CONTACTS 
0201 - Intel Corporation                                                                                                                                                  O-00000704 0201
020101 - Mr. Morris E. Jones;

SUBJECTS
Get Command Improved with New Parameters to Specify Line Number to Be
Medit Get Command Create Alternate with Parameters to Read in a File
Failure Medit After 386 Cycles Get Command Profile May be Overloading
Get Command Causes Medit to Fail after 386 Cycles May Indicate Cause
Fail Medit Get Command 386 Cycles May Indicate Cause of Medit Memory
Memory Fails 386 Cycles Get E Commands Open Profile Each Pass Get Co

1108 -
1108 -    ..
1109 - Summary/Objective
1110 -
111001 - Follow up ref SDS 2 0000, ref SDS 1 AT6E.
111002 -
111003 - Developed a test application routine to read in 10 lines, which was
111004 - successful, and so was incorporated into macro file 03501 for the
111005 - Subject Index profile, reported in another record today. ref SDS 2
111006 - RR3O
111008 -  ..
111009 - Testing however showed that Medit fails with repeated use of get
111010 - command in this function after exactly 386 cycles, which means reading
111011 - 3,860 lines. ref SDS 2 C96O
111012 -
111013 -     [On 040528 getting failure in Subject Index after reading only
111014 -     1,177 lines, which is less than half of 3,860, when recl 3000 is
111015 -     set. ref SDS 4 F15O
111017 -      ..
111018 -     [On 040528 report that justify Alt F9 calling macro 49 causes
111019 -     Medit memory to fail after repeated uses on a paragraph with a lot
111020 -     of lines. ref SDS 4 UL8G
111022 -      ..
111023 -     [On 040706 further problems led to revising default recl, and
111024 -     plans to meet with Morris on 040724 about fixing the problem.
111025 -     ref SDS 5 VR9J
111027 -      ..
111028 -     [On 040910 Morris fixed problem with get command. ref SDS 6 0001
111029 -
111031 -  ..
111032 - Submitted ref DIT 1 0001 to Morris saying....
111033 -
111034 -    1.  Preliminary testing seems to indicate that the new feature you
111035 -        created may work even better than expected. ref DIT 1 0001
111037 -         ..
111038 -    2.  Developed a procedure for scrolling through the file x number
111039 -        of lines, per pass, where in our case x = 10.  Think the same
111040 -        can be done for scrolling up and down 1 line at a time, and
111041 -        should be able to rig something that jumps to the top and
111042 -        bottom of the file and to specific lines.  Also can fashion a
111043 -        file search method.  All sounds helpful for memory management;
111044 -        so, we'll see how it turns out over the next few days.
111046 -         ..
111047 -    3.  Testing the initial tool to scroll up 10 lines, shows
111048 -        repeatable failure of get command after 386 cycles [reported in
111049 -        another record. ref SDS 2 C95J]  Attached test files will let
111050 -        you observe the problem.  tst is the macro; tst3 is the file to
111051 -        open and call tst -- it just starts off with the first screen
111052 -        of stuff;  and tst1 is the data file where the macro tst gets
111053 -        the stuff to read into tst3.
111055 -         ..
111056 -    4.  Theoretically, the limitation of processing 3,700 lines is not
111057 -        fatal for the initial application, but for other things we may
111058 -        need more capacity.  Is there an inherent limitation on the
111059 -        number of times get can be used?
111060 -
111061 -            [On 040910 Morris fixed problem with get command.
111062 -            ref SDS 6 0001
111064 -  ..
111065 - Morris wrote back saying....
111066 -
111067 -    1.  Subject: RE: Test Problem with Get
111068 -        Date: Fri, 14 Nov 2003 15:46:36 -0800
111069 -        From: "Morris Jones"
111070 -        To: "Rod Welch"
111072 -         ..
111073 -    2.  There is no limitation on the number of times you can use with
111074 -        get.  Not sure why things are failing...
111076 -         ..
111077 -    3.  Morris
111078 -
111079 -
111081 -  ..
1111 -
1112 -
1113 - 1625
1114 -
111401 - Called and talked to Morris.
111402 -
111403 - He said the problem with get failing after a lot of cycles (actually
111404 - 386, per above, ref SDS 0 3F9U) suggests a memory problem.  He said
111405 - that when the get command runs, Medit calls a file profile to assign
111406 - function keys, and feels this may be causing memory loss even though
111407 - "profile" is not on the path and so not actually opened.
111409 -  ..
111410 - He suggested running the test code with a "purge profile" command.
111411 -
111412 -     [On 040910 Morris fixed problem with get command. ref SDS 6 0001
111413 -
111414 -
111416 -  ..
1115 -
1116 -
1117 - 1648
1118 -
111801 - This afternoon did an experiment by creating a file "profile" in the
111802 - current directory and then ran from DOS.....
111803 -
111804 -        C>e fred
111806 -  ..
111807 - The function keys in "profile" were active, which aligns with Morris'
111808 - explanation about profile being called when Medit is run.
111810 -  ..
111811 - Did another test after opening fred by calling the "free" command
111812 - Morris created to show available memory.  After executing the command
111813 - purge profile, the memory diagnostic showed a smaller number.
111815 -  ..
111816 - Called "get" command against a filename.  This did not result in
111817 - restoring commands in "profile," which suggests that Medit does not
111818 - load "profile" with the "get" command.  None-the-less "get" may use
111819 - similar code to the "e" command and so may be having an adverse impact
111820 - on memory even though the full functionality of loading the "profile"
111821 - does not occur.
111823 -  ..
111824 - Tried the experiment Morris suggested of putting...
111825 -
111826 -                      purge profile
111827 -
111828 - ...in the code for scrolling through the sorted index using the get
111829 - command, shown in another record. ref SDS 2 EE3F  Still got a failure
111830 - on cycle 387, which is processing 3700 lines.
111832 -  ..
111833 - Possibly the "profile" system is using up memory and this is
111834 - manifested in the 387 cycles that the new get command routine exposed.
111835 -
111836 -
111837 -
111839 -  ..
1119 -
1120 -
1121 - 1749
1122 -
112201 - Submitted ref DIT 2 0001 to Morris saying....
112202 -
112203 -    1.  Maybe we can recover some compiler space by removing the call
112204 -        to "profile" since you are not using Medit and SDS develops a
112205 -        profile based on the applications using calls to macros.
112206 -        ref DIT 2 0001
112208 -         ..
112209 -    2.  If you are loading up something each time a file is opened,
112210 -        this could account for memory problems.
112212 -         ..
112213 -    3.  I am curious why a get command would give rise to an automatic
112214 -        call to profile, since stuff is being read into an existing
112215 -        environment. [...reported in another record today. ref SDS 2
112216 -        EE3F]  In a simple case, say you open a file fred, and lets
112217 -        assume Medit calls "profile" which sets up the function keys
112218 -        according to the list in your Medit directory. Suppose you
112219 -        assign F8 to temporarily do something to test and study the get
112220 -        command. ref SDS 0 LF43  Then, each time you call get, you lose
112221 -        the custom assignment because it runs "profile." Not sure why
112222 -        it should work that way, but in any case, it would be a relief,
112223 -        if we have discovered after all this time a source of memory
112224 -        problems that can be solved by eliminating the call to profile.
112226 -         ..
112227 -    4.  If you have time after dinner, please call on this stuff.
112228 -
112229 -           [On 031116 letter to Morris suggests disconnecting automatic
112230 -           call to "profile" may help solve memory problems.
112231 -           ref SDS 3 YY7M
112232 -
112233 -           [On 040910 Morris fixed problem with get command.
112234 -           ref SDS 6 0001
112235 -
112236 -
112238 -  ..
1123 -
1124 -
1125 - 1834
1126 -
112601 - Morris wrote back saying he cannot work on Medit because he has work
112602 - related to his assignment at San Jose State....
112603 -
112604 -    1.  Subject: RE: Test Problem with Get Profile Experiment Failed
112605 -        Date: Fri, 14 Nov 2003 18:34:03 -0800
112606 -        From: "Morris Jones"
112607 -        To: "Rod Welch"
112609 -         ..
112610 -    2.  Rod,
112611 -
112612 -        I have to write a midterm for tomorrow, and get ready for a ham
112613 -        radio test in the morning.  No time now.
112615 -         ..
112616 -    3.  Morris
112617 -
112618 -            [On 040910 Morris fixed problem with get command.
112619 -            ref SDS 6 0001
112620 -
112621 -
112622 -
1127 -