hobbes.nmsu.edu/2008-06-02_hobbes.nmsu.edu.zip/dos/viz423.zip/VIZ.DDT

Text File | 1992-02-24 | 11KB | 208 lines

                      VIz
                      ---
             Copyright (c) 1989-1992, FM de Monasterio
              Licensed Material - All Rights Reserved


    BUG REPORTS

  * When using release 1.40 thru 1.53 with the DOS line editor DOSEDIT.COM, a
    bug is triggered when returning from the help screen of VIz when the video
    mode 01 has been selected and the <PgUp> key is pressed.  This bug results
    in a loss of the cursor.  This behavior is not triggered in other DOS line
    editors (such as STACK.COM) and has been compensated in VIz release 1.54+.

  * When using QEMM and VIz 1.53+ there was a report of conflicts when loading
    VIz in high-DOS memory that caused the (386) machine to hang. The problem,
    however, was produced by insufficient room in high-DOS memory to allow for
    the expansion at loading time when using QEMM's LOADHI (see above).

  * When using release 1.54 with an ANSI driver the background color attribute
    was not always preserved past the 24th row. The problem has been corrected
    in release 1.55+.

  * When using release 1.54, changes of the video page by means other than VIz
    resulted in inactivation of the accelerator when returning to page 0. This
    problem has been solved in release 1.55+.

  * When running the full PCTECH benchmarks with release 1.56 (or 1.56a) on an
    EGA system, there are reports of missing characters at the bottom part of
    the graphics display of data comparisons. Such a problem is triggered by a
    strange video page change request by the benchmark while still in graphics
    mode; it does not occur in some EGA systems and in none of the VGA systems
    tested so far.  The problem has been solved in release 1.57.

  * A bug was introduced in release 1.58 that resulted in garbled scrolling of
    screen windows. It has been solved in release 1.59.

  * Release 1.59 has been reported to affect the functioning of Xerox/Olivetti
    machines; the bug seems to reside in the accelerator module. Such machines
    run release 1.57 apparently without this problem.  Since the machines have
    a proprietary CGA-like display it is expected the problem may be solved in
    the upcoming CGA release of VIz.

  * Users of DOS release 4.0x should be aware of reports of problems in using
    ANSI.SYS.  If you experience such problems, be aware that some are likely
    to be experienced with the ANSI-compatible driver of VIz as well.

  * Release 2.20 attempted to solve the lack of screen ANSI-color preservation
    when the DOS' CLS command is issued by linking together the ANSI attribute
    and the attribute of (old) switch /C.  This approach, however, resulted in
    conflicting screen colors with some programs that scrolled screen windows.
    The problem has been solved by removing the linkage and by implementing a
    different color selection scheme via switch /C that uses the EGA/VGA color
    palette (see Switch C above) in release 2.30.

  * When using switch F with a VGA, releases 2.31-2.32 toggled between 50 and
    28 rows, instead of 50 and 25 rows.  release 2.33 has corrected this bug.

  * Releases 2.31-32 would report a CRC error when using some monochrome card
    systems.  release 2.33 has corrected this bug.

  * When using the "view" option of NORTON COMMANDER (F3) the first 2 lines of
    the text displayed remain unchanged if the <PgDn> key is pressed while the
    remaining lines are scrolled. This was a bug in the scroll down service of
    VIz, fixed as of release 2.40.

  * When inserting <comments> in WORD PERFECT 5.0+ text, additional lines than
    those requested by the utility were scrolled into the text (so that "text"
    from the video page 1 also appeared in the WP text). This was a bug in the
    scroll down service of VIz, fixed as of release 2.40.

  * When running VIz 2.41 with switch /X+ implemented (DOS int 29h directed to
    the "teletype" service of the accelerator), there was a bug in the service
    for the bell (ASCII 7) that could result in erratic displays, such as mode
    changes. The problem has been fixed as of release 2.42.

  * Release 2.42 contained a bug that could trigger improper stack restoration
    on DOS return after an illegal video function request; this has been fixed
    in release 2.43.

  * When running VIz 2.20+ and the disk utility DAAG.EXE the attribute of some
    areas of the display is changed after the window for file extension counts
    is displayed and then erased.

  * VIz releases 2.41-2.43 contain an error in the DOS release checking service
    that results in an error message followed by aborted execution for releases
    3.00-3.09 and 4.00-4.09; this has been corrected in release 2.44.

  * VIz release 2.44 and prior ones providing control of <PrtSc> key contain an
    error making VIz disregard any further switches after switch S. It has been
    corrected in release 2.50.

  * VIz releases 2.41-2.50 contain an error in the EGA status checking service,
    which resulted in a CRC failure when some adapters were encountered. It has
    been corrected in release 2.51.

  * A cursor bug has been reported when using VIz release 2.51 and ANSI drivers
    in DOS 4.01 (multicolor cursor).

  * A bug in releases 2.30 through 3.20 returned incomplete video font data for
    subfunction 30h of function 11h of int 10h, producing garbled displays with
    some utilities such as MORE FONTS. It has been corrected in release 3.21+.

  * The utility SHELL.COM of WordPerfect Library makes repeated settings of the
    same video page per each line of screen displayed. Because VIz page service
    also participated in the reloading of the EGA/VGA color palette (switch C),
    such repeated settings of the same page produced a haltering video display.
    Starting with release 3.21, this resident page service does not participate
    in palette reloading, as other non-optimized, higher-language utilities may
    show the same behavior.

  * There have been reports of problems in loading into high-DOS memory the EXE
    file format of release 3.10-3.21 that seemed to confuse high-memory loaders
    such as 386LOAD.COM.  The previous .COM format has been restored in release
    3.40 (see VIZ.DOC).

  * Some programs, written in C-language, when encountering an invalid argument
    in the command line generated the display of 65,535 characters when switch
    /Z+ was selected in release 3.40 of VIz.  This problem, due to an incorrect
    value in a CPU register by these programs has been solved in release 3.50+.

  * An assembly macro in the BBS release was not updated in going from release
    3.40 to 3.50, and garbled part of the display triggered by switch /C?; this
    bug did not affect other releases.

  * A mistake during compilation of release 3.60 avoided the updating of switch
    /V status; corrected in release 3.61.

  * A flaw in programming logic resulted in erratic reactivation of VIz after a
    change to page 0 from another video page (releases 3.21+); this problem has
    been corrected in release 3.63.

  * In some EGA cards, selecting the alternate BIOS print-screen service yields
    a print-screen malfunction (e.g. only a line feed is printed). All releases
    up to 3.71 automatically selected this alternate BIOS service when a VGA or
    EGA card was detected. Starting with release 3.72, the alternate service is
    selected if switch /S+, /S-, or just /S, is requested at installation time.

  * On first installation there was an error in cursor positioning if the video
    had not been scrolled first; this error has been corrected in release 4.00.

  * There has been a report of a VIz-4.00 (BBS release) problem with an Orchid
    Prodesigner II Plus VGA card, in which VIz did not seem to turn itself off
    when the video mode was switched to an unspecified 'hi-resolution' graphic
    mode.  So far, such a problem has not been replicated in other VGA systems
    and awaits further study.

  * The octal number input service (used by switch /Bn, for example) contained
    an error that converted a single-digit input such as '7' into '70' instead
    of '07;' this error has been corrected in release 4.03.

  * There have been reports of a bug in the 086 release of release 4.12, which
    has been traced to an incorrect assembly macro assignment that resulted in
    double instruction-prefix opcode.

  * A bug in the video adaptor testing of VIz 4.13 crashed some machines with
    a CGA card; it has been fixed in release 4.14.

  * Release 4.15 fixes several errors which started to accumulate with release
    4.12. Of these, the most serious was an assembler (phase?) error resulting
    in the incorrect segment override during the updating of a data byte; as a
    consequence, the program changed the contents of (one-byte of) memory out-
    side its boundaries. Depending on system conditions, such error might have
    negligible to obvious effects (such as affecting the batch file processing
    by COMMAND.COM). ** USERS OF release 4.12 OR HIGHER SHOULD UPGRADE TO 4.15
    EVEN IF NO PROBLEMS HAVE BEEN EXPERIENCED **

  * In release 4.15, an error message was displayed when switch /A+ was issued
    after the ANSI resident was installed, unless switch /A- was issued first.
    This harmless error has been corrected in release 4.16.

  * In release 4.16, a potential error could be triggered by repeated pressing
    of the <F10><Alt> keys during the Status/Usage/Help display.

  * In the 486 version of release 4.16, an error resulted in the scrambling of
    the screen attributes when the 14-line screen font was selected.

  * In the 386/486 version of release 4.17 an error resulted in the scrambling
    of the screen attributes when the 35-line screen font was selected.

  * Releases prior to 4.20 cannot selfload in a UMB when the UMB allocation is
    controlled by MS-DOS 5.0 (DOS=UMB linkage), and cannot create or update an
    environment variable (in the global DOS environment block) when MS-DOS 5.0
    executes from the HMA.

  * Fixed a potential problem in using XMS services with DOS 2.x, and error in
    the cursor position upon exiting the display panel of switch C? in release
    4.20.  Modified implementation of StdOut processing to allow redirection.

  * Changed ANSI.SYS detection service to include the ANSI.SYS of MS-DOS 5, in
    which the attribute word at offset 4 was changed from 8013h to C053h.

  * Some utilities clear the screen via INT 10h, scroll-up function 6h, but do
    not reset the cursor its home position through an INT 10h call.  When such
    programs "clear the screen" upon their termination, it may be necessary to
    issue a CLS command to insure proper positioning of the cursor.

  * In release 4.22 the flags register is not always completely restored after
    checking for StdOut redirection, which results in a run-time error in some
    programs accessing INT21h. The problem seems to occur when other residents
    also intercept this interrupt service.

                -----------------
         ┌─────────────────────────────────────────────────────────────┐
         │ Trademarked names are the property of the respective owners │
         └─────────────────────────────────────────────────────────────┘
 [END]