Tristan Matthews | 0461646 | 2013-11-14 16:09:34 -0500 | [diff] [blame] | 1 | .TH PCREPRECOMPILE 3 |
| 2 | .SH NAME |
| 3 | PCRE - Perl-compatible regular expressions |
| 4 | .SH "SAVING AND RE-USING PRECOMPILED PCRE PATTERNS" |
| 5 | .rs |
| 6 | .sp |
| 7 | If you are running an application that uses a large number of regular |
| 8 | expression patterns, it may be useful to store them in a precompiled form |
| 9 | instead of having to compile them every time the application is run. |
| 10 | If you are not using any private character tables (see the |
| 11 | .\" HREF |
| 12 | \fBpcre_maketables()\fP |
| 13 | .\" |
| 14 | documentation), this is relatively straightforward. If you are using private |
| 15 | tables, it is a little bit more complicated. However, if you are using the |
| 16 | just-in-time optimization feature of \fBpcre_study()\fP, it is not possible to |
| 17 | save and reload the JIT data. |
| 18 | .P |
| 19 | If you save compiled patterns to a file, you can copy them to a different host |
| 20 | and run them there. This works even if the new host has the opposite endianness |
| 21 | to the one on which the patterns were compiled. There may be a small |
| 22 | performance penalty, but it should be insignificant. However, compiling regular |
| 23 | expressions with one version of PCRE for use with a different version is not |
| 24 | guaranteed to work and may cause crashes, and saving and restoring a compiled |
| 25 | pattern loses any JIT optimization data. |
| 26 | . |
| 27 | . |
| 28 | .SH "SAVING A COMPILED PATTERN" |
| 29 | .rs |
| 30 | .sp |
| 31 | The value returned by \fBpcre_compile()\fP points to a single block of memory |
| 32 | that holds the compiled pattern and associated data. You can find the length of |
| 33 | this block in bytes by calling \fBpcre_fullinfo()\fP with an argument of |
| 34 | PCRE_INFO_SIZE. You can then save the data in any appropriate manner. Here is |
| 35 | sample code that compiles a pattern and writes it to a file. It assumes that |
| 36 | the variable \fIfd\fP refers to a file that is open for output: |
| 37 | .sp |
| 38 | int erroroffset, rc, size; |
| 39 | char *error; |
| 40 | pcre *re; |
| 41 | .sp |
| 42 | re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL); |
| 43 | if (re == NULL) { ... handle errors ... } |
| 44 | rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size); |
| 45 | if (rc < 0) { ... handle errors ... } |
| 46 | rc = fwrite(re, 1, size, fd); |
| 47 | if (rc != size) { ... handle errors ... } |
| 48 | .sp |
| 49 | In this example, the bytes that comprise the compiled pattern are copied |
| 50 | exactly. Note that this is binary data that may contain any of the 256 possible |
| 51 | byte values. On systems that make a distinction between binary and non-binary |
| 52 | data, be sure that the file is opened for binary output. |
| 53 | .P |
| 54 | If you want to write more than one pattern to a file, you will have to devise a |
| 55 | way of separating them. For binary data, preceding each pattern with its length |
| 56 | is probably the most straightforward approach. Another possibility is to write |
| 57 | out the data in hexadecimal instead of binary, one pattern to a line. |
| 58 | .P |
| 59 | Saving compiled patterns in a file is only one possible way of storing them for |
| 60 | later use. They could equally well be saved in a database, or in the memory of |
| 61 | some daemon process that passes them via sockets to the processes that want |
| 62 | them. |
| 63 | .P |
| 64 | If the pattern has been studied, it is also possible to save the normal study |
| 65 | data in a similar way to the compiled pattern itself. However, if the |
| 66 | PCRE_STUDY_JIT_COMPILE was used, the just-in-time data that is created cannot |
| 67 | be saved because it is too dependent on the current environment. When studying |
| 68 | generates additional information, \fBpcre_study()\fP returns a pointer to a |
| 69 | \fBpcre_extra\fP data block. Its format is defined in the |
| 70 | .\" HTML <a href="pcreapi.html#extradata"> |
| 71 | .\" </a> |
| 72 | section on matching a pattern |
| 73 | .\" |
| 74 | in the |
| 75 | .\" HREF |
| 76 | \fBpcreapi\fP |
| 77 | .\" |
| 78 | documentation. The \fIstudy_data\fP field points to the binary study data, and |
| 79 | this is what you must save (not the \fBpcre_extra\fP block itself). The length |
| 80 | of the study data can be obtained by calling \fBpcre_fullinfo()\fP with an |
| 81 | argument of PCRE_INFO_STUDYSIZE. Remember to check that \fBpcre_study()\fP did |
| 82 | return a non-NULL value before trying to save the study data. |
| 83 | . |
| 84 | . |
| 85 | .SH "RE-USING A PRECOMPILED PATTERN" |
| 86 | .rs |
| 87 | .sp |
| 88 | Re-using a precompiled pattern is straightforward. Having reloaded it into main |
| 89 | memory, you pass its pointer to \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP in |
| 90 | the usual way. This should work even on another host, and even if that host has |
| 91 | the opposite endianness to the one where the pattern was compiled. |
| 92 | .P |
| 93 | However, if you passed a pointer to custom character tables when the pattern |
| 94 | was compiled (the \fItableptr\fP argument of \fBpcre_compile()\fP), you must |
| 95 | now pass a similar pointer to \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP, |
| 96 | because the value saved with the compiled pattern will obviously be nonsense. A |
| 97 | field in a \fBpcre_extra()\fP block is used to pass this data, as described in |
| 98 | the |
| 99 | .\" HTML <a href="pcreapi.html#extradata"> |
| 100 | .\" </a> |
| 101 | section on matching a pattern |
| 102 | .\" |
| 103 | in the |
| 104 | .\" HREF |
| 105 | \fBpcreapi\fP |
| 106 | .\" |
| 107 | documentation. |
| 108 | .P |
| 109 | If you did not provide custom character tables when the pattern was compiled, |
| 110 | the pointer in the compiled pattern is NULL, which causes \fBpcre_exec()\fP to |
| 111 | use PCRE's internal tables. Thus, you do not need to take any special action at |
| 112 | run time in this case. |
| 113 | .P |
| 114 | If you saved study data with the compiled pattern, you need to create your own |
| 115 | \fBpcre_extra\fP data block and set the \fIstudy_data\fP field to point to the |
| 116 | reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA bit in the |
| 117 | \fIflags\fP field to indicate that study data is present. Then pass the |
| 118 | \fBpcre_extra\fP block to \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP in the |
| 119 | usual way. If the pattern was studied for just-in-time optimization, that data |
| 120 | cannot be saved, and so is lost by a save/restore cycle. |
| 121 | . |
| 122 | . |
| 123 | .SH "COMPATIBILITY WITH DIFFERENT PCRE RELEASES" |
| 124 | .rs |
| 125 | .sp |
| 126 | In general, it is safest to recompile all saved patterns when you update to a |
| 127 | new PCRE release, though not all updates actually require this. |
| 128 | . |
| 129 | . |
| 130 | . |
| 131 | .SH AUTHOR |
| 132 | .rs |
| 133 | .sp |
| 134 | .nf |
| 135 | Philip Hazel |
| 136 | University Computing Service |
| 137 | Cambridge CB2 3QH, England. |
| 138 | .fi |
| 139 | . |
| 140 | . |
| 141 | .SH REVISION |
| 142 | .rs |
| 143 | .sp |
| 144 | .nf |
| 145 | Last updated: 26 August 2011 |
| 146 | Copyright (c) 1997-2011 University of Cambridge. |
| 147 | .fi |