1 
  2 09/21/87  misc_commands
  3 Known errors in the current release of misc_commands.
  4 #         Associated TR's
  5 Description
  6 
  7 528  phx20941
  8 If the segment requested for replacing is a non-existent segment then
  9 the archive replace operation refuses to create new archive component.
 10 This is true, but the archive command must print out an error message
 11 to users.
 12 
 13 
 14 526  phx17201
 15 Assume there is a null link named foo, then the get_effective_access
 16 command returns error_table_$action_not_performed for a null link.  It
 17 should return error_table_$noentry for a null link.  For example:
 18 
 19    ! link >foo
 20    ! gea foo
 21      get_effective_access: The request operation is not supported.
 22      >udd>m>Lippard>foo
 23 
 24 525  phx17148
 25 If users call the suffix_name_$new_suffix entrypoint and supply the
 26 input argument suffix that begins with a period character (".") then
 27 the returned new suffix begins with two consecutive periods ("..") and
 28 the return error code value is zero.
 29 
 30 
 31 518  phx14448
 32 Assume there is a "-working_dir" search path contained in the "info"
 33 search list.  The help command gets error when it hits the -working_dir
 34 search path.  For example:
 35 
 36   ! create_dir test_dir
 37   ! cwd test_dir
 38   ! delete_dir [wd]
 39   ! help
 40     Error:  get_wdir_ error by change_wdir_$get_wdir_ |1065
 41             (>sss>bound_fscom1_)
 42             No working directory set for this process
 43 
 44 This is incompatible with its behavior.  If there is a pathname in the
 45 search path that references a non-existent directory, then the help
 46 command should issue a warning message and then continues to search.
 47 
 48 
 49 517  phx05578
 50 The list commands incorrectly reports effective mode for inconsistent
 51 MSFs.  Currently, it reports "rw" modes for a specified inconsistent
 52 MSF.  It must report only the "r" mode for a specified inconsistent
 53 MSF.
 54 
 55 516  phx14711
 56 The previous error list entry has been lost.  This problem is a limited
 57 display field.  See the TR.  This problem is repaired in the exl
 58 version of list.
 59 
 60 515  phx20750
 61 exec_com [verified] gets confused by unbalanced parentheses in
 62  quoted strings .
 63 
 64 
 65 514  phx20749
 66 print command [verified]. The print command should not change tty_ modes.
 67 
 68 513  phx20747 segs active function [verified].   The order of the
 69  segments listed has been incompatibly changed without notice.
 70 
 71 512  phx20744 fortran_abs command [verified]
 72  "fa" command doesn't accept -card control arg.  .
 73 
 74 511  phx20699
 75 display_audit_file [verified].  Generates error when asked to display
 76  an audit file with nothing in it other than the header.
 77 
 78 
 79 510  phx20694
 80 Problem description:  The "display_audit_file" command accepts a
 81  non-standard file name for "-output_file" control argument with a
 82  command like:
 83 
 84    !  daf -of nuke..
 85 
 86 Proposed solution:
 87 
 88 The "display_audit_file" command will not accept strange names when
 89 "-output_file" control argument is specified.
 90 
 91 The "check_star_name_" system subroutine will be called in order to
 92 produce correct error codes.  The control mask contains CHECK_STAR_
 93 REJECT_WILD which declared in check_star_name.incl.pl1.
 94 
 95 
 96 508  phx20673 enter_output_request (eor) [verified]
 97  If you request that a file be indented when printed and that it
 98  print with centered labels, the labels are also indented off center.
 99 
100 507  phx20811 phx20658 phx19390 phx17261
101 Problem description:  The reorder_archive command references
102  error_table_$record_quota_ overflow.  This generates a linkage fault.
103 
104 Proposed solution:
105 
106 Change the reorder_archive.pl1 to reference the error_table_$rqover
107 entry point which is defined in the error_table_.alm.  Note that this
108 status code is already documented in the Multics Programer's Reference
109 Manual, page 7-23.
110 
111 
112 506  phx20645 date_compiled [verified].  Prints more than was asked for.
113 
114 505  phx20639 access_commands [verified].  The -exclude facility does not
115  exclude ACL entries with explicit tags if a "*" is given for the tag in
116  the argument list.  If the user exceeds the allowed ten -ex arguments,
117  the command can malfunction.
