Serious anomalies with in-place edit and delete for attachments

R6’s in-place editing and deletion of attachments, while a welcome innovation, has created serious problems for certain application designs. In Notes 4/5, you could prevent inadvertent editing or deleting of attachments under specified conditions by using QueryOpen and QueryModeChange to prevent a document from going into edit mode.

Unfortunately, due to some glaring anomalies in R6, it is now much more difficult to provide this protection. I thought that it would be useful to document these for the benefit of anyone else struggling with the same problems. (These results were obtained on version 6.5).

Anomaly 1.

Note in read mode. Right-click attachment and select Edit. Querymodechange fires, but script sets Continue = False (because application designer does not want to permit editing).

Result: note stays in read mode, but attachment is still opened for editing in registered application. After editing attachment, user closes note. Asked whether to save changes (and querysave fires - this could never happen from read mode before). If answer is yes, error message says “You must be in edit mode to attach files to this document”. On the positive side, this prevents modification of the attachment, but it does mean that user’s editing efforts have been wasted.

Preferred behaviour: if Continue = False in querymodechange, attachment should not be opened for editing.

Anomaly 2.

Note in read mode. Right-click attachment and select Delete. Querymodechange fires, but script sets Continue = False for same reason.

Result: note stays in read mode, but attachment is deleted. If note is closed, user is asked whether to save changes (and querysave fires). If answer = Yes, note is saved with attachment removed. In this case, therefore, attachments have not been protected from deletion.

Preferred behaviour: if Continue = False in querymodechange, no attempt should be made to delete the attachment.

Anomaly 3.

Same situation as anomaly 2, except that attachment is double-clicked to open Attachment Properties and Delete is selected from there.

Result: note goes into edit mode, without firing querymodechange or postmodechange!!! Attachment is also deleted. It is now possible to edit anything on the note, with the application designer having no way of detecting it has happened.

In considering possible solutions, you encounter several problems.

Anomaly 1 is difficult to avoid, except by providing a warning in querymodechange. Looks very inelegant from user viewpoint.

Anomaly 2&3 needs a Continue = False in querysave. This is not sufficient however. You must be absolutely sure that nothing will cause a backend save between the time that you pressed delete and the document is closed again. Otherwise you are left with an attachment icon and no $File.

Also, if the note goes into edit mode due to anomaly 3, the user might be able to change the conditions under which editing is not permitted, so querysave protection could be lost.

If anyone can think of better solutions, please let us know. And if anyone is in a better position than me to feed these points to Lotus, please do so. They really must be fixed.

Subject: Any update on this?

Subject: Would someone like to submit this to Lotus?

As we are still only evaluating R6, we do not currently have a support contract. Perhaps someone else who is concerned by these issues could submit them?

Subject: RE: Would someone like to submit this to Lotus?

I have submitted this related issue:

http://www-10.lotus.com/ldd/nd6forum.nsf/55c38d716d632d9b8525689b005ba1c0/47cf03a5bbdcb0a285256e83003f942f?OpenDocument

Subject: We have now submitted these issues to Lotus …

and they have assigned SPR numbers to them. They seem to acknowledge the problem, but give no immediate indication of what they might do about it.

Subject: RE: We have now submitted these issues to Lotus …

Can you post the SPR in here so we can view its progress / send support of the fact this is a big issue!!

Subject: Progress on in-place edit anomalies

Just decided to check what progress has been made on these anomalies.

When we submitted this problem a PMR was assigned, but I cannot find any reference to it when I search. They told us that an SPR had been assigned, but would not show us what it said - they said SPRs were confidential documents.

However they did say that the problem was similar to SPR SIWA5LG2AN and ROHI5J9CV4. Both are supposedly fixed in release 6.5.4.

I have just checked the behaviour of 6.5.4 against my original list of anomalies, with the following results.

Anomaly 1 - unchanged.

Anomaly 2 - behaviour is different, but it is still possible to delete an attachment when querymodechange = false. Anomaly still exists. Comment in the SPR says ‘querymodechange should not be used as a security feature’. Perhaps not, but it was useful as a way of preventing inadvertent editing under particular conditions, but is rather less useful now. Its behaviour should at least be consistent.

Anomaly 3. Improved. No longer goes into edit mode. Now the same as anomaly 2.

I do not have access to version 7 yet, so do not know whether there have been any improvements there.

Subject: Oh-oh… Serious anomalies with in-place edit and delete for attachments

Serios Anomaly #4 (related to 3): OLE object embedded on form loads on open, to fill Notes window. Essential to write information into object cell when doc changes from Edit to Read mode. Ctrl-E flops between modes without Query/PostModeChange firing.

The behavior is exhibited in the MS Object Library which ships with Notes (if any of you cats at Iris are listening in)(and who has dared say “cats” for the last 20 years?).

anyway, we do have a support contract, and I have sent this in – don’t think we have any resources for the near term except our own wits, tho’. Best of luck! (Cats)

Subject: Looks like serious bugs - please contact Lotus Support.