Re: v.11 Desktop Anomalies & bug list (legitimate contributions welcome from all)
DROPPED TABLES - SETS BASED ON DROPPED TABLES DON'T KNOW THE DIFFERENCE & CONTINUE TO WORK AS EXPECTED!
If this isn't a bug, it certainly "compromises" the reliability & usefulness of the backup/zip routines built into a5!
(They don't backup these dropped tables.)
v.11 (Build 3211 Addins: 4072) ~ 01/24/2013
History: Unknown.
Status: DOCUMENTED & SUBMITTED
If you drop a table (accidentally), any set relying on that table can't tell the difference. (And consequently, all layouts based on that set will continue to run even though a table in the set has been dropped and is presumably missing. ~ It definitely doesn't show up in the CP! In most instances, (when you drop tables) any reference to any layout object (aka form report etc.) that was "tied" to the dropped table will cause an error when running the db. (when code references the dropped layout object) This is not the case with sets.
Here's the real problem/threat: Sending, backing up & zipping the database that has a table that was dropped inadvertently creates a broken copy with a missing table! Make a backup (using the a5 built in tools) OF A WORKING DATABASE, and the backup is going to be missing those tables. (Because a5 backups don't backup anything that is "dropped" from the CP.) The backup database won't work because of the missing table (that was merely dropped in the original), and there is no way to fix it/get it back! (Unless you have the original somewhere which still has the dropped objects sitting in the directory.)
Note: I've always thought (and suggested several times) that the backup/zip routine (by default) should provide the ability to back up dropped objects! ~ I had recommended that Alpha added a checkbox to the zip/backup routine which is by default, "checked." (I submitted this as a bug/suggestion/request several years ago.) And while this wouldn't fix the issue of sets not realizing a table has been dropped, it would ensure (by default) that you get a backup of the entire "working" database. (A backup including the dropped table which you could re-add.)
It seems to me, the backup/zip routines need to be modified as suggested! Ultimately, this scenario renders the zip/backup routines useless!
Because of this "problem", I'll bet there are a few zipped databases "out there" that are useless without the originals! (And nobody will realize this until a catastrophic failure occurs & they need to restore from scratch!)
DROPPED TABLES - SETS BASED ON DROPPED TABLES DON'T KNOW THE DIFFERENCE & CONTINUE TO WORK AS EXPECTED!
If this isn't a bug, it certainly "compromises" the reliability & usefulness of the backup/zip routines built into a5!
(They don't backup these dropped tables.)
v.11 (Build 3211 Addins: 4072) ~ 01/24/2013
History: Unknown.
Status: DOCUMENTED & SUBMITTED
If you drop a table (accidentally), any set relying on that table can't tell the difference. (And consequently, all layouts based on that set will continue to run even though a table in the set has been dropped and is presumably missing. ~ It definitely doesn't show up in the CP! In most instances, (when you drop tables) any reference to any layout object (aka form report etc.) that was "tied" to the dropped table will cause an error when running the db. (when code references the dropped layout object) This is not the case with sets.
Here's the real problem/threat: Sending, backing up & zipping the database that has a table that was dropped inadvertently creates a broken copy with a missing table! Make a backup (using the a5 built in tools) OF A WORKING DATABASE, and the backup is going to be missing those tables. (Because a5 backups don't backup anything that is "dropped" from the CP.) The backup database won't work because of the missing table (that was merely dropped in the original), and there is no way to fix it/get it back! (Unless you have the original somewhere which still has the dropped objects sitting in the directory.)
Note: I've always thought (and suggested several times) that the backup/zip routine (by default) should provide the ability to back up dropped objects! ~ I had recommended that Alpha added a checkbox to the zip/backup routine which is by default, "checked." (I submitted this as a bug/suggestion/request several years ago.) And while this wouldn't fix the issue of sets not realizing a table has been dropped, it would ensure (by default) that you get a backup of the entire "working" database. (A backup including the dropped table which you could re-add.)
It seems to me, the backup/zip routines need to be modified as suggested! Ultimately, this scenario renders the zip/backup routines useless!
Because of this "problem", I'll bet there are a few zipped databases "out there" that are useless without the originals! (And nobody will realize this until a catastrophic failure occurs & they need to restore from scratch!)
Comment