Boost logo

Boost-Commit :

Subject: [Boost-commit] svn:boost r68193 - sandbox/guild/pool/libs/pool/doc
From: pbristow_at_[hidden]
Date: 2011-01-16 13:54:34


Author: pbristow
Date: 2011-01-16 13:54:31 EST (Sun, 16 Jan 2011)
New Revision: 68193
URL: http://svn.boost.org/trac/boost/changeset/68193

Log:
made diagrams .png from snips of originals (but now don't work in pdf!)

removed /details files from references section.

Still missing links to references etc.

Binary files modified:
   sandbox/guild/pool/libs/pool/doc/pool.pdf
Text files modified:
   sandbox/guild/pool/libs/pool/doc/jamfile.v2 | 8
   sandbox/guild/pool/libs/pool/doc/pool.qbk | 229 +++------------------------------------
   2 files changed, 27 insertions(+), 210 deletions(-)

Modified: sandbox/guild/pool/libs/pool/doc/jamfile.v2
==============================================================================
--- sandbox/guild/pool/libs/pool/doc/jamfile.v2 (original)
+++ sandbox/guild/pool/libs/pool/doc/jamfile.v2 2011-01-16 13:54:31 EST (Sun, 16 Jan 2011)
@@ -20,9 +20,9 @@
 doxygen autodoc
    :
       [ glob ../../../boost/pool/*.hpp ]
- [ glob ../../../boost/pool/detail/*.hpp ]
- #[ glob ../../../boost/pool/detail/*.inc ] # These might better be named .ipp?
- # The .inc are not recognised correctly by doxywizard leaving error message.
+ #[ glob ../../../boost/pool/detail/*.hpp ]
+ #[ glob ../../../boost/pool/detail/*.ipp ]
+ # Renamed as .inc are not recognised correctly by doxywizard leaving error message.
       # Warning: include file boost/pool/detail/pool_construct.inc not found, perhaps you forgot to add its directory to INCLUDE_PATH?
       # and problem with 'no type' in jamfile.
       # error: target { ../../../boost/pool/detail/pool_construct.inc. } has no type
@@ -40,6 +40,8 @@
                         <doxygen:param>EXTRACT_PRIVATE=NO
                         <doxygen:param>EXPAND_ONLY_PREDEF=YES
                         <xsl:param>"boost.doxygen.reftitle=Boost.Pool Reference"
+
+
    ;
 
 xml pool : pool.qbk ;

Modified: sandbox/guild/pool/libs/pool/doc/pool.pdf
==============================================================================
Binary files. No diff available.

Modified: sandbox/guild/pool/libs/pool/doc/pool.qbk
==============================================================================
--- sandbox/guild/pool/libs/pool/doc/pool.qbk (original)
+++ sandbox/guild/pool/libs/pool/doc/pool.qbk 2011-01-16 13:54:31 EST (Sun, 16 Jan 2011)
@@ -1,6 +1,6 @@
 [/
  / Copyright (c) 2000 - 2006 Stephen Cleary
- / Copyright (c) 2011 Paul A. Bristow
+ / Copyright (c) 2011 Paul A. Bristow (conversion to Quickbook format)
  / Distributed under the Boost Software License, Version 1.0.
  / (See accompanying file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
  /]
@@ -34,10 +34,10 @@
 [def __requires [*Requires:]]
 
 [def __pool_interfaces [link boost_pool.pool.interfaces Pool Interfaces]]
+[def __pool_references [link boost_pool.pool.appendices.references References]]
 
 [def __todo [link boost_pool.appendices.todo TODO]]
 
-
 [template mu[]'''&#x3BC;'''] [/ µ Greek small letter mu]
 [template plusminus[]'''&#x00B1;'''] [/ ? plus or minus sign]
 
@@ -255,26 +255,18 @@
 
 [section:testing Building the Test Programs]
 
