[BACK]Return to pcrestack.html CVS log [TXT] [DIR] Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / pcre / doc / html

Diff for /embedaddon/pcre/doc/html/pcrestack.html between versions 1.1.1.1 and 1.1.1.2

version 1.1.1.1, 2012/02/21 23:05:52 version 1.1.1.2, 2012/02/21 23:50:25
Line 16  man page, in case the conversion went wrong. Line 16  man page, in case the conversion went wrong.
 PCRE DISCUSSION OF STACK USAGE  PCRE DISCUSSION OF STACK USAGE
 </b><br>  </b><br>
 <P>  <P>
When you call <b>pcre_exec()</b>, it makes use of an internal function calledWhen you call <b>pcre[16]_exec()</b>, it makes use of an internal function
<b>match()</b>. This calls itself recursively at branch points in the pattern,called <b>match()</b>. This calls itself recursively at branch points in the
in order to remember the state of the match so that it can back up and try apattern, in order to remember the state of the match so that it can back up and
different alternative if the first one fails. As matching proceeds deeper andtry a different alternative if the first one fails. As matching proceeds deeper
deeper into the tree of possibilities, the recursion depth increases. Theand deeper into the tree of possibilities, the recursion depth increases. The
 <b>match()</b> function is also called in other circumstances, for example,  <b>match()</b> function is also called in other circumstances, for example,
 whenever a parenthesized sub-pattern is entered, and in certain cases of  whenever a parenthesized sub-pattern is entered, and in certain cases of
 repetition.  repetition.
Line 33  the recursive call would immediately be passed back as Line 33  the recursive call would immediately be passed back as
 current call (a "tail recursion"), the function is just restarted instead.  current call (a "tail recursion"), the function is just restarted instead.
 </P>  </P>
 <P>  <P>
The above comments apply when <b>pcre_exec()</b> is run in its normalThe above comments apply when <b>pcre[16]_exec()</b> is run in its normal
 interpretive manner. If the pattern was studied with the  interpretive manner. If the pattern was studied with the
 PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling was successful, and  PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling was successful, and
the options passed to <b>pcre_exec()</b> were not incompatible, the matchingthe options passed to <b>pcre[16]_exec()</b> were not incompatible, the matching
 process uses the JIT-compiled code instead of the <b>match()</b> function. In  process uses the JIT-compiled code instead of the <b>match()</b> function. In
 this case, the memory requirements are handled entirely differently. See the  this case, the memory requirements are handled entirely differently. See the
 <a href="pcrejit.html"><b>pcrejit</b></a>  <a href="pcrejit.html"><b>pcrejit</b></a>
 documentation for details.  documentation for details.
 </P>  </P>
 <P>  <P>
The <b>pcre_dfa_exec()</b> function operates in an entirely different way, andThe <b>pcre[16]_dfa_exec()</b> function operates in an entirely different way,
uses recursion only when there is a regular expression recursion or subroutineand uses recursion only when there is a regular expression recursion or
call in the pattern. This includes the processing of assertion and "once-only"subroutine call in the pattern. This includes the processing of assertion and
subpatterns, which are handled like subroutine calls. Normally, these are never"once-only" subpatterns, which are handled like subroutine calls. Normally,
very deep, and the limit on the complexity of <b>pcre_dfa_exec()</b> isthese are never very deep, and the limit on the complexity of
controlled by the amount of workspace it is given. However, it is possible to<b>pcre[16]_dfa_exec()</b> is controlled by the amount of workspace it is given.
write patterns with runaway infinite recursions; such patterns will causeHowever, it is possible to write patterns with runaway infinite recursions;
<b>pcre_dfa_exec()</b> to run out of stack. At present, there is no protectionsuch patterns will cause <b>pcre[16]_dfa_exec()</b> to run out of stack. At
against this.present, there is no protection against this.
 </P>  </P>
 <P>  <P>
The comments that follow do NOT apply to <b>pcre_dfa_exec()</b>; they areThe comments that follow do NOT apply to <b>pcre[16]_dfa_exec()</b>; they are
relevant only for <b>pcre_exec()</b> without the JIT optimization.relevant only for <b>pcre[16]_exec()</b> without the JIT optimization.
 </P>  </P>
 <br><b>  <br><b>
Reducing <b>pcre_exec()</b>'s stack usageReducing <b>pcre[16]_exec()</b>'s stack usage
 </b><br>  </b><br>
 <P>  <P>
 Each time that <b>match()</b> is actually called recursively, it uses memory  Each time that <b>match()</b> is actually called recursively, it uses memory
Line 94  subject strings is to write repeated parenthesized sub Line 94  subject strings is to write repeated parenthesized sub
 than one character whenever possible.  than one character whenever possible.
 </P>  </P>
 <br><b>  <br><b>
Compiling PCRE to use heap instead of stack for <b>pcre_exec()</b>Compiling PCRE to use heap instead of stack for <b>pcre[16]_exec()</b>
 </b><br>  </b><br>
 <P>  <P>
 In environments where stack memory is constrained, you might want to compile  In environments where stack memory is constrained, you might want to compile
 PCRE to use heap memory instead of stack for remembering back-up points when  PCRE to use heap memory instead of stack for remembering back-up points when
<b>pcre_exec()</b> is running. This makes it run a lot more slowly, however.<b>pcre[16]_exec()</b> is running. This makes it run a lot more slowly, however.
 Details of how to do this are given in the  Details of how to do this are given in the
 <a href="pcrebuild.html"><b>pcrebuild</b></a>  <a href="pcrebuild.html"><b>pcrebuild</b></a>
 documentation. When built in this way, instead of using the stack, PCRE obtains  documentation. When built in this way, instead of using the stack, PCRE obtains
 and frees memory by calling the functions that are pointed to by the  and frees memory by calling the functions that are pointed to by the
<b>pcre_stack_malloc</b> and <b>pcre_stack_free</b> variables. By default, these<b>pcre[16]_stack_malloc</b> and <b>pcre[16]_stack_free</b> variables. By
point to <b>malloc()</b> and <b>free()</b>, but you can replace the pointers todefault, these point to <b>malloc()</b> and <b>free()</b>, but you can replace
cause PCRE to use your own functions. Since the block sizes are always thethe pointers to cause PCRE to use your own functions. Since the block sizes are
same, and are always freed in reverse order, it may be possible to implementalways the same, and are always freed in reverse order, it may be possible to
customized memory handlers that are more efficient than the standard functions.implement customized memory handlers that are more efficient than the standard
 functions.
 </P>  </P>
 <br><b>  <br><b>
Limiting <b>pcre_exec()</b>'s stack usageLimiting <b>pcre[16]_exec()</b>'s stack usage
 </b><br>  </b><br>
 <P>  <P>
 You can set limits on the number of times that <b>match()</b> is called, both in  You can set limits on the number of times that <b>match()</b> is called, both in
total and recursively. If a limit is exceeded, <b>pcre_exec()</b> returns antotal and recursively. If a limit is exceeded, <b>pcre[16]_exec()</b> returns an
 error code. Setting suitable limits should prevent it from running out of  error code. Setting suitable limits should prevent it from running out of
 stack. The default values of the limits are very large, and unlikely ever to  stack. The default values of the limits are very large, and unlikely ever to
 operate. They can be changed when PCRE is built, and they can also be set when  operate. They can be changed when PCRE is built, and they can also be set when
<b>pcre_exec()</b> is called. For details of these interfaces, see the<b>pcre[16]_exec()</b> is called. For details of these interfaces, see the
 <a href="pcrebuild.html"><b>pcrebuild</b></a>  <a href="pcrebuild.html"><b>pcrebuild</b></a>
 documentation and the  documentation and the
<a href="pcreapi.html#extradata">section on extra data for <b>pcre_exec()</b></a><a href="pcreapi.html#extradata">section on extra data for <b>pcre[16]_exec()</b></a>
 in the  in the
 <a href="pcreapi.html"><b>pcreapi</b></a>  <a href="pcreapi.html"><b>pcreapi</b></a>
 documentation.  documentation.
 </P>  </P>
 <P>  <P>
 As a very rough rule of thumb, you should reckon on about 500 bytes per  As a very rough rule of thumb, you should reckon on about 500 bytes per
recursion. Thus, if you want to limit your stack usage to 8Mb, yourecursion. Thus, if you want to limit your stack usage to 8Mb, you should set
should set the limit at 16000 recursions. A 64Mb stack, on the other hand, canthe limit at 16000 recursions. A 64Mb stack, on the other hand, can support
support around 128000 recursions.around 128000 recursions.
 </P>  </P>
 <P>  <P>
 In Unix-like environments, the <b>pcretest</b> test program has a command line  In Unix-like environments, the <b>pcretest</b> test program has a command line
 option (<b>-S</b>) that can be used to increase the size of its stack. As long  option (<b>-S</b>) that can be used to increase the size of its stack. As long
 as the stack is large enough, another option (<b>-M</b>) can be used to find the  as the stack is large enough, another option (<b>-M</b>) can be used to find the
 smallest limits that allow a particular pattern to match a given subject  smallest limits that allow a particular pattern to match a given subject
string. This is done by calling <b>pcre_exec()</b> repeatedly with differentstring. This is done by calling <b>pcre[16]_exec()</b> repeatedly with different
 limits.  limits.
 </P>  </P>
 <br><b>  <br><b>
   Obtaining an estimate of stack usage
   </b><br>
   <P>
   The actual amount of stack used per recursion can vary quite a lot, depending
   on the compiler that was used to build PCRE and the optimization or debugging
   options that were set for it. The rule of thumb value of 500 bytes mentioned
   above may be larger or smaller than what is actually needed. A better
   approximation can be obtained by running this command:
   <pre>
     pcretest -m -C
   </pre>
   The <b>-C</b> option causes <b>pcretest</b> to output information about the
   options with which PCRE was compiled. When <b>-m</b> is also given (before
   <b>-C</b>), information about stack use is given in a line like this:
   <pre>
     Match recursion uses stack: approximate frame size = 640 bytes
   </pre>
   The value is approximate because some recursions need a bit more (up to perhaps
   16 more bytes).
   </P>
   <P>
   If the above command is given when PCRE is compiled to use the heap instead of
   the stack for recursion, the value that is output is the size of each block
   that is obtained from the heap.
   </P>
   <br><b>
 Changing stack size in Unix-like systems  Changing stack size in Unix-like systems
 </b><br>  </b><br>
 <P>  <P>
Line 163  limit on stack size by code such as this: Line 190  limit on stack size by code such as this:
 </pre>  </pre>
 This reads the current limits (soft and hard) using <b>getrlimit()</b>, then  This reads the current limits (soft and hard) using <b>getrlimit()</b>, then
 attempts to increase the soft limit to 100Mb using <b>setrlimit()</b>. You must  attempts to increase the soft limit to 100Mb using <b>setrlimit()</b>. You must
do this before calling <b>pcre_exec()</b>.do this before calling <b>pcre[16]_exec()</b>.
 </P>  </P>
 <br><b>  <br><b>
 Changing stack size in Mac OS X  Changing stack size in Mac OS X
Line 189  Cambridge CB2 3QH, England. Line 216  Cambridge CB2 3QH, England.
 REVISION  REVISION
 </b><br>  </b><br>
 <P>  <P>
Last updated: 26 August 2011Last updated: 21 January 2012
 <br>  <br>
Copyright &copy; 1997-2011 University of Cambridge.Copyright &copy; 1997-2012 University of Cambridge.
 <br>  <br>
 <p>  <p>
 Return to the <a href="index.html">PCRE index page</a>.  Return to the <a href="index.html">PCRE index page</a>.
Removed from v.1.1.1.1  
changed lines
  Added in v.1.1.1.2

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>