/* gzlog.h
 Copyright (C) 2004, 2008, 2012 Mark Adler, all rights reserved
 version 2.2, 14 Aug 2012

 This software is provided 'as-is', without any express or implied
 warranty.  In no event will the author be held liable for any damages
 arising from the use of this software.

 Permission is granted to anyone to use this software for any purpose,
 including commercial applications, and to alter it and redistribute it
 freely, subject to the following restrictions:

 1. The origin of this software must not be misrepresented; you must not
    claim that you wrote the original software. If you use this software
    in a product, an acknowledgment in the product documentation would be
    appreciated but is not required.
 2. Altered source versions must be plainly marked as such, and must not be
    misrepresented as being the original software.
 3. This notice may not be removed or altered from any source distribution.

 Mark Adler    [email protected]
*/

/* Version History:
  1.0  26 Nov 2004  First version
  2.0  25 Apr 2008  Complete redesign for recovery of interrupted operations
                    Interface changed slightly in that now path is a prefix
                    Compression now occurs as needed during gzlog_write()
                    gzlog_write() now always leaves the log file as valid gzip
  2.1   8 Jul 2012  Fix argument checks in gzlog_compress() and gzlog_write()
  2.2  14 Aug 2012  Clean up signed comparisons
*/

/*
  The gzlog object allows writing short messages to a gzipped log file,
  opening the log file locked for small bursts, and then closing it.  The log
  object works by appending stored (uncompressed) data to the gzip file until
  1 MB has been accumulated.  At that time, the stored data is compressed, and
  replaces the uncompressed data in the file.  The log file is truncated to
  its new size at that time.  After each write operation, the log file is a
  valid gzip file that can decompressed to recover what was written.

  The gzlog operations can be interupted at any point due to an application or
  system crash, and the log file will be recovered the next time the log is
  opened with gzlog_open().
*/

#ifndef GZLOG_H
#define GZLOG_H

/* gzlog object type */
typedef void gzlog;

/* Open a gzlog object, creating the log file if it does not exist.  Return
  NULL on error.  Note that gzlog_open() could take a while to complete if it
  has to wait to verify that a lock is stale (possibly for five minutes), or
  if there is significant contention with other instantiations of this object
  when locking the resource.  path is the prefix of the file names created by
  this object.  If path is "foo", then the log file will be "foo.gz", and
  other auxiliary files will be created and destroyed during the process:
  "foo.dict" for a compression dictionary, "foo.temp" for a temporary (next)
  dictionary, "foo.add" for data being added or compressed, "foo.lock" for the
  lock file, and "foo.repairs" to log recovery operations performed due to
  interrupted gzlog operations.  A gzlog_open() followed by a gzlog_close()
  will recover a previously interrupted operation, if any. */
gzlog *gzlog_open(char *path);

/* Write to a gzlog object.  Return zero on success, -1 if there is a file i/o
  error on any of the gzlog files (this should not happen if gzlog_open()
  succeeded, unless the device has run out of space or leftover auxiliary
  files have permissions or ownership that prevent their use), -2 if there is
  a memory allocation failure, or -3 if the log argument is invalid (e.g. if
  it was not created by gzlog_open()).  This function will write data to the
  file uncompressed, until 1 MB has been accumulated, at which time that data
  will be compressed.  The log file will be a valid gzip file upon successful
  return. */
int gzlog_write(gzlog *log, void *data, size_t len);

/* Force compression of any uncompressed data in the log.  This should be used
  sparingly, if at all.  The main application would be when a log file will
  not be appended to again.  If this is used to compress frequently while
  appending, it will both significantly increase the execution time and
  reduce the compression ratio.  The return codes are the same as for
  gzlog_write(). */
int gzlog_compress(gzlog *log);

/* Close a gzlog object.  Return zero on success, -3 if the log argument is
  invalid.  The log object is freed, and so cannot be referenced again. */
int gzlog_close(gzlog *log);

#endif