------------
This tutorial is designed to take you through the common features of the SQL
completion plugin so that: >
a) You gain familiarity with the plugin
b) You are introduced to some of the more common features
c) Show how to customize it to your preferences
d) Demonstrate "Best of Use" of the plugin (easiest way to configure).
First, create a new buffer: >
:e tutorial.sql
Static features
---------------
To take you through the various lists, simply enter insert mode, hit:
<C-C>s (show SQL statements)
At this point, you can page down through the list until you find "select".
If you are familiar with the item you are looking for, for example you know
the statement begins with the letter "s". You can type ahead (without the
quotes) "se" then press:
<C-Space>t
Assuming "select" is highlighted in the popup list press <Enter> to choose
the entry. Now type:
* fr<C-C>a (show all syntax items)
choose "from" from the popup list.
When writing stored procedures using the "type" list is useful. It contains
a list of all the database supported types. This may or may not be true
depending on the syntax file you are using. The SQL Anywhere syntax file
(sqlanywhere.vim) has support for this: >
BEGIN
DECLARE customer_id <C-C>T <-- Choose a type from the list
Dynamic features
----------------
To take advantage of the dynamic features you must first install the
dbext.vim plugin (http://vim.sourceforge.net/script.php?script_id=356). It
also comes with a tutorial. From the SQL completion plugin's perspective,
the main feature dbext provides is a connection to a database. dbext
connection profiles are the most efficient mechanism to define connection
information. Once connections have been setup, the SQL completion plugin
uses the features of dbext in the background to populate the popups.
What follows assumes dbext.vim has been correctly configured, a simple test
is to run the command, :DBListTable. If a list of tables is shown, you know
dbext.vim is working as expected. If not, please consult the dbext.txt
documentation.
Assuming you have followed the dbext-tutorial you can press <C-C>t to
display a list of tables. There is a delay while dbext is creating the table
list. After the list is displayed press <C-W>. This will remove both the
popup window and the table name already chosen when the list became active. >
4.3.1 Table Completion: *sql-completion-tables*
Press <C-C>t to display a list of tables from within the database you
have connected via the dbext plugin.
NOTE: All of the SQL completion popups support typing a prefix before pressing
the key map. This will limit the contents of the popup window to just items
beginning with those characters. >
4.3.2 Column Completion: *sql-completion-columns*
The SQL completion plugin can also display a list of columns for particular
tables. The column completion is trigger via <C-C>c.
NOTE: The following example uses <C-Right> to trigger a column list while
the popup window is active. This map is only available on the Windows
platforms since *nix does not recognize CTRL and the right arrow held down
together. If you wish to enable this functionality on a *nix platform choose
a key and create one of these mappings (see |sql-completion-maps| for further
details on where to create this imap): >
imap <buffer> <your_keystroke> <C-R>=sqlcomplete#DrillIntoTable()<CR>
imap <buffer> <your_keystroke> <C-Y><C-\><C-O>:call sqlcomplete#Map('column')<CR><C-X><C-O>
Example of using column completion:
- Press <C-C>t again to display the list of tables.
- When the list is displayed in the completion window, press <C-Right>,
this will replace the list of tables, with a list of columns for the
table highlighted (after the same short delay).
- If you press <C-Left>, this will again replace the column list with the
list of tables. This allows you to drill into tables and column lists
very quickly.
- Press <C-Right> again while the same table is highlighted. You will
notice there is no delay since the column list has been cached. If you
change the schema of a cached table you can press <C-C>R, which
clears the SQL completion cache.
- NOTE: <C-Right> and <C-Left> have been designed to work while the
completion window is active. If the completion popup window is
not active, a normal <C-Right> or <C-Left> will be executed.
Lets look how we can build a SQL statement dynamically. A select statement
requires a list of columns. There are two ways to build a column list using
the SQL completion plugin. >
One column at a time:
< 1. After typing SELECT press <C-C>t to display a list of tables.
2. Choose a table from the list.
3. Press <C-Right> to display a list of columns.
4. Choose the column from the list and press enter.
5. Enter a "," and press <C-C>c. Generating a column list
generally requires having the cursor on a table name. The plugin
uses this name to determine what table to retrieve the column list.
In this step, since we are pressing <C-C>c without the cursor
on a table name the column list displayed will be for the previous
table. Choose a different column and move on.
6. Repeat step 5 as often as necessary. >
All columns for a table:
< 1. After typing SELECT press <C-C>t to display a list of tables.
2. Highlight the table you need the column list for.
3. Press <Enter> to choose the table from the list.
4. Press <C-C>l to request a comma separated list of all columns
for this table.
5. Based on the table name chosen in step 3, the plugin attempts to
decide on a reasonable table alias. You are then prompted to
either accept of change the alias. Press OK.
6. The table name is replaced with the column list of the table is
replaced with the comma separate list of columns with the alias
prepended to each of the columns.
7. Step 3 and 4 can be replaced by pressing <C-C>L, which has
a <C-Y> embedded in the map to choose the currently highlighted
table in the list.
There is a special provision when writing select statements. Consider the
following statement: >
select *
from customer c,
contact cn,
department as dp,
employee e,
site_options so
where c.
In INSERT mode after typing the final "c." which is an alias for the
"customer" table, you can press either <C-C>c or <C-X><C-O>. This will
popup a list of columns for the customer table. It does this by looking back
to the beginning of the select statement and finding a list of the tables
specified in the FROM clause. In this case it notes that in the string
"customer c", "c" is an alias for the customer table. The optional "AS"
keyword is also supported, "customer AS c". >
4.3.3 Procedure Completion: *sql-completion-procedures*
Similar to the table list, <C-C>p, will display a list of stored
procedures stored within the database. >
4.3.4 View Completion: *sql-completion-views*
Similar to the table list, <C-C>v, will display a list of views in the
database.
4.4 Completion Customization *sql-completion-customization*
----------------------------
The SQL completion plugin can be customized through various options set in
your |vimrc|: >
omni_sql_no_default_maps
< - Default: This variable is not defined
- If this variable is defined, no maps are created for OMNI
completion. See |sql-completion-maps| for further discussion.
>
omni_sql_use_tbl_alias
< - Default: a
- This setting is only used when generating a comma separated
column list. By default the map is <C-C>l. When generating
a column list, an alias can be prepended to the beginning of each
column, for example: e.emp_id, e.emp_name. This option has three
settings: >
n - do not use an alias
d - use the default (calculated) alias
a - ask to confirm the alias name
<
An alias is determined following a few rules:
1. If the table name has an '_', then use it as a separator: >
MY_TABLE_NAME --> MTN
my_table_name --> mtn
My_table_NAME --> MtN
< 2. If the table name does NOT contain an '_', but DOES use
mixed case then the case is used as a separator: >
MyTableName --> MTN
< 3. If the table name does NOT contain an '_', and does NOT
use mixed case then the first letter of the table is used: >
mytablename --> m
MYTABLENAME --> M
omni_sql_ignorecase
< - Default: Current setting for|ignorecase|
- Valid settings are 0 or 1.
- When entering a few letters before initiating completion, the list
will be filtered to display only the entries which begin with the
list of characters. When this option is set to 0, the list will be
filtered using case sensitivity. >
omni_sql_include_owner
< - Default: 0, unless dbext.vim 3.00 has been installed
- Valid settings are 0 or 1.
- When completing tables, procedure or views and using dbext.vim 3.00
or higher the list of objects will also include the owner name.
When completing these objects and omni_sql_include_owner is enabled
the owner name will be replaced. >
omni_sql_precache_syntax_groups
< - Default:
['syntax','sqlKeyword','sqlFunction','sqlOption','sqlType','sqlStatement']
- sqlcomplete can be used in conjunction with other completion
plugins. This is outlined at |sql-completion-filetypes|. When the
filetype is changed temporarily to SQL, the sqlcompletion plugin
will cache the syntax groups listed in the List specified in this
option.
>
4.5 SQL Maps *sql-completion-maps*
------------
The default SQL maps have been described in other sections of this document in
greater detail. Here is a list of the maps with a brief description of each.
Static Maps
-----------
These are maps which use populate the completion list using Vim's syntax
highlighting rules. >
<C-C>a
< - Displays all SQL syntax items. >
<C-C>k
< - Displays all SQL syntax items defined as 'sqlKeyword'. >
<C-C>f
< - Displays all SQL syntax items defined as 'sqlFunction. >
<C-C>o
< - Displays all SQL syntax items defined as 'sqlOption'. >
<C-C>T
< - Displays all SQL syntax items defined as 'sqlType'. >
<C-C>s
< - Displays all SQL syntax items defined as 'sqlStatement'. >
Dynamic Maps
------------
These are maps which use populate the completion list using the dbext.vim
plugin. >
<C-C>t
< - Displays a list of tables. >
<C-C>p
< - Displays a list of procedures. >
<C-C>v
< - Displays a list of views. >
<C-C>c
< - Displays a list of columns for a specific table. >
<C-C>l
< - Displays a comma separated list of columns for a specific table. >
<C-C>L
< - Displays a comma separated list of columns for a specific table.
This should only be used when the completion window is active. >
<C-Right>
< - Displays a list of columns for the table currently highlighted in
the completion window. <C-Right> is not recognized on most Unix
systems, so this maps is only created on the Windows platform.
If you would like the same feature on Unix, choose a different key
and make the same map in your vimrc. >
<C-Left>
< - Displays the list of tables.
<C-Left> is not recognized on most Unix systems, so this maps is
only created on the Windows platform. If you would like the same
feature on Unix, choose a different key and make the same map in
your vimrc. >
<C-C>R
< - This maps removes all cached items and forces the SQL completion
to regenerate the list of items.
Customizing Maps
----------------
You can create as many additional key maps as you like. Generally, the maps
will be specifying different syntax highlight groups.
If you do not wish the default maps created or the key choices do not work on
your platform (often a case on *nix) you define the following variable in
your |vimrc|: >
let g:omni_sql_no_default_maps = 1
Do no edit ftplugin/sql.vim directly! If you change this file your changes
will be over written on future updates. Vim has a special directory structure
which allows you to make customizations without changing the files that are
included with the Vim distribution. If you wish to customize the maps
create an after/ftplugin/sql.vim (see |after-directory|) and place the same
maps from the ftplugin/sql.vim in it using your own key strokes. <C-C> was
chosen since it will work on both Windows and *nix platforms. On the windows
platform you can also use <C-Space> or ALT keys.
4.6 Using with other filetypes *sql-completion-filetypes*
------------------------------
Many times SQL can be used with different filetypes. For example Perl, Java,
PHP, Javascript can all interact with a database. Often you need both the SQL
completion as well as the completion capabilities for the current language you
are editing.
This can be enabled easily with the following steps (assuming a Perl file): >
1. :e test.pl
2. :set filetype=sql
3. :set ft=perl
Step 1
------
Begins by editing a Perl file. Vim automatically sets the filetype to
"perl". By default, Vim runs the appropriate filetype file
ftplugin/perl.vim. If you are using the syntax completion plugin by following
the directions at |ft-syntax-omni| then the |'omnifunc'| option has been set to
"syntax#Complete". Pressing <C-X><C-O> will display the omni popup containing
the syntax items for Perl.
Step 2
------
Manually setting the filetype to 'sql' will also fire the appropriate filetype
files ftplugin/sql.vim. This file will define a number of buffer specific
maps for SQL completion, see |sql-completion-maps|. Now these maps have
been created and the SQL completion plugin has been initialized. All SQL
syntax items have been cached in preparation. The SQL filetype script detects
we are attempting to use two different completion plugins. Since the SQL maps
begin with <C-C>, the maps will toggle the |'omnifunc'| when in use. So you
can use <C-X><C-O> to continue using the completion for Perl (using the syntax
completion plugin) and <C-C> to use the SQL completion features.
Step 3
------
Setting the filetype back to Perl sets all the usual "perl" related items back
as they were.
vim:tw=78:ts=8:ft=help:norl: