Error with notification count

[JonoH]

It's currently being investigated by the vendor, but I'm off now for a couple of weeks. I may still be around but I'm not going to be reading my emails for updates. On my return I'll let you know where we are Thanks all

[jab1]

No problem @JonoH - it's with Khorus, so we wait for them.

[Jaggies]

Notification count was at just 1 following @jab1 respnding to my post.
I'll see if anything changes when I next visit the Forum.

[jab1]

@Jaggies The error only appears the first time I open the forum in a morning - I am permanently logged in.

[