118 
119 504  phx20630 exec_com [verified].
120  Bad parsing of arguments in &set command.
121 
122 
123 503  phx20628 delete [verified].  The delete command incorrectly
124  prints the message for error_table_$badstar when given a blank
125  argument.
126 
127 502  phx20573 exec_com &trace [verified].  exec_com &trace &prefix
128  will not accept a quoted string with embedded blanks.  Worked okay
129  in MR10, fails in MR11.  .
130 
131 501  phx20481 copy [verified].  copy -all gives incorrect error message.
132 
133 499  phx19798 cv_ptr_ [investigating].  Doesn't work on inner ring
134  virtual pointer strings.  .
135 
136 
137 498  phx19667 cv_float_ subroutine [verified].  cv_float_ doesn't
138  accept "1.e3" as first argument ("1.0e3" is OK however).
139 
140 497  phx19503 ioa_ [verified].  The ioa_ control string ^va gives
141  different results than ^a when a BS ("010"b3) character is passed.
142 
143 496  phx19233 truncate [verified].  Fails to truncate an MSF, even
144  though I have rw access to it.
145 
146 
147 495  phx17260
148 Problem description:  pa_search_list.pl1 references the undefined
149  error code, error_table_$programming_error.
150 
151 Proposed solution:  Remove error_table_$programming_error from
152  pa_search_list.pl1
153 
154 494  phx170221 error_table_ and company [verified].  There are
155  error_table_ codes which are not referenced in the system.  Worse
156  yet, there are codes referenced which are not defined!
157 
158 
159 493  phx14945
160 Problem description:  When an user fails to supply a path for
161  -output_file, then compare_ascii command uses the argument follows
162  -output_file as the pathname.  For example:
163 
164    ! cpa foo bar -of -he
165 
166 directs the output of the comparison to the "-he" file.
167 
168 Proposed solution:
169 
170 Use error_table_$noarg to produce error code when an user does not
171 supply an output file for -of.
172 
173 
174 491  phx14364 compare_ascii [verified].  Indication of a large
175  block of differences when there aren't any.
176 
177 484  phx20624 find_common_acl_names_ [verified].  This subroutine
178  returns error_table_$badstar where error_table_$bad_name should
179  be returned.  As time and resources permit this problem will be
180  evaluated and better defined at a future date.
181 
182 
183 483  phx20622
184 Problem description:  The "rename" command misuses the
185  "error_table_$badstar" error code in two different ways.
186 
187    a.  "rename" displays an incorrect error message if the "-name" control
188        argument is NOT specified for nonstandard entrynames. For example:
189 
190        ! rename foo  *.FOO
191          rename:  Invalid syntax in starname. *.FOO
192 
193    b.  So does for an ojbect type such as a mailbox.  For example:
194 
195        ! rename TLNguyen.mbx foo
196          rename:  Invalid syntax in starname. foo
197 
198 
199 Proposed solution:
200 
201 The "check_star_name_" must be used in order to produce correct error
202 code.  The control mask should contain CHECK_STAR_IGNORE_ EQUAL since
203 the target pathname accepts equal convention.
204 
205 Change "error_table_$badstar" to "error_table_$not_seg_type" to report
206 an invalid suffix (.foo) has been given for an object type such as a
207 mailbox (.mbx)
208 
209 
210 481  phx20597 archive [verified].  archive aborts unpleasantly
211  when it tries to add a module that will push the archive over
212  it's maximum segment length.  As time and resources permit this
213  problem will be evaluated and better defined at a future date.
214 
215 480  phx20572 calendar [verified].  calendar does not properly
216  handle centuries other than 1900.  As time and resources permit
217  this problem will be evaluated and better defined at a future date.
218 
219 478  phx20562 locate_words [verified].  Displays the wrong lines
220  when -lines is used.  As time and resources permit this problem
221  will be evaluated and better defined at a future date.
222 
223 
224 475  phx20430 ldw command [verified].  There appears to be an
225  internal inconsistancy in >udd>Doc>lib>Doc.dict which causes
226  the ldw command to loop.  As time and resources permit this problem
227  will be evaluated and better defined at a future date.
228 
229 472  phx20415 sort_strings [verified].  Fails when one of the
230  arguments to be sorted contains a newline character.  As time
231  and resources permit this problem will be evaluated and better
232  defined at a future date.
233 
234 470  phx20375 general [verified].  (ls -dtcm) "list file -dtcm"
235  gives incorrect info when file is a MSF.  As time and resources
236  permit this problem will be evaluated and better defined at a
237  future date.
238 
239 
240 469  phx20366
241 online documentation, cv_float_check_.info [investigating]
242  This help file appears to be obsolete.  The entry is not in the
243  manual, I'm unable to locate it on the system, and the described
244  entry does not seem to offer anything not done by cv_float_.
245  As time and resources permit this problem will be evaluated and better
246  defined at a future date.
247 
248 468  phx20363 delete_dir or suffix_forum_ mode [verified].  A ring1
249  process is unable to delete a directory whose subhierarchy contains
250  a forum meeting.  As time and resources permit this problem will be
251  evaluated and better defined at a future date.
252 
253 467  phx20305 dump_segment [error].  dump_segment_$string conflict .
254  As time and resources permit this problem will be evaluated and better
255  defined at a future date.
256 
257 
258 466  phx20291 Documentation/help files [verified].  help files for
259  sort_ and merge_ are missing.  As time and resources permit this
260  problem will be evaluated and better defined at a future date.
261 
262 465  phx20241 fortran_abs [investigating].  fortran_abs incorrectly
263  enforces (-card & (^-quote | ^-point)).  As time and resources permit
264  this problem will be evaluated and better defined at a future date.
265 
266 464  phx20071 calc [investigating].  The calc command does not
267  diagnose random characters between a function name and the associated
268  argument for the function.  As time and resources permit this problem
269  will be evaluated and better defined at a future date.
270 
271 461  phx19950 get_shortest_path [verified].  Faults in inner ring in
272  particular conditions.  As time and resources permit this problem will
273  be evaluated and better defined at a future date.
274 
275 
276 459  phx19848 cross_reference [verified].  cross_reference
277  incorrectly reports the namesof source segment that were
278  components of archives.  As time and resources permit this
279  problem will be evaluated and better defined at a future date.
280 
281 
282 458  phx19836
283 Problem description:  If you attempt to run delete_acl on
284  "Data_Managment.Daemon" access control name from a DM file with
285  a command like:
286 
287    !  da dmf Data_Managment.Daemon
288 
289  then no warning error message is printed.  It is true that users are
290  not supposed to remove Data_Management.Daemon's access from a DM file,
291  but users should be told that he cannot do such thing.
292 
293 
294 Proposed solution:
295 
296 Change dm_error_.alm and fm_attribute_.pl1 to return an appropriate
297 error message to an user when he/she tries to manipulate ACL of
298 Data_Management.Daemon.* from a DM file.
299 
300 I.  Add the folowing error code to dm_error_.alm
301 
302    dm_error_$acl_restriction
303    (ACL change invalid "rw Data_Management.Daemon.*" required on
304     DM files.)
305 
306    dm_error_$acl_entry_no_rw
307    (ACL element denies "rw Data_Management.Daemon.*" required on
308     DM files.)
309 
310    dm_error_$acl_entry_no_star
311    (DM files allow only Data_Management.Daemon ACL entry with
312     a star instance tag.)
313 
314 
315 II.  Change the fm_attribute_.pl1 such that
316 
317   1. For add_acl_entries,
318 
319      if Data_Management.Daemon.* ACL exists:
320         if mode is rw then the operation is OK;
321         else reject that element and abort the operation;
322 
323   2. For delete_acl_entries,
324 
325      if ACL contains element for Data_Management.Daemon.*
326         then reject that element and abort the operation;
327 
328 
329   3. For replace_acl
330 
331      if ACL entry is for Data_Management.Daemon.*:
332         if mode is rw then opeation is OK;
333         else reject that element and abort the operation;
334 
335      if ACL does not contain entry for Data_Management.Daemon.*
336         then if sysdaemon_sw is on
337              then add rw Data_Management.Daemone.* term;
338              else reject the operation;
339 
340 
341 III.  Finally, I have changed the status of this TR from
342       "dev" back to "error" because of the following reasons:
343 
344       1.  The current delete_acl entry in acl_commands_.pl1
345           invokes fs_util_$replace_acl instead of fs_util_$
346           delete_acl_entries.  Therefore, in order to fix
347           the problem correctly, Gary Dixon said
348 
349 
350           a. to change the delete_acl so that the
351              delete_acl will pass matched individual access name
352              and its code to fm_attribute_$delete_acl_entries.
353              As a result, fm_attribute will return the deleted
354              access name and the code back to the delete_acl
355              so that the error message will be included the
356              deleted access name which is typed in by an user.
357              Currently, the delete_acl only passed UN-matched
358              individual access names and their error codes to
359              fm_attribute_$replace_acl.  Therefore, fs_atribute_
360              does not have information that the user typed
361              in.  So it only return the new error code back to
362              the delete_acl.  As a result, the error message does
363              not have specified access names that the user wants
364              to delete.
365 
366 
367           b. to change the set_acl to call fs_util_$add_acl_entries
368              to add a new acl entry to the current acl entries.
369              Currently, the set_acl calls fs_util_$replace_acl
370              to add a new acl entry to the current acl entries.
371 
372 
373 453  phx19616
374 Problem description:
375 
376 Try to reference the path1 when moving a specified segment (which is
377 already existed in path1) to a specified parent directory position.
378 Assume that this new position has multiple names and the user has "sma"
379 access modes to a specified parent directory.
380 
381 
382 For example:
383 
384  ! cd dir1; an dir1 dir1.name; cwd dir1
385  ! cd dir2; cwd dir2; cr seg1
386  !  move seg1  <
387     move:  Name duplication.  Old name dir1 remove from >udd>m>barmar>dir1.name
388     move:  Some directory in path specified doesn't exist.
389            >udd>m>barmar>dir2>seg1
390     Segment  >udd>m>barmar>dir2>seg1  not deleted
391 
392 
393 452  phx19589 trace [error].  References a non-existent thingie.
394  As time and resources permit this problem will be evaluated and better
395  defined at a future date.
396 
397 
398 449  phx19518
399 Problem description:
400 
401 If you attempt to run "delete_acl" against the component within a forum
402 meeting with a command like:
403 
404    !  da xxx.forum>Proceedings.1 Margulies.. Holmstedt.. -brief
405 
406 in a ring 1 process like Backup.SysDaemon, then the ACL get changed
407 (and xmode get lost) even though there are no matches for Margulies..
408 or Holmstedt..  on the ACL.
409 
410 NB:  suffix_forum_ is NOT involved here because fs_util_ does NOT
411 realize that the segment in question is a component of a forum meeting.
412 
413 
414 Proposed solution:
415 
416 The following conversion between Frank Martinson and Gary Dixon
417 explained why I have changed the "not_error" status back to "error"
418 status:
419 
420 
421 I'm sorry, but this is still an error.  I am therefore reopening the
422 TR.  I don't know what Frank Martinson may have done to reverify the
423 TR.  I had no problem reproducing the problem, as follows:
424    ! fcr
425    ! test1.forum
426    ! test.forum
427    ! Martinson
428    ! Beattie
429    !  .  This creates a test forum.  Use list_acl to examine its ACL.
430 Then login as Ring_1_Repair and change to the wdir containing the new
431 forum.
432    ! cwd test1.forum
433    !  da Attendees Brunelle..  Then logout from Ring_1_Repair.  If you
434 now use list_acl from your own process on the test1.forum, you will see
435 that now all ACL entries show null extended access mode.  The
436 delete_acl command has modified the ACL, even though there was no match
437 in the ACL for Brunelle.  In the process of modifying the ACL, it has
438 destroyed the extended access modes in all ACL entries.
439 
440 Therefore, the following proposal still holds:  the "delete_acl" must
441 not change the ACL when no matching entries are found.
442 
443 
444 447  phx19473 >sss>pl1.dcl [verified].  There is an entry for
445  assign_$computational which does not exist and no entry for
446  assign_$computational_ which does exist.  As time and resources permit
447  this problem will be evaluated and better defined at a future date.
448 
449 446  phx19461 library_print [verified].  lpr does not properly
450  handle ** as a search name.  As time and resources permit this problem
451  will be evaluated and better defined at a future date.
452 
453 445  phx19431 check_info_segs command [change_pending].  Unable to
454  update time value in value seg without listing all the changed info
455  segs.  As time and resources permit this problem will be evaluated
456  and better defined at a future date.
457 
458 
459 443  phx19277 l_set_acl, l_set_ring_brackets, etc [verified].
460  These commands don't work with inner-ring MSFs.  As time and
461  resources permit this problem will be evaluated and better
462  defined at a future date.
463 
464 439  phx19109 list command [verified].  Doesn't rtrim the pathname
465  in ""Directory empty"" message.  The exprimental library version
466  of this command fixes this problem.
467 
468 438  phx19106 save_dir_info [verified].  Produces error saving quota
469  of a dir immediately below the root.  As time and resources permit
470  this problem will be evaluated and better defined at a future date.
471 
472 
473 436  phx19051 system active function [investigating].  The system
474  active function / command should use the process date_time format
475  not the date format and the time  format.  As time and resources permit
476  this problem will be evaluated and better defined at a future date.
477 
478 435  phx19039 mrpg [investigating].  Incorrect code generated by mrpg.
479  As time and resources permit this problem will be evaluated and better
480  defined at a future date.
481 
482 434  phx19026 trim_wordlist [verified].  Words in the dictionary with
483  an apostrophe do not trim out words that are the same except for being
484  capitalized.  As time and resources permit this problem will be
485  evaluated and better defined at a future date.
486 
487 
488 433  phx18987 trace command [error].  The command "trace foo$*" prints
489  a confusing message when "foo" is a name on a bound segment that does
490  not contain a component "foo".  As time and resources permit this
491  problem will be evaluated and better defined at a future date.
492 
493 432  phx18949 memo [verified].  Repeating memo's cause strange this to
494  happen if a user with multip is logged on twice.  As time and resources
495  permit this problem will be evaluated and better defined at a future date.
496 
497 431  phx18928 describe_entry_type [verified].  Should have an option to
498  display all accepted switch names,  rather than just the primary name
499  and default switch setting.  As time and resources permit this problem
500  will be evaluated and better defined at a future date.
501 
502 
503 430  phx18917 response AF [error].  The response active function trims
504  the question.  As time and resources permit this problem will be
505  evaluated and better defined at a future date.
506 
507 427  phx18830 peruse_crossref [verified].  Should accept names of the
508  form *system$xmail_data_ptr to list references to external vars whose
509  names do not include $.  As time and resources permit this problem will
510  be evaluated and better defined at a future date.
511 
512 425  phx18729 >doc>is>list_acl.info [verified].  The info file is
513  seriously out of synch with the command.
514 
515 423  phx18336 copy command does not copy an msf correctly [verified].
516  copy command does not copy a multi-segment file correctly if there are
517  gaps in the sequence numbers of the component segments.
518 
519 
520 421  phx18231 commands [verified].  calc shouldn't set up a pi handler
521  when invoked as an active function.
522 
523 420  phx18216 swn & swf commands [verified].  Documentation out of
524  sync with software.
525 
526 419  phx18187 mrpg [investigating].  mrpg fail with fault_tag_1 ascii
527  data where pointer expected.
528 
529 416  phx17988 copy_dir & move_dir [investigating].  Using copy_dir or
530  move_dir with a target which is contained in the subtree being
531  processed is pathological {pathname-ological}.
532 
533 415  phx17983 library tools [investigating].  "archive" and
534  "library_info" report different times updated for archive components.
535 
536 
537 414  phx17976
538 Problem description:  The "copy_dir" command adds the names of the
539  source directory to the target directory.  This causes unexpected
540  "Do you want to delete ..." messages to appear for apparently no reason.
541 
542 For example:  the current working directory is
543              >udd>m>TLNguyen>copy_dir_dev
544  ! ls -a
545    segments = 1,  lengths = 0
546    r w   0   save
547 
548   directories = 1
549   sma  source
550        s
551 
552 
553  ! copy_dir source  <save
554    copy_dir:  creating   >udd>m>TLNguyen>save
555 
556  ! ls <save -dir
557    sma   save
558          source
559          s
560 
561  !copy_dir <save  source -replace
562   copy_dir:  Name duplication.  Do you want to delete the old segment
563              >udd>m>TLNguyen>copy_dir_dev>save?
564 
565 
566 Proposed solution:
567 
568 This is not an error.  The copy_dir command works correct as its
569 documentation.  Since the user did not specified the "-primary" control
570 argument the copy_dir command copied all the names of the source
571 directory to a specified target directory.
572 
573 For the particular exampled describled above.  The copy_dir command
574 copies all the names of the source directory named "source" and "s" to
575 a specified target directory named "save".  As a result, "save" has
576 multiple or additional names which are "source" and "s".  Further-
577 more, the last error message is correct because the "nd_handler_"
578 subroutine is called to resolve the name duplication and "source"
579 already exists in the user'working directory,
580 >udd>m>TLNguyen>copy_dir_dev, and "source" is a single segment name (or
581 only one name).  Therefore, "nd_handler_" asks the user if he/she wants
582 to delete the old segment.
583 
584 Finally, the "copy_dir" command works correct according to its
585 documentation description.
586 
587 
588 413  phx17969 standard commands [investigating].  The "move"
589  (copy) command currently installed in >sss makes some
590  invalid assumptions about the suffix_XXX_$move capabilities.
591 
592 412  phx17939 general_ready [error].  It is impossible to get the
593  gr command to produce the date properly in UK format.
594 
595 411  phx17904 repeat_line (rpl) active function [error].  Imbedded
596  characters in prompt.
597 
598 410  phx17901 peruse_crossref [verified].  Peruse_crossref takes an
599  out_of_bounds when given a name which begins with an "invalid"
600  character.
601 
602 
603 408  phx17779 tools [investigating].  cross_reference (cref) does
604  not handle archive format errors in the middle of archives and faults
605  with an attempt to reference through null pointer.
606 
607 407  phx17771 cross_reference command [verified].  Deletes the
608  output_file if it gets an error initiating it.
609 
610 406  phx17743 library maintenance software [investigating].  The -time
611  control argument to library_cleanup does not accept date_time strings.
612 
613 404  phx17636 fortran_abs command [verified].  Does not accept valid
614  fortran arguments.
615 
616 403  phx17608 l_delete [verified].  l_delete loops trying to delete a
617  segment with its copy_switch on.
618 
619 
620 402  phx17575 l_delete [verified].  Gives wrong entry type in message.
621 
622 401  phx17483 cref, crossreferencer [verified].  A command ends up no
623  being cref'd as it and its synonym sho up as always indicating the user
624  should see the other entry.
625 
626 400  phx17234 peruse_crossref [verified].  Can't find references to
627  cv_dec_check_.
628 
629 399  phx17082 copy_dir [investigating].  Doesn't copy iacl of higher
630  rings {lower privilege} than ring of execution .
631 
632 398  phx17000 calc command [error].  Calc does it's arithmetic in binary,
633  and gives incorrect answers for simple decimal arithmetic.  When you add
634  up a list containing dollars and cents, the answer isn't correct.
635 
636 
637 397  phx16970 copy_cards command [investigating].  copy_cards
638  bungles starnames.
639 
640 396  phx16644 runoff_abs [change_pending] Linked TRs:  phx16311
641  runoff_abs does not work.
642 
643 395  phx16628 ou_ (object utils) [investigating].  Produces an
644  object segment with a looping definitions section.
645 
646 394  phx16521 general_ready active function [investigating].
647  [general_ready] returns its string with a trailing newline.
648 
649 393  phx16310 calc [verified].  Problem displaying small numbers.
650 
651 
652 391  phx16109 area_status [verified].  Prints incorrect offsets for
653  areas in the middle of segments.
654 
655 390  phx16050 algol_68_abs [change_pending].  The a68a command does not
656  recognise the -no_check control arg.
657 
658 389  phx15962 debug [verified].  If used in the video system, often
659  complains of "Undefined order request."
660 
661 388  phx15814 print command with -page_length control argument [verified].
662  print with -pl mishandles backspaces on a terminal .
663 
664 
665 387  phx15461 commands [error].  A number of commands do not check
666  that they are erroneously used as active functions.
667 
668 385  phx15189 list command [verified] Linked TRs:  phx05578.
669  Lists acl of an msf incorrectly.
670 
671 384  phx14921 list command [verified].  If given a control arg of -^d
672  will take a conversion error in formline_ or report error incorrectly.
673 
674 383  phx14676 sss -- copy_cards [verified].  This module is showing
675  its age and could be cleaned up quite a bit.
676 
677 382  phx14627 peruse_crossref [investigating]. I tried to
678  peruse_crossref backup_volume_record.incl.pl1.
679 
680 
681 381  phx14116
682 Problem description:  If you attempt to run "set_acl" on a specified
683  incosistent multi-segment file (MSF) with the following commands like:
684 
685    !  cd foo_dir
686    !  sbc foo_dir 1
687    !  sa  foo_dir rw
688       set_acl:  MSF is inconsistent
689 
690 
691 Proposed solution:
692 
693 I.  Change the msf_manager_ to ignore the "Entry not found" error code
694 when it calls hcs_$list_acl to list acl of the component zero within an
695 imcompleted (inconsistent) MSF.  Since the inconsistent MSF does not
696 have any component within it so hcs returns "Entry not found" to
697 msf_manager_.
698 
699 Currently, msf_manager_ will returns error_table_$inconsistent_msf to
700 the caller after hcs_ returns "Entry not found" to msf_manager_.
701 
702 
703 II.  Finally, Gary Dixon said that the solution only accepted according
704 to the existing implementation of the source program.  He wants me to
705 change the source to allow IACL-for_segment work on MSF directories.
706 So, for the case of an inconsistent MSF which does not have the zero
707 component then the set_acl will take the ACL of MSF since the ACL of
708 MSF is the same as the ACL of each component within the MSF .
709 Therefore, I will change status of this TR back to error and I will
710 work on this one again when my project leader says ok.
711 
712 
713 378  phx12360 list_mdir [error].  Tells you there are no master
714  directories when you know there are.
715 
716 369  phx09588 calc [error].  calc sets up a pi handler when invoked
717  as an active function.
718 
719 362  phx18653 ascii_to_bcd_ [verified].  (ascii_to_bcd_) subroutine
720  missing.
721 
722 
723 357  phx05578
724 Problem description:  The list command chooses the wrong place to look
725 for the ACL of an MSF.  That is the list command uses the directory
726 access mode, rather than using the ACL of component zero of the MSF.
727 As a result, the list command reports a misleading mode for an MSF with
728 an inconsistent ACL.
729 
730 For instance:  Assume that the correct mode was "r".  Then the list
731               command reports it as "rw".
732 
733 
734 347  phx19100
735 The the list commands handling of the command "ls < **" looks the same
736 as "ls <<**".
737 
738 This is not an error.
739 
740 By specifying "ls <" the containing directory is set to "<<" so the the
741 directory directly above the current on can be listed.  By adding the
742 "**" entry specifier afterward, the entries ** are being requested for
743 the same directory of the previous selection.
744 
745 345  phx11304
746 memo has strange bug-like incompatible changes.
747 
748 343  phx20419
749 (archive) Appends suffix at first occurrence of a space in the archive
750 name.
751 
752 
753 341  phx20322
754 (exec_com) Doesn't close [...]'s properly.
755 
756 340  phx20310 phx20323
757 (archive) Saying "ac r foo nonexistent_seg" produces no error message.
758 
759 339  phx20303
760 Problem description:
761 
762 In create_branch_info.incl.pl1, the version constants are not declared
763 "options (constant)", which causes them to be allocated in the static
764 option and be writable.
765 
766 Proposed solution:  Declare the version 2 as options (constant).
767 
768 
769 329  phx19927
770 (exec_com) &on catches only the first occurrence of any condition.
771 
772 327  phx12135
773 (list) "list -first 0" lists all entries.
774 
775 325  phx19818
776 (archive) Erroneous "All components have been deleted" message when
777 extracting into nonexistent directory.
778 
779 324  phx19725
780 (exec_com) Mis-read quoted string causes error about unbalanced
781 brackets.
782 
783 
784 315  phx19493
785 (convert_ec) Bugs in equal convention and error messages.
786 
787 314  phx19408
788 (signal) Gets a garbage code value when given a non-system error table
789 code with -code.
790 
791 313  none
792 (value_) Gets not_in_read_bracket when value seg contains no values.
793 See 05/29/85 mail from Noah Davids in misc_commands.sv.mbx.
794 
795 309  phx19322
796 terminate_segno does not check that that the segment number is within
797 the range of non-hardcore segment numbers used in the process.
798 
799 
800 308  phx18884
801 (delete_dir) Given "dd x.*" where x.two has its safety switch on, you
802 get the starname query but not a separate query for x.two.
803 
804 307  phx19128
805 (walk_subtree) Shouldn't walk through extended objects implemented as
806 directories.  Also add -inase to allow walk through.
807 
808 305  phx19192
809 (exec_com) &goto from a nested handler does not properly exit.
810 
811 
812 297  none
813 (exec_com) Null &else clause makes parser forget it's in a nested &if.
814 
815 294  phx18806
816 (validate_info_seg) Doesn't complain if multiple paragraphs between
817 double blank lines add up to more than 15 lines.
818 
819 287  phx18651
820 (date_deleter) Should return et$nomatch if no entries match a starname.
821 
822 279  phx18522
823 (exec_com) Traces &set lines with quotes incorrectly.  To be fixed in
824 &version 3.
825 
826 
827 277  phx18455
828 (get_quota) Given a starname, treats MSF's as directories.  Should
829 ignore them.
830 
831 273  phx18264
832 (exec_com) Extra quotes on value of &r2.  This problem is solved, but
833 TR remains open pending fix of minor &trace problem that omits some
834 text.
835 
836 270  phx18171
837 (date_deleter) Does not respect the safety switch.  Suggest
838 -query_protected (default), -delete_protected, and -skip_protected.
839 
840 266  phx18142
841 (where) Takes fault referencing an initiated seg deleted by another
842 process.
843 
844 
845 248  phx17505
846 (exec_com) Gets end-of-data executing a certain teco program.  Have
847 asked for shorter test case.
848 
849 247  phx17505
850 (exec_com) Problem feeding teco a control that causes it to get
851 end-of-data.
852 
853 236  phx17200
854 (exec_com) Error expanding &if &[...] nested inside &if-&then-&do.
855 
856 227  none
857 Commands that append default suffixes to pathname args (eg.  pl1, rdm)
858 often print the un-suffixed name for "Entry name too long."
859 
860 
861 210  phx16271 phx16419
862 When a user tries shortest_path on a component pathname, it returns the
863 input arguments with nothing done to them.  This is actually a problem in the
864 supporting subr.
865 
866 206  phx05178
867 (where) "wh hphcs_" gets moderr but "wh hphcs_ -all" works fine.  Also,
868 moderr error msg when correct should print full pathname.
869 
870 205  phx07909
871 (create_area) message describing location should not be printed by
872 default
873 
874 
875 204  phx11484
876 (print_bind_map) prints a line of octal zeroes for the pathname of a
877 nonexistent bindfile
878 
879 203  phx03385 phx13618
880 (logout) should leave ips signals masked after detaching I/O switches
881 
882 202  phx03255
883 (logout) absentee logout message is printed too early (before
884 signalling finish, etc.)
885 
886 183  phx15989
887 (delete) With -query_all, should list which have safety switch on and
888 omit the subsequent individual queries.
889 
890 
891 169  phx15640
892 (switch_on) Won't set the damaged switch of a directory.  - Not an
893 error in the command since damaged_switch is meaningless for dirs.
894 Forwarded to Ranzenbach to fix set.pl1 not to set damaged switch on a
895 dir.
896 
897 159  phx15386
898 (exec_com) "&trace &control on &command off" should work like you
899 expect.
900 
901 127  phx14665
902 (rename) -name isn't calling expand_pathname_, also bug in error
903 message.
904 
905 
906 123  phx14570
907 Problem description:  The "value_list" command no longer accepts
908  long starnames.  For example:
909 
910    ! vls -match this_sure_is_a_long_name.this_one_is_too.*.how_about_that -var
911      value_list:  Invalid syntax in star name.
912 
913 Proposed solution:  Long starnames will be accepted.
914 
915 The "check_star_name_" system subroutine will be called to handle
916 starnames.  The control mask contains CHECK_STAR_IGNORE_LENGTH to
917 accept long starnames.
918 
919 
920 119  phx14527
921 (exec_com) <whitespace>&+ doesn't work.  Found a spurious char before
922 the &+.
923 
924 114  phx14401
925 (qedx) Now emits extra blank line after "gp/.../" output.
926 
927 95  phx02582
928 (status) Active function error for "Segment contents not in VTOC".
929 Should work at least when called with -dsw.
930 
931 82  phx14082
932 (value_) value_set adds new value rather than replacing old one.
933 
934 
935 71  MIT
936 (where) Should say which bound component if >1 match for entry point
937 name.  (JSLove.PDO at MIT)
938 
939 65  phx13256
940 (value_get) Locking problem, see TR text.
941 
942 17  phx09397
943 (command_query_) Should ignore blank lines.