5.6.11 Release Notes

(All) # 1162
    In rclerk/rreport (only), the functions FIELDNAME(), FIELDEDIT(),
    FIELDLEN(),  and FIELDVAL() would all fail with fatal "reference
    to a field that doesn't exist" error if given a non-existent
    field.  In dclerk/dreport, such a condition was handled by
    returning a null string.?
(All) #1091
    If you have a protected lookup in @WBRK processing, and don't modify
    or WRITE the lookup, the record would remain locked.
(All) #1093
    A selection set comparing a memo field equal to "" doesn't work.
(All) #1107
    DOKEY from a CALLed table would not do anything.
(All) #1113
    If you use a file I/O function in output processing, and assign the
    result to a dummy field not defined in automatic processing, the I/O
    will occur multiple times.
(fileProGI) #1121
    Under some circumstances, GI could change the contents of a field on
    the screen, even if the user never entered that field.
(Linux) #1122
    dmakemenu could crash when pressing DEL/Ctrl-C after using the F5
    shell script editor.
(All) #1123
    dmoedef could crash when loading an output format with a sort field
    higher than any field number in the file.
(Linux) #1124
    F9/search in *cabe wouldn't properly update the display when the user
    presses Backspace in the search field.
(*nix) #1132
    Improved signal handling, to properly handle things such as closing
    a window in FacetTerm while filePro is still running.
(fileProGI) #1133
    Under fileProGI, if you use the MENU command with an array, where some
    elements were never assigned a value, filePro could crash.  (For example,
    DIM the array for 10 elements, and only assign a value to the first 5.)
(Linux) #1135
    ddefine could crash on some Linux systems on some circumstances.
(fileProGI) #1136
    Under fileProGI, if you have PFPT=ON set, and do not have any values
    set for PFTEMP, TEMP, and TMP, filePro can crash when executing a
    command with the "-pq" flag.
(*nix) #1139
    showlock didn't display correct information on some 64-bit *nix
    releases.
(All) #1150
    It was possible to get an erroneous "subscript out of bounds" error
    when referring to an array which was aliased to dummy fields or an
    associated field group.
(All) #1151
    On some platforms, it was possible for opendir() to reurn a list of
    filenames of the correct length, but which contained some blank
    entries instead of the names.
(All) #1152
    If you assign a value to a memo field via processing, and then refer
    to that same memo field before writing the record, the memo field
    can become corrupted, or cause filePro to crash.
(All) #1153
    When building an old-style index, dmxaint can fail with an "invalid
    argument" error should the index exceed 2GB in size, even on
    platforms that support 64-bit I/O.
(All)
    When restructuring a file larger than 2GB in size, ddefine could
    corrupt the data.
(All) #1154
    ddefine will now update the record counter during restructure only
    once a second, rather than every 100 records.
(All) #1155
    When calling ENCRYPT() without a nonce, and then calling GETNONCE()
    to read the value, it is possible for the same nonce to be generated
    more than once.
(All) #1140
    Under Windows, if a filePro program was run from the task scheduler,
    and an error occurred, filePro would wait forever for the non-existent
    user to press Enter to continue.
(All) #1141
    If you have a MEMO EDIT command within @WUK processing, and have
    the "-d" flag on the command line, the memo editor prompts were
    left on-screen upon saving the memo.
(All) #1142
    Quikstart didn't properly support dummy fields declared as (16,memo).
    See also #915.
(All) #1143
    In some cases, if you got a symtax error on a "large" processing
    table in rcabe, the program would crash upon pressing Enter at the
    error message.  (Note: rcabe only.)
(fileProGI) #1147
    When running under GI, filePro on Windows will now include "local
    printer" as a destination when using "-pt".
(All) #1154
    When restructuring a file, ddefine/autoshuf will now update the
    record counter only once a second, rather than every 100 records.
(All) #1159
    Under "just the right combination of conditions", using *cabe's
    lookup wizard would change the lookup type to "Z-Fuzzy" within
    the wizard.
(All) #1161
    If you have an index built on a field with a user edit (ie: not a 
    system edit), and you add or delete edits prior to the edit
    definition, dclerk might not respect the edit in the "enter index
    search data" dialog after the initial scan.
    Note that this is typically harmless, but it can interfere with
    the behavior of right-justified edits.