============================================

New in 5.8.02.XX Update Subscription Program

============================================

New SELECTBOX() function.

New Form Filtering for IUA(clerk) Form command.

Ability to define screens for *clerk with "truncated/scrolling" fields.

Added timeout option to user, and related TIMEOUT() test.

PDF Enhancement to <PAGE> for MARGINT|MARGINB|MARGINL|MARGINR=”margin”

============================================

New in 5.8.01.XX Update Subscription Program

============================================

New #1312

    Automatic indexes can now be built using a selection set.  As

    records are created or modified, they will be placed in the

    index only if they pass the selection criteria.

New

    Dual Write or Mirroring.  Extended options in define files (ddefine)

    now has an option to mirror a file.  Use of PFDIR2 and PFDATA2 settings

    determine the path of the mirror directory.

New #1324

    PFPQ=ON   Acts as if the "-PQ" flag was passed to *clerk/*report.

================================

5.8.XX.04

================================

(All) #1054

    Erroneous "feature not licensed" errors correct.

(All) #1313

    filePro user menus can now use a pseudo-environment variable

    "@VR" to place the filePro version on the screen.  You can use

    either Windows ("%@VR%") or Unix/Linux ("$@VR") syntax.

(fileProGI) #1317

    filePro GI client displayed "garbage" for the "Enter" prompt.

(All) #1344

    A spurious "requested feature not licensed" message would

    sometimes be included in filePro error messages.

(Quikstart) #1345

    If you have a "locked(-)" test without a lookup-dash, rcabe

    did not give a syntax error.

(Windows 64-bit) #1347

    Pressing F8/Options in printer maintenance would cash the

    64-bit Windows version of filePro.

(All) #1350

    dxmaint -m "" would not override a PFQUAL setting.

(Linux) #1352

    filePro now turs off stty "iexten" mode on systems which

    support it.

(All) #1353

    FPML now supports user-defined margins to the <PAGE> tag.

(All) #1357

    New PFPDFCOMPRESSMODE config variable to control PDF compression.

(Windows) #1360

    The Windows version of dexpand didn't properly handle key files

    larger than 4G in size.

(Windows 64-bit) #1361

    When concatenating a string with a memo field from a lookup

    file, 64-bit Windows filePro could crash.

================================

5.8.XX.03

================================

(All) #402

    If you select a record via *clerk's "6 - fuzzy search" menu

    choice, and then press Break/Del/Ctrl-C, *clerk exits, rather

    than returning to the menu.

(fPSQL) #1053

    fPSQL did not properly display user edits in the F5 field list.

(fileProGI) #1336

    When running filePro on a Windows server, the "Enter" prompt

    displayed "garbage" characters.

(5.8.01.01 only) #1337

    It was possible that the first lookup to a file would corrupt

    indexes built on field you modified.

(All) #1338

    The CREATE() function did not respect the umask, always using a

    umask of 0177 instead.  Now it will respect PFUMASK if set.

================================

5.8.XX.02

================================

(All) #402

    If you select a record via *clerk's "6 - fuzzy search" menu

    choice, an then press Break/Del/Ctrl-C, *clerk exits, rather

    than returning to the *clerk menu.

(64-bit only) #1332

    Spellcheck didn't work properly on 64-bit implementations.

(GIclient only) #1336

    When running fileProGI against a Windows version of filePro,

    then "Enter" prompt displayed as garbage.

(5.8.01.01 only) #1337

    It was possible that the first lookup to a file could

    corrupt indexes built on field you modified.

================================

5.8.XX.01

================================

(All) # 1332

    Aspell ws replaced with hunspell for the 64 bit version of

    filePro

(Linux) #1330

    The 64-bit Linux version of xfer didn't copy screens correctly.

(Linux) #1327  

    On 64-bit Linux. closing a lookup file could crash with a corrupt heap.  

    ("*** glibc detected ***" error message.)            

(Linux) #1316

    Fuzzy browse lookups could crash on 64 bit Linux

(All) #1286

    Setting PFIDLEN=32 will cause @ID, @CB, and @UB to have a length of

    32 rather than 8.  (The only legal values are currently "8" and "32".

    Any other value is undefined.)

(All) #1295

    Alien and ODBC files can't be encrypted by filePro.  ddefine will

    now prevent you from marking such files as encrypted.

(All) #1302

    filePro now defaults to PFMAXALLOC=128 and PFMAXASIZE=128000, for

    up to 16MB of RAM for sorting.

(All) #1304

    If you have a fatal error (such as "file not found" on import) in

    a CALLed table, filePro might crash upon exit.

(All) #1305

    If you print a form from *clerk via "F"orn (not from FORM within

    processing), and that form's output processing does a lookup to

    the same file as an open lookup in input/auto processing, and

    then CLOSEs that lookup, filePro can crash if the input/auto

    processing re-executes that lookup.

(All) #1314

    The FPML documentation was incorrect.  The "<PAGE>" tag is

    documented as taking an "ORIENTATION" attribute, but it should

    have said "ORIENT".  filePro will now allow either spelling.

(Windows 10) #1318

    On Windows 10, if you set filePro's TextNormal attribute to a

    value from 0x80-0xff, there could be "ghost" characters left

    on the screen.

(Windows, 64-bit) #1321

    The 64-bit Windows version of rclerk/rreport could crash when

    using OPENDIR().

(All) #1322

    If you have a very long filename that "looks like" a qualified

    filename (such as "key [conflict at 2014-12-09_23-28-27]"),

    filePro could crash.

================================

5.8.00.00

================================

(Linux) #1292

        On some Linux systems, filePro would see a different MAC

        address than reported by "ifconfig" or "ip addr".

(All) #1294

        If you edit a prc file outside of filePro, and there is

        no newline character at the end of the last line, filePro

        may crash on exit or "switch files" in *clerk.

(All) #1295

        Alien files can't be encrypted.  Ddefine now prevents you

        from marking such files as encrypted.

(OSR5) #1297

        On SCO OSR5, if you have a fuzzy browse lookup, and pass

        it a zero-length key, filePro can crash.

(OSR5) #1299

        Fuzzy search on SCO OSR5 returned incorrect results.

(Windows) #1301

        When using PFIXS=ON or the "-jy" flag, it was possible                    

        for Windows to throw an "uninitialized variable" error

        when scanning for records using an indexed field.

(All) #1304

        If you have a fatal error (such as "file not found" on

        import) in a CALLed table, filePro might crash upon exit.

(All) #1305

        If you print a form from *clerk via "F"orm (not from FORM

        within processing), and that form's output processing does

        a lookup to the same file as an open lookup in input/auto

        processing, and CLOSEs that lookup, filePro can crash if

        the input/auto processing re-executes that lookup.

(All) Version 5.8

       New function to get error code for ENCRYPT()/DECRYPT()

       failure.

            status = CRYPTERROR( [format] )

       If "format" is omitted, or zero, then the value is

       returned as a numeric error number, or zero for "no

       error".  If "format" is "1", then the value is returned

       as a string.  Other values for "format" are undefined.

       (If ENCRYPT/DEcrypt fails, a null string -- "" -- is

       returned.)

(All) Version 5.8

        New function to set the ODBC query timeout:

    old = @ODBC.handle.TIMEOUT(timeout)

        where "timeout" is the desired timeout in seconds.  The    

        function returns the old timeout value, if available, or ""

        if not.

        The default timeout is 15 seconds.  Not all ODBC sources

        allow the timeout to be set.  Setting the timeout to zero

        disables any timeout fuctionality, and can cause filePro to

        simply wait forever.

        This only affects future queries.  Also, some ODBC sources

        share the timeout between all handles attached to the same

        ODBC_CONNECTION handle.