bug fix report

Danielle et jean Thierry-Mieg mieg at kaa.crbm.cnrs-mop.fr
Tue Apr 30 10:55:11 EST 1996


Final test fro distribution 4.3

Test was ran under 4.3h
 I found two missing feature that i added to 4.3i
 The constraints which was missing in 4.1 and is only supported in parse.c, 
 The new feature is to be able to follow complex-questions, 
	i.e. Follow A#B   or  Follow A = "aaa"
 Everything now works correctly as far as I know.


Detlef found after release 4.1 that the main remaining problems were:
1)
the query used in IGD for single-chromosome subset selection is
still not parsed, nor does the query
2)
where find Probe probe_clone_data = "vector_type" AND NEXT = "Plasmid"
work.

Then he gave a  list of current problems sorted by priority. I checked
all of then in 4.3 and added in his lists some comments

-------------------------------------------------------
:NAME:    tagvqry                 :PRIORITY: 5
:SUMMARY: queries on values after tags

works in 4.3h
Example:
acedb> clear
// 0 Active Objects 
acedb> query Find n R = aaa AND NEXT = bbb

// Found 1 objects
// 1 Active Objects 
acedb> show

N : "nn1"
  R   aaa     bbb

A better syntax now is:

acedb>  query Find n R:2 = bbb
// Found 1 objects
-------------------------------------------------------
:NAME:    parsubq                 :PRIORITY: 4
:SUMMARY: parameter substitution in queries in complex cases

works in 4.3h
-------------------------------------------------------
:NAME:    abnmqry                 :PRIORITY: 3
:SUMMARY: find M "<someth>" should find the M object called <someth>

works in 4.3h
-------------------------------------------------------
:NAME:    cttab                 :PRIORITY: 3
:SUMMARY: constructed types and parameter passing in table definitions

works in 4.3h
-------------------------------------------------------
:NAME:    nnmqry                 :PRIORITY: 3
:SUMMARY: queries for objects whose name is a number

works in 4.3h
-------------------------------------------------------
:NAME:    aqdump                 :PRIORITY: 2
:SUMMARY: turn attached query results on/off in write command

works in 4.3h
-------------------------------------------------------
:NAME:    atqry                 :PRIORITY: 2
:SUMMARY: attached queries from root of objects; show attached data

works in 4.3h
example:
acedb> find A1

// Found 1 objects in this class
// 1 Active Objects 
acedb> show

A1 : "a1"
  T   a2
  U
  V   W   a2
  X   a4
  Y

// 1 object dumped
// 1 Active Objects 
acedb> show -a

A1 : "a1"
T        "a2" "this should be picked up by the attached query"
T        "a2" "multi-valued should work"
T        "a2"
U        "just to make sure that only part of the a2 object is attached"
U        "this should be displayed by the U @'>T\; B' query in model of A1"
U
V        W "a2" "this checks attach from within a constructed type"
V        W "a2"
X        "a4" "a2"
X        "a4"
//!! Internal error : condMatch should not parse FIND
Y   // This crash is correct, you have to attach locally, not a whole class.

// 1 object dumped
// 1 Active Objects 
acedb> 

-------------------------------------------------------
:NAME:    cnstrain                 :PRIORITY: 2
:SUMMARY: constraints should be evaluated when loading data

works in 4.3i
 was nullified in 4.1 except on read new models, actually
we do not use it for nematode or anywhere else that i know
-------------------------------------------------------
:NAME:    dblqot                 :PRIORITY: 2
:SUMMARY: double quotes in data and names and queries

works in 4.3h
example:
acedb> parse
// Type in your data in .ace format, then CTRL-D to finish
N  "quo\"ed"
M  "bizre{ data\"\\"

// 1 objects read with 0 errors// 1 Active Objects 
acedb>  find N  quo"ed

// Found 1 objects in this class
// 1 Active Objects 
acedb> show

N : "quo\"ed"
  M   bizre{ data"\

// 1 object dumped

-------------------------------------------------------
:NAME:    ppcont                 :PRIORITY: 2
:SUMMARY: loading erroneous data with pparse

works in 4.3
-------------------------------------------------------
:NAME:    followct                 :PRIORITY: 1
:SUMMARY: query that uses follow into a constructed type: follow A#B

works in 4.3i 
 i had forgotten that one, i just fixed it now.
This was never in the acedoc, but i remember Otto wanted it for a long time.

The same system now allows:

N : "nn1"
  R   aaa     bbb
  M   mm1     cc1
      mm2     cc2
              cc3

acedb> query Follow M = mm1 // -> cc1
acedb> query Follow M = mm2 // -> cc2 , cc3


-------------------------------------------------------
:NAME: tables
:SUMMARY: some tests on complex tables

 works in 4.3h
 you must "" otherwise the text a is mixed with the tag a
 the correct condition is "%3" = "%2"

-------------------------------------------------------


Everything else in Detlef script is ok except:

/* :SUMMARY:  "help query" shows query help text
:PRIORITY: 1
:STATUS: not clear
:FEATURE:
:HISTORY: 
:MODEL: 
:DATA: 
:RESULT: 
fails in: ace950522a4n --
    "help query" gives wrong result on first invocation, works on subsequent

:SCRIPT: */
help: dbinit wspec.mini // help system
  tace
    help1: do -e help query
    help2: do -e help query
  exit

Comment: I cannot reproduce the bug outside of aquila.

******

/* :SUMMARY: model command does not work for renamed classes
:PRIORITY: 1
:STATUS: fail, 

Comment: but is it a bug ? the local model file should be remaned

******

/* :SUMMARY: constructed types and parameter passing in table definitions
:PRIORITY: 3
:STATUS: nearly ok, 

comment: The table definition should say Next_Key rather than text 
	to access a a#b in table. This is a half bug, because it works but the
        interface is not natural.

******

CONCLUSION:
No remaining bug:






More information about the Acedb mailing list