From: bct Date: Sun, 23 Jan 2022 23:18:21 +0530 Subject: Patch to build with pcre2 --- README | 20 ++++++++++++++++++++ configure.ac | 25 ++++++++++++++++++++----- src/pull_by_re.c | 47 +++++++++++++++++++++++++++++++++++++---------- src/pull_by_re.h | 4 +--- src/pullseq.c | 44 +++----------------------------------------- src/search_header.c | 21 ++++++++++++++------- src/search_header.h | 6 ++++-- 7 files changed, 99 insertions(+), 68 deletions(-) diff --git a/README b/README index b720bae..64fa546 100644 --- a/README +++ b/README @@ -100,3 +100,23 @@ INSTALL: ./configure # configure the application based on your system make # will build the application make install # will install in /usr/local by default + + NOTE: If you have PCRE (perl-compatible regular expression library) + installed in a non-standard location (e.g. on a mac using brew), the + ./configure script will fail. You'll need to update your CFLAGS and + LDFLAGS env settings to define where your PCRE library files were + installed. + + For example, on a mac with pcre installed by brew, you can do this: + + pcre-config --cflags + -I/usr/local/Cellar/pcre/8.39/include + + Then you can just add this to a env CFLAGS variable and run the + configure command, like so... + + export CFLAGS="-I/usr/local/Cellar/pcre/8.39/include" + ./configure + + If your pcre library is installed somewhere else, you just update + the CFLAGS env variable accordingly. diff --git a/configure.ac b/configure.ac index b8a0e95..2cc154f 100644 --- a/configure.ac +++ b/configure.ac @@ -2,7 +2,7 @@ # Process this file with autoconf to produce a configure script. AC_PREREQ([2.69]) -AC_INIT([pullseq], [1.0.1], [bct.x42@gmail.com]) +AC_INIT([pullseq], [1.1.0], [bct.x42@gmail.com]) #AC_CONFIG_AUX_DIR(config) AM_INIT_AUTOMAKE([foreign]) AC_CONFIG_SRCDIR([src/pullseq.c]) @@ -13,11 +13,27 @@ AC_PROG_CC # Checks for libraries. AC_CHECK_LIB([z],[gzopen]) -AC_CHECK_LIB([pcre],[pcre_compile]) +AC_CHECK_LIB([pcre2-8],[pcre2_config_8], [], [ + echo " Error! You need to have libpcre2 installed." + echo " If you have PCRE2 installed in a non-standard" + echo " location (e.g. as with brew on a mac), you'll" + echo " need to update the CFLAGS/LDFLAGS env" + echo " variables. See the README for more details." + exit -1 + ]) # Checks for header files. -AC_CHECK_HEADERS([inttypes.h limits.h stddef.h stdlib.h string.h]) -AC_CHECK_HEADERS([pcre.h]) +AC_CHECK_HEADERS([pcre2.h inttypes.h limits.h stddef.h stdlib.h string.h]) + +# this no longer works, but the above does... will debug later +#AC_CHECK_HEADERS([pcre2.h], [], [ +# echo " Error! You need to have libpcre2 headers installed." +# echo " If you have PCRE2 installed in a non-standard" +# echo " location (e.g. as with brew on a mac), you'll" +# echo " need to update the CFLAGS/LDFLAGS env" +# echo " variables. See the README for more details." +# exit -1 +# ]) # Checks for typedefs, structures, and compiler characteristics. AC_C_INLINE @@ -37,4 +53,3 @@ AC_CHECK_FUNCS([strlen memset strchr strerror strtol]) AC_CONFIG_FILES([Makefile src/Makefile]) AC_OUTPUT - diff --git a/src/pull_by_re.c b/src/pull_by_re.c index 1da147c..9632797 100644 --- a/src/pull_by_re.c +++ b/src/pull_by_re.c @@ -1,9 +1,11 @@ +#define PCRE2_CODE_UNIT_WIDTH 8 + #include #include #include #include #include -#include +#include #include "pull_by_re.h" #include "file_read.h" @@ -18,13 +20,18 @@ __KSEQ_READ(static) extern char const *progname; extern int verbose_flag; -int pull_by_re(char *input_file, pcre *re, pcre_extra *re_extra, int min, int max, int length, int exclude, int convert, int just_count) { +int pull_by_re(char *input_file, char *aStrRegex, int min, int max, int length, int exclude, int convert, int just_count) { gzFile fp; int count=0,l; int excluded = 0; int is_fasta = 0; /* assume fastq */ kseq_t *seq; + /* pcre2 variables */ + pcre2_code *re; // the regex object + PCRE2_SIZE erroroffset; + int errornumber; + /* open fasta file */ fp = gzopen(input_file,"r"); if (!fp) { @@ -40,30 +47,50 @@ int pull_by_re(char *input_file, pcre *re, pcre_extra *re_extra, int min, int ma gzrewind(fp); kseq_rewind(seq); /* rewind to beginning for main loop */ - if (verbose_flag) { - if (is_fasta) - fprintf(stderr, "Input is FASTA format\n"); - else - fprintf(stderr, "Input is FASTQ format\n"); - } + if (verbose_flag) { + if (is_fasta) + fprintf(stderr, "Input is FASTA format\n"); + else + fprintf(stderr, "Input is FASTQ format\n"); + } + + /* initialize the re */ + re = pcre2_compile( + aStrRegex, /* the pattern */ + PCRE2_ZERO_TERMINATED, /* indicates pattern is zero-terminated */ + 0, /* default options */ + &errornumber, /* for error num */ + &erroroffset, /* err offset */ + NULL); /* default compile context */ + + if (re == NULL) { + PCRE2_UCHAR buffer[256]; + pcre2_get_error_message(errornumber, buffer, sizeof(buffer)); + fprintf(stderr, "PCRE2 compilation failed at offset %d: %s\n", (int)erroroffset, buffer); + exit(EXIT_FAILURE); + } /* search through list and see if this header matches */ while((l = kseq_read(seq)) >= 0) { if (exclude) { - if (search_header(re, re_extra, seq->name.s) || search_header(re, re_extra, seq->comment.s)) + if (search_header(re, seq->name.s) || search_header(re, seq->comment.s)) excluded++; else { /* regex doesn't match, so check size/print */ count += size_filter(seq, is_fasta, min, max, length, convert, just_count); } } else { - if (search_header(re, re_extra, seq->name.s) || search_header(re, re_extra, seq->comment.s)) { + if (search_header(re, seq->name.s) || search_header(re, seq->comment.s)) { /* regex matches so check size/print */ count += size_filter(seq, is_fasta, min, max, length, convert, just_count); } else excluded++; } } /* end of seq traversal */ + + /* tear down re */ + pcre2_code_free(re); /* free up the re */ + kseq_destroy(seq); gzclose(fp); /* done reading file so close */ diff --git a/src/pull_by_re.h b/src/pull_by_re.h index bd86f8c..845c362 100644 --- a/src/pull_by_re.h +++ b/src/pull_by_re.h @@ -1,8 +1,6 @@ #ifndef PULL_BY_SIZE_H #define PULL_BY_SIZE_H -#include - -int pull_by_re(char *input_file, pcre *re, pcre_extra *re_extra, int min, int max, int length, int exclude, int convert, int just_count); +int pull_by_re(char *input_file, char *aStrRegex, int min, int max, int length, int exclude, int convert, int just_count); #endif diff --git a/src/pullseq.c b/src/pullseq.c index 505ac32..870c417 100644 --- a/src/pullseq.c +++ b/src/pullseq.c @@ -4,7 +4,6 @@ #include #include #include -#include #include "global.h" #include "pullseq.h" @@ -59,10 +58,6 @@ int main(int argc, char *argv[]) { int length = 50; long value; char *end; - pcre *reCompiled = NULL; - pcre_extra *pcreExtra; - const char *pcreErrorStr; - int pcreErrorOffset; char *aStrRegex = NULL; extern char *optarg; /* external from getopt */ @@ -265,32 +260,6 @@ int main(int argc, char *argv[]) { } } - /* if a regex search was requested, set up the pcre engine */ - if (aStrRegex) { - // First, the regex string must be compiled. Support extended - // regex strings (e.g. m//x) - reCompiled = pcre_compile(aStrRegex, PCRE_CASELESS, &pcreErrorStr, &pcreErrorOffset, NULL); - - if (reCompiled == NULL) { - fprintf(stderr, "ERROR: Could not compile '%s': %s\n", aStrRegex, pcreErrorStr); - exit(EXIT_FAILURE); - } else { - if (verbose_flag) - fprintf(stderr, "Successfully compiled regex from string '%s'\n", aStrRegex); - } - - // Optimize the regex - pcreExtra = pcre_study(reCompiled, 0, &pcreErrorStr); - - /* pcre_study() returns NULL for both errors and when it can not optimize the regex. - The last argument is how one checks for errors (it is NULL if everything works, - and points to an error string otherwise. */ - if(pcreErrorStr != NULL) { - printf("ERROR: Could not study regex string '%s': %s\n", aStrRegex, pcreErrorStr); - exit(EXIT_FAILURE); - } - } - if (names || names_from_stdin) { if (names) { names_fp = fopen(names,"r"); @@ -302,8 +271,8 @@ int main(int argc, char *argv[]) { names_fp = stdin; } count = pull_by_name(in, names_fp, min, max, length, exclude, convert, just_count); - } else if (reCompiled) { - count = pull_by_re(in, reCompiled, pcreExtra, min, max, length, exclude, convert, just_count); + } else if (aStrRegex) { + count = pull_by_re(in, aStrRegex, min, max, length, exclude, convert, just_count); } else { count = pull_by_size(in, min, max, length, convert, just_count); } @@ -317,15 +286,8 @@ int main(int argc, char *argv[]) { if (names_fp) fclose(names_fp); - if (aStrRegex) { + if (aStrRegex) free(aStrRegex); - // Free up the regular expression. - pcre_free(reCompiled); - - // Free up the EXTRA PCRE value (may be NULL at this point) - if (pcreExtra != NULL) - pcre_free(pcreExtra); - } if (verbose_flag) fprintf(stderr,"Pulled %i entries\n",count); diff --git a/src/search_header.c b/src/search_header.c index bf1e27c..29298b1 100644 --- a/src/search_header.c +++ b/src/search_header.c @@ -1,29 +1,36 @@ +#define PCRE2_CODE_UNIT_WIDTH 8 + #include #include -#include +#include #include "search_header.h" #include "global.h" -int search_header(pcre *re, pcre_extra *re_extra, char *str) { +/* re is a compiled pcre2 regex */ +int search_header(pcre2_code *re, char *str) { int pcreExecRet; - int pcre_ovector[MAX_CAPTURE_COUNT]; + pcre2_match_data *match_data; if (str == NULL) { return 0; } + match_data = pcre2_match_data_create_from_pattern(re, NULL); // init structure for result - pcreExecRet = pcre_exec(re, re_extra, + /* run the match */ + pcreExecRet = pcre2_match(re, str, strlen(str), // length of header string 0, // Start looking at this point 0, // pcre exec OPTIONS - pcre_ovector, // capture groups vector - MAX_CAPTURE_COUNT); // Length of output capture groups + match_data, // pcre2_match_data + NULL); // default match context + + pcre2_match_data_free(match_data); /* Release memory used for the match */ if (pcreExecRet < 0) { - if (pcreExecRet == PCRE_ERROR_NOMATCH) + if (pcreExecRet == PCRE2_ERROR_NOMATCH) return 0; else fprintf(stderr, "Problem with your regex (%d)\n", pcreExecRet); diff --git a/src/search_header.h b/src/search_header.h index b8d6b05..23f36e1 100644 --- a/src/search_header.h +++ b/src/search_header.h @@ -1,10 +1,12 @@ #ifndef SEARCH_HEADER_H #define SEARCH_HEADER_H -#include +#define PCRE2_CODE_UNIT_WIDTH 8 + +#include #define MAX_CAPTURE_COUNT 30 -int search_header(pcre *re, pcre_extra *re_extra, char *str); +int search_header(pcre2_code *re, char *str); #endif