-The subdirectory ['build] contains subdirectories for several different platforms.
-These subdirectories contain all necessary work-around code for that platform,
-as well as makefiles or IDE project files as appropriate.
-
-Read the `readme.txt` in the proper subdirectory, if it exists.
-
-The standard makefile targets are ['all], ['clean] (which deletes any intermediate files),
-and ['veryclean] (which deletes any intermediate files and executables).
-All intermediate and executable files are built in the same directory as the makefile/project file.
-If there is a project file supplied instead of a makefile,
-['clean] and ['veryclean] shell scripts/batch files will be provided.
+A jamfile.v2 is provided which can be run is the usual way, for example:
+
+[@boost:\libs\pool\test> bjam -a >pool_test.log]
 
 [endsect] [/section:testing Building the Test Programs]
 
-[section:pooling Pool in more depth]
+[section:pooling Pool in More Depth]
 
-[section:concepts - Basic ideas behind pooling]
+[section:concepts Basic ideas behind pooling]
 
-[: ['Dynamic memory allocation has been a fundamental part
-of most computer systems since roughly 1960...] [@#ref1 1] ]
+['Dynamic memory allocation has been a fundamental part
+of most computer systems since roughly 1960...] __pool_references 1
 
 Everyone uses dynamic memory allocation.
 If you have ever called malloc or new, then you have used dynamic memory allocation.
@@ -306,72 +298,9 @@
 must ['save] some information in it - usually just its size.
 Then, when the block is free'd, the memory manager can easily tell how large it is.
 
-[br]
-[section:mbna Memory block, allocated (not by process)]
-
-[/<caption> <span class="emphasis"><em>Memory block, not allocated</em></span> </caption>]
+[$/images/pc1.png]
 
-'''
-<table cellspacing="0" border="3" rules="none" style="float: left; clear: both;" summary="Memory not used by process">
- <tr>
- <td style="background-color: white; text-align: center;">Memory block, not allocated</td>
- </tr>
-
- <tr>
- <td style="background-color: red; text-align: center;">Memory not belonging to process</td>
- </tr>
-
- <tr>
- <td style=
- "padding: 1 0; background-color: silver; text-align: center;">
- Memory used internally by memory allocator algorithm (usually 8-12 bytes)</td>
- </tr>
-
- <tr>
- <td style=
- "padding: 2 0; background-color: gray; text-align: center">Unused memory</td>
- </tr>
-
- <tr>
- <td style="background-color: red; text-align: center;">Memory not belonging to process</td>
- </tr>
- </table>
-'''
-[endsect]
-[br]
-[/h4 Memory block, allocated (used by process)]
-[br]
-[section:mba Memory block, allocated (in use by process)]
-
-[/<caption> <span class="emphasis"><em>Memory block, allocated (used by program)</em></span> </caption>]
-
-'''
- <table cellspacing="0" border="3" rules="none" style= "float: right; clear: both;" summary="Memory used by program">
- <tr>
- <td style="background-color: white; text-align: center;">Memory block, allocated (used by process)</td>
- </tr>
-
- <tr>
- <td style="background-color: red; text-align: center;">Memory not belonging to process</td>
- </tr>
-
- <tr>
- <td style="background-color: silver; text-align: center;">Memory used internally by memory allocator algorithm (usually 4 bytes)</td>
- </tr>
-
- <tr>
- <td style=
- "padding: 3 0; background-color: yellow; text-align: center">Memory usable by program</td>
- </tr>
-
- <tr>
- <td style="background-color: red; text-align: center;">Memory not belonging to process</td>
- </tr>
- </table>
-'''
-[br]
-[br]
-[endsect]
+[$/images/pc2.png]
 
 [h5 Dynamic memory allocation is often inefficient]
 
@@ -405,7 +334,7 @@
 However, this is done at the loss of generality;
 simple segregated storage only can allocate memory chunks of a single size.
 
-[endsect] [/section:concepts - Basic ideas behind pooling]
+[endsect] [/section:concepts Basic ideas behind pooling]
 
 [section:simple Simple Segregated Storage]
 
@@ -416,14 +345,8 @@
 Where the block comes from is not important until implementation time.
 A Pool is some object that uses Simple Segregated Storage in this fashion.
 To illustrate:
-[table:mb1 Memory block, split into chunks
-[ [Memory not belonging to process] ]
-[ [Chunk 0] ]
-[ [Chunk 1] ]
-[ [Chunk 2] ]
-[ [Chunk 3] ]
-[ [Memory not belonging to process] ]
-]
+
+[$/images/pc3.png]
 
 Each of the chunks in any given block are always the same size.
 This is the fundamental restriction of Simple Segregated Storage:
@@ -435,22 +358,7 @@
 Simple Segregated Storage works by interleaving a free list within the unused chunks.
 For example:
 
-[table:mb2 Memory block, with no chunks allocated
-[ [Memory not belonging to process] ]
-[ [Chunk 0; points to Chunk 1] ]
-[ [Chunk 1; points to Chunk 2] ]
-[ [Chunk 2; points to Chunk 3] ]
-[ [Chunk 3; end-of-list] ]
-[ [Memory not belonging to process] ]
-]
-[table:mb3 Memory block, with two chunks allocated
-[ [Memory not belonging to process] ]
-[ [Chunk 0; points to Chunk 2] ]
-[ [Chunk 1 (in use by process)] ]
-[ [Chunk 2; end-of-list] ]
-[ [Chunk 3 (in use by process)] ]
-[ [Memory not belonging to process] ]
-]
+[$/images/pc4.png]
 
 By interleaving the free list inside the chunks,
 each Simple Segregated Storage only has the overhead of a single pointer
@@ -464,6 +372,8 @@
 Memory deallocation may be as simple as adding that chunk to the front of the free list, a O(1) operation.
 However, more complicated uses of Simple Segregated Storage may require a sorted free list, which makes deallocation O(N).
 
+[$/images/pc5.png]
+
 Simple Segregated Storage gives faster execution and less memory overhead
 than a system-supplied allocator, but at the loss of generality.
 A good place to use a Pool is in situations
@@ -607,94 +517,20 @@
 
 Here's an example memory block, where `requested_size == sizeof(void *) == sizeof(size_type) == 4`:
 
-[table:mb41 Memory block containing 4 chunks, showing overlying array structures; FLP = Interleaved Free List Pointer
-[ [Sections] [size_type alignment] [void * alignment] [requested_size alignment] ]
-[ [ Memory not belonging to process] ]
-[ [ Chunks section (16 bytes)] [(4 bytes)] [FLP for Chunk 1 (4 bytes)] [Chunk 1 (4 bytes)] ]
-[ [(4 bytes)] [FLP for Chunk 2 (4 bytes)] [Chunk 2 (4 bytes)] ]
-[ [(4 bytes)] [FLP for Chunk 3 (4 bytes)] [Chunk 3 (4 bytes)] ]
-[ [(4 bytes)] [FLP for Chunk 4 (4 bytes)] [Chunk 4 (4 bytes)] ]
-[ [Pointer to next Block (4 bytes)] [(4 bytes)] [Pointer to next Block (4 bytes)] ]
-[ [Size of next Block (4 bytes)] [Size of next Block (4 bytes)] ]
-[ [ Memory not belonging to process] ]
-]
+[$/images/mb1.png]
 
 To show a visual example of possible padding,
 here's an example memory block where
 requested_size == 8 and sizeof(void *) == sizeof(size_type) == 4:
 
-[table:mb4r Memory block containing 4 chunks, showing overlying array structures; FLP = Interleaved Free List Pointer
-[ [Sections] [size_type alignment] [void * alignment] [requested_size alignment] ]
-[ [ Memory not belonging to process] ]
-[ [ Chunks section (32 bytes)] [(4 bytes)] [FLP for Chunk 1 (4 bytes)] [ Chunk 1 (8 bytes)] ]
-[ [(4 bytes)] [(4 bytes)] ]
-[ [(4 bytes)] [FLP for Chunk 2 (4 bytes)] [ Chunk 2 (8 bytes)] ]
-[ [(4 bytes)] [(4 bytes)] ]
-[ [(4 bytes)] [FLP for Chunk 3 (4 bytes)] [ Chunk 3 (8 bytes)] ]
-[ [(4 bytes)] [(4 bytes)] ]
-[ [(4 bytes)] [FLP for Chunk 4 (4 bytes)] [ Chunk 4 (8 bytes)] ]
-[ [(4 bytes)] [(4 bytes)] ]
-[ [Pointer to next Block (4 bytes)] [(4 bytes)] [Pointer to next Block (4 bytes)] ]
-[ [Size of next Block (4 bytes)] [Size of next Block (4 bytes)] ]
-[ [ Memory not belonging to process] ]
-]
+[$/images/mb2.png]
 
 Finally, here is a convoluted example where the requested_size is 7,
 `sizeof(void *) == 3, and sizeof(size_type) == 5`,
 showing how the least common multiple guarantees alignment requirements
 even in the oddest of circumstances:
 
-[table:mb22 Memory block containing 2 chunks, showing overlying array structures
-[ [Sections] [size_type alignment] [void * alignment] [requested_size alignment] ]
-[ [ Memory not belonging to process ] ]
-[ [ Chunks section (210 bytes)] [(5 bytes)] [ Interleaved free list pointer for Chunk 1 (15 bytes; 3 used)] [ Chunk 1 (105 bytes; 7 used)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ (15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [(15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ (15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [(15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ (15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [(15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ Interleaved free list pointer for Chunk 2 (15 bytes; 3 used)] [ Chunk 2 (105 bytes; 7 used)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [(15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ (15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [(15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ (15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [(15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes)] [ (15 bytes)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes) ] ]
-[ [ Pointer to next Block (15 bytes; 3 used)] [(5 bytes)] [ Pointer to next Block (15 bytes; 3 used)] ]
-[ [(5 bytes)] ]
-[ [(5 bytes) ] ]
-[ [Size of next Block (5 bytes; 5 used)] [Size of next Block (5 bytes; 5 used)] ]
-[ [ Memory not belonging to process] ]
-]
+[$/images/mb3.png]
 
 [section:chunks How Contiguous Chunks are Handled]
 
@@ -714,29 +550,8 @@
 Here's an example memory block,
 where `requested_size == 1` and `sizeof(void *) == sizeof(size_type) == 4`:
 
-[table: mb4c1 Memory block containing 4 chunks; requested_size is 1
-[ [Sections] [size_type alignment] [void * alignment] [requested_size alignment] ]
-[ [ Memory not belonging to process] ]
-[ [ Chunks section (16 bytes)] [(4 bytes)] [FLP to Chunk 2 (4 bytes)] [Chunk 1 (4 bytes)] ]
-[ [(4 bytes)] [FLP to Chunk 3 (4 bytes)] [Chunk 2 (4 bytes)] ]
-[ [(4 bytes)] [FLP to Chunk 4 (4 bytes)] [Chunk 3 (4 bytes)] ]
-[ [(4 bytes)] [FLP to end-of-list (4 bytes)] [Chunk 4 (4 bytes)] ]
-[ [Pointer to next Block (4 bytes)] [(4 bytes)] [Ptr to end-of-list (4 bytes)] ]
-[ [Size of next Block (4 bytes)] [0 (4 bytes)] ]
-[ [ Memory not belonging to process] ]
-]
-
-[table:mb7 After user requests 7 contiguous elements of requested_size
-[ [Sections] [size_type alignment] [void * alignment] [requested_size alignment] ]
-[ [ Memory not belonging to process] ]
-[ [ Chunks section (16 bytes)] [(4 bytes)] [(4 bytes)] [4 bytes in use by program] ]
-[ [(4 bytes)] [(4 bytes)] [3 bytes in use by program (1 byte unused)] ]
-[ [(4 bytes)] [FLP to Chunk 4 (4 bytes)] [Chunk 3 (4 bytes)] ]
-[ [(4 bytes)] [FLP to end-of-list (4 bytes)] [Chunk 4 (4 bytes)] ]
-[ [Pointer to next Block (4 bytes)] [(4 bytes)] [Ptr to end-of-list (4 bytes)] ]
-[ [Size of next Block (4 bytes)] [0 (4 bytes)] ]
-[ [ Memory not belonging to process] ]
-]
+
+[$/images/mb4.png]
 
 Then, when the user deallocates the contiguous memory,
 we can split it up into chunks again.


Boost-Commit list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk