Fixed groupings for LLADD API docs.

This commit is contained in:
Sears Russell 2004-07-06 20:59:36 +00:00
parent 540360338d
commit 370ad33210
8 changed files with 48 additions and 9 deletions

View file

@ -44,6 +44,8 @@ terms specified in this license.
* *
* A standard header file, adopted from Autobook. * A standard header file, adopted from Autobook.
* *
* http://sources.redhat.com/autobook/
*
* The idea behind this file is twofold. First, we want to keep as * The idea behind this file is twofold. First, we want to keep as
* much of the #ifdef portability nonsense in here as possible. * much of the #ifdef portability nonsense in here as possible.
* Second, we allow users to #include headers that in turn #include * Second, we allow users to #include headers that in turn #include
@ -51,10 +53,8 @@ terms specified in this license.
* the right thing' and build, even though they do not #include the * the right thing' and build, even though they do not #include the
* config.h file that all of the LLADD stuff uses. * config.h file that all of the LLADD stuff uses.
* *
* @todo: Need to make sure everyone actually includes this thing, and * @todo Need to make sure every .c file actually includes this thing, and
* also includes constants.h * also includes constants.h, and that no .h files include config.h
*
* @ingroup LLADD_CORE
* *
* $Id$ * $Id$
*/ */

View file

@ -55,8 +55,13 @@ BEGIN_C_DECLS
@todo Was getting some memory over-runs from the fact that I didn't @todo Was getting some memory over-runs from the fact that I didn't
know the exact length of a raw log entry. This seems to be fixed know the exact length of a raw log entry. This seems to be fixed
now. (Need to run through electric fence / determine the 'right now.
way' to do this...)
@todo Is there a better way to deal with sizeof() and log entries?
@ingroup LLADD_CORE
$Id$
*/ */
typedef struct { typedef struct {

View file

@ -3,9 +3,27 @@
#ifndef __ALLOC_H #ifndef __ALLOC_H
#define __ALLOC_H #define __ALLOC_H
/**
@file
@ingroup OPERATIONS
$Id$
*/
Operation getAlloc(); Operation getAlloc();
Operation getDealloc(); Operation getDealloc();
/**
Allocate a record.
@param The transaction responsible for the allocation @param The
size of the new record to be allocated. (Talloc may allocate a
blob if the record will not easily fit on a page.)
*/
recordid Talloc(int xid, long size); recordid Talloc(int xid, long size);
/** @todo Currently, we just leak store space on dealloc. */
void Tdealloc(int xid, recordid rid); void Tdealloc(int xid, recordid rid);
#endif #endif

View file

@ -50,6 +50,8 @@ terms specified in this license.
* *
* @see decrement.h * @see decrement.h
* *
* @ingroup OPERATIONS
*
* $Id$ * $Id$
*/ */

View file

@ -20,6 +20,8 @@
result in poor performance when the bucket size is large, and result in poor performance when the bucket size is large, and
transactions are short. transactions are short.
@ingroup OPERATIONS
$Id$ $Id$
*/ */

View file

@ -43,7 +43,7 @@ terms specified in this license.
/** /**
* @file * @file
* *
* interface for dealing with pages * interface for dealing with slotted pages
* *
* @ingroup LLADD_CORE * @ingroup LLADD_CORE
* $Id$ * $Id$

View file

@ -45,7 +45,13 @@ terms specified in this license.
* *
* The minimal subset of LLADD necessary to implement transactional consistency. * The minimal subset of LLADD necessary to implement transactional consistency.
* *
* This module includes the standard API (excluding operations), the logger, the buffer mananger, and recovery code. * This module includes the standard API (excluding operations), the
* logger, the buffer mananger, and recovery code.
*
* In theory, the other .h files that are installed in /usr/include
* aren't needed for application developers.
*
* @todo Move as much of the stuff in lladd/ to src/lladd/ as possible.
* *
*/ */
/** /**

View file

@ -7,8 +7,14 @@ BEGIN_C_DECLS
/** /**
@file @file
blobManager - Provides blob handling blobManager - Provides blob handling.
Plan for modularity: Exactly one blob manager per storeFile. Plan for modularity: Exactly one blob manager per storeFile.
Alternatively, we could use the bufferManager's record length
field to allow for more than one blob manager per buffer
manager. (Right now, this field is set to a special value for
blobs; thie could be extended for other types of records.)
Blob manager interacts with page manger via page manager's Blob manager interacts with page manger via page manager's
public api. public api.