Source-Changes-HG archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

[src/trunk]: src/usr.sbin/makemandb Add support for compressed man pages in a...



details:   https://anonhg.NetBSD.org/src/rev/a2f6216044d7
branches:  trunk
changeset: 777372:a2f6216044d7
user:      joerg <joerg%NetBSD.org@localhost>
date:      Thu Feb 16 20:58:55 2012 +0000

description:
Add support for compressed man pages in all the usual formats.

diffstat:

 usr.sbin/makemandb/Makefile    |    6 +-
 usr.sbin/makemandb/makemandb.c |  113 ++++++++++++++++++++++++++++++++--------
 2 files changed, 93 insertions(+), 26 deletions(-)

diffs (262 lines):

diff -r 18c179671087 -r a2f6216044d7 usr.sbin/makemandb/Makefile
--- a/usr.sbin/makemandb/Makefile       Thu Feb 16 20:58:23 2012 +0000
+++ b/usr.sbin/makemandb/Makefile       Thu Feb 16 20:58:55 2012 +0000
@@ -1,4 +1,4 @@
-# $NetBSD: Makefile,v 1.1 2012/02/07 19:13:32 joerg Exp $
+# $NetBSD: Makefile,v 1.2 2012/02/16 20:58:55 joerg Exp $
 
 .include <bsd.own.mk>
 
@@ -22,8 +22,8 @@
 MDOCMLOBJDIR!= cd ${MDOCDIR}/lib/libmandoc && ${PRINTOBJDIR}
 MDOCMLLIB=     ${MDOCMLOBJDIR}/libmandoc.a
 
-DPADD.makemandb+=      ${MDOCMLLIB}
-LDADD.makemandb+=      -L${MDOCMLOBJDIR} -lmandoc
+DPADD.makemandb+=      ${MDOCMLLIB} ${LIBARCHIVE} ${LIBBZ2} ${LIBLZMA}
+LDADD.makemandb+=      -L${MDOCMLOBJDIR} -lmandoc -larchive -lbz2 -llzma
 DPADD+=                ${LIBSQLITE3} ${LIBM} ${LIBZ} ${LIBUTIL}
 LDADD+=                -lsqlite3 -lm -lz -lutil
 
diff -r 18c179671087 -r a2f6216044d7 usr.sbin/makemandb/makemandb.c
--- a/usr.sbin/makemandb/makemandb.c    Thu Feb 16 20:58:23 2012 +0000
+++ b/usr.sbin/makemandb/makemandb.c    Thu Feb 16 20:58:55 2012 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: makemandb.c,v 1.4 2012/02/15 23:36:10 joerg Exp $      */
+/*     $NetBSD: makemandb.c,v 1.5 2012/02/16 20:58:55 joerg Exp $      */
 /*
  * Copyright (c) 2011 Abhinav Upadhyay <er.abhinav.upadhyay%gmail.com@localhost>
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps%bsd.lv@localhost>
@@ -17,7 +17,7 @@
  */
 
 #include <sys/cdefs.h>
-__RCSID("$NetBSD: makemandb.c,v 1.4 2012/02/15 23:36:10 joerg Exp $");
+__RCSID("$NetBSD: makemandb.c,v 1.5 2012/02/16 20:58:55 joerg Exp $");
 
 #include <sys/stat.h>
 #include <sys/types.h>
@@ -26,6 +26,7 @@
 #include <ctype.h>
 #include <dirent.h>
 #include <err.h>
+#include <archive.h>
 #include <libgen.h>
 #include <md5.h>
 #include <stdio.h>
@@ -94,12 +95,13 @@
 static void append(secbuff *sbuff, const char *src);
 static void init_secbuffs(mandb_rec *);
 static void free_secbuffs(mandb_rec *);
-static int check_md5(const char *, sqlite3 *, const char *, char **);
+static int check_md5(const char *, sqlite3 *, const char *, char **, void *, size_t);
 static void cleanup(mandb_rec *);
 static void set_section(const struct mdoc *, const struct man *, mandb_rec *);
 static void set_machine(const struct mdoc *, mandb_rec *);
 static int insert_into_db(sqlite3 *, mandb_rec *);
-static void begin_parse(const char *, struct mparse *, mandb_rec *);
+static void begin_parse(const char *, struct mparse *, mandb_rec *,
+                        const void *, size_t len);
 static void pmdoc_node(const struct mdoc_node *, mandb_rec *);
 static void pmdoc_Nm(const struct mdoc_node *, mandb_rec *);
 static void pmdoc_Nd(const struct mdoc_node *, mandb_rec *);
@@ -591,6 +593,59 @@
        sqlite3_finalize(inner_stmt);
 }
 
+/* read_and_decompress --
+ *     Reads the given file into memory. If it is compressed, decompres
+ *     it before returning to the caller.
+ */
+static int
+read_and_decompress(const char *file, void **buf, size_t *len)
+{
+       size_t off;
+       ssize_t r;
+       struct archive *a;
+       struct archive_entry *ae;
+
+       if ((a = archive_read_new()) == NULL)
+               errx(EXIT_FAILURE, "memory allocation failed");
+
+       if (archive_read_support_compression_all(a) != ARCHIVE_OK ||
+           archive_read_support_format_raw(a) != ARCHIVE_OK ||
+           archive_read_open_filename(a, file, 65536) != ARCHIVE_OK ||
+           archive_read_next_header(a, &ae) != ARCHIVE_OK)
+               goto archive_error;
+       *len = 65536;
+       *buf = emalloc(*len);
+       off = 0;
+       for (;;) {
+               r = archive_read_data(a, (char *)*buf + off, *len - off);
+               if (r == ARCHIVE_OK) {
+                       archive_read_close(a);
+                       *len = off;
+                       return 0;
+               }
+               if (r <= 0) {
+                       free(*buf);
+                       break;
+               }
+               off += r;
+               if (off == *len) {
+                       *len *= 2;
+                       if (*len < off) {
+                               warnx("File too large: %s", file);
+                               free(*buf);
+                               archive_read_close(a);
+                               return -1;
+                       }
+                       *buf = erealloc(*buf, *len);
+               }
+       }
+
+archive_error:
+       warnx("Error while reading `%s': %s", file, archive_error_string(a));
+       archive_read_close(a);
+       return -1;
+}
+
 /* update_db --
  *     Does an incremental updation of the database by checking the file_cache.
  *     It parses and adds the pages which are present in file_cache,
@@ -605,7 +660,9 @@
        sqlite3_stmt *stmt = NULL;
        const char *file;
        char *errmsg = NULL;
-       char *buf = NULL;
+       char *md5sum;
+       void *buf;
+       size_t buflen;
        int new_count = 0;      /* Counter for newly indexed/updated pages */
        int total_count = 0;    /* Counter for total number of pages */
        int err_count = 0;      /* Counter for number of failed pages */
@@ -623,14 +680,21 @@
                errx(EXIT_FAILURE, "Could not query file cache");
        }
 
+       buf = NULL;
        while (sqlite3_step(stmt) == SQLITE_ROW) {
+               free(buf);
                total_count++;
                rec->device = sqlite3_column_int64(stmt, 0);
                rec->inode = sqlite3_column_int64(stmt, 1);
                rec->mtime = sqlite3_column_int64(stmt, 2);
                file = (const char *) sqlite3_column_text(stmt, 3);
-               md5_status = check_md5(file, db, "mandb_meta", &buf);
-               assert(buf != NULL);
+               if (read_and_decompress(file, &buf, &buflen)) {
+                       err_count++;
+                       buf = NULL;
+                       continue;
+               }
+               md5_status = check_md5(file, db, "mandb_meta", &md5sum, buf, buflen);
+               assert(md5sum != NULL);
                if (md5_status == -1) {
                        warnx("An error occurred in checking md5 value"
                              " for file %s", file);
@@ -646,13 +710,13 @@
                        struct stat sb;
                        stat(file, &sb);
                        if (S_ISLNK(sb.st_mode)) {
-                               free(buf);
+                               free(md5sum);
                                link_count++;
                                continue;
                        }
-                       update_existing_entry(db, file, buf, rec,
+                       update_existing_entry(db, file, md5sum, rec,
                            &new_count, &link_count, &err_count);
-                       free(buf);
+                       free(md5sum);
                        continue;
                }
 
@@ -664,10 +728,10 @@
                         */
                        if (mflags.verbosity > 1)
                                printf("Parsing: %s\n", file);
-                       rec->md5_hash = buf;
+                       rec->md5_hash = md5sum;
                        rec->file_path = estrdup(file);
                        // file_path is freed by insert_into_db itself.
-                       begin_parse(file, mp, rec);
+                       begin_parse(file, mp, rec, buf, buflen);
                        if (insert_into_db(db, rec) < 0) {
                                warnx("Error in indexing %s", file);
                                err_count++;
@@ -676,7 +740,8 @@
                        }
                }
        }
-       
+       free(buf);
+
        sqlite3_finalize(stmt);
        
        if (mflags.verbosity) {
@@ -715,7 +780,8 @@
  *  parses the man page using libmandoc
  */
 static void
-begin_parse(const char *file, struct mparse *mp, mandb_rec *rec)
+begin_parse(const char *file, struct mparse *mp, mandb_rec *rec,
+    const void *buf, size_t len)
 {
        struct mdoc *mdoc;
        struct man *man;
@@ -723,7 +789,7 @@
 
        rec->xr_found = 0;
 
-       if (mparse_readfd(mp, -1, file) >= MANDOCLEVEL_FATAL) {
+       if (mparse_readmem(mp, buf, len, file) >= MANDOCLEVEL_FATAL) {
                warnx("%s: Parse failure", file);
                return;
        }
@@ -1679,7 +1745,8 @@
  *  1: If the hash exists in the database.
  */
 static int
-check_md5(const char *file, sqlite3 *db, const char *table, char **buf)
+check_md5(const char *file, sqlite3 *db, const char *table, char **md5sum,
+    void *buf, size_t buflen)
 {
        int rc = 0;
        int idx = -1;
@@ -1687,8 +1754,8 @@
        sqlite3_stmt *stmt = NULL;
 
        assert(file != NULL);
-       *buf = MD5File(file, NULL);
-       if (*buf == NULL) {
+       *md5sum = MD5Data(buf, buflen, NULL);
+       if (*md5sum == NULL) {
                warn("md5 failed: %s", file);
                return -1;
        }
@@ -1698,19 +1765,19 @@
        rc = sqlite3_prepare_v2(db, sqlstr, -1, &stmt, NULL);
        if (rc != SQLITE_OK) {
                free(sqlstr);
-               free(*buf);
-               *buf = NULL;
+               free(*md5sum);
+               *md5sum = NULL;
                return -1;
        }
 
        idx = sqlite3_bind_parameter_index(stmt, ":md5_hash");
-       rc = sqlite3_bind_text(stmt, idx, *buf, -1, NULL);
+       rc = sqlite3_bind_text(stmt, idx, *md5sum, -1, NULL);
        if (rc != SQLITE_OK) {
                warnx("%s", sqlite3_errmsg(db));
                sqlite3_finalize(stmt);
                free(sqlstr);
-               free(*buf);
-               *buf = NULL;
+               free(*md5sum);
+               *md5sum = NULL;
                return -1;
        }
 



Home | Main Index | Thread Index | Old Index