Return to pager-invariants.txt CVS log

Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / sqlite3 / doc

sqlite3

    1:  *** Throughout this document, a page is deemed to have been synced
    2:      automatically as soon as it is written when PRAGMA synchronous=OFF.
    3:      Otherwise, the page is not synced until the xSync method of the VFS
    4:      is called successfully on the file containing the page.
    5: 
    6:  *** Definition:  A page of the database file is said to be "overwriteable" if
    7:      one or more of the following are true about the page:
    8:  
    9:      (a)  The original content of the page as it was at the beginning of
   10:           the transaction has been written into the rollback journal and
   11:           synced.
   12:  
   13:      (b)  The page was a freelist leaf page at the start of the transaction.
   14:  
   15:      (c)  The page number is greater than the largest page that existed in
   16:           the database file at the start of the transaction.
   17:  
   18:  (1) A page of the database file is never overwritten unless one of the
   19:      following are true:
   20:  
   21:      (a) The page and all other pages on the same sector are overwriteable.
   22:  
   23:      (b) The atomic page write optimization is enabled, and the entire
   24:          transaction other than the update of the transaction sequence
   25:          number consists of a single page change.
   26:  
   27:  (2) The content of a page written into the rollback journal exactly matches
   28:      both the content in the database when the rollback journal was written
   29:      and the content in the database at the beginning of the current
   30:      transaction.
   31:  
   32:  (3) Writes to the database file are an integer multiple of the page size
   33:      in length and are aligned to a page boundary.
   34:  
   35:  (4) Reads from the database file are either aligned on a page boundary and
   36:      an integer multiple of the page size in length or are taken from the
   37:      first 100 bytes of the database file.
   38:  
   39:  (5) All writes to the database file are synced prior to the rollback journal
   40:      being deleted, truncated, or zeroed.
   41:  
   42:  (6) If a master journal file is used, then all writes to the database file
   43:      are synced prior to the master journal being deleted.
   44:  
   45:  *** Definition: Two databases (or the same database at two points it time)
   46:      are said to be "logically equivalent" if they give the same answer to
   47:      all queries.  Note in particular the the content of freelist leaf
   48:      pages can be changed arbitarily without effecting the logical equivalence
   49:      of the database.
   50:  
   51:  (7) At any time, if any subset, including the empty set and the total set,
   52:      of the unsynced changes to a rollback journal are removed and the 
   53:      journal is rolled back, the resulting database file will be logical
   54:      equivalent to the database file at the beginning of the transaction.
   55:  
   56:  (8) When a transaction is rolled back, the xTruncate method of the VFS
   57:      is called to restore the database file to the same size it was at
   58:      the beginning of the transaction.  (In some VFSes, the xTruncate
   59:      method is a no-op, but that does not change the fact the SQLite will
   60:      invoke it.)
   61:  
   62:  (9) Whenever the database file is modified, at least one bit in the range
   63:      of bytes from 24 through 39 inclusive will be changed prior to releasing
   64:      the EXCLUSIVE lock.
   65: 
   66: (10) The pattern of bits in bytes 24 through 39 shall not repeat in less
   67:      than one billion transactions.
   68: 
   69: (11) A database file is well-formed at the beginning and at the conclusion
   70:      of every transaction.
   71: 
   72: (12) An EXCLUSIVE lock must be held on the database file before making
   73:      any changes to the database file.
   74: 
   75: (13) A SHARED lock must be held on the database file before reading any
   76:      content out of the database file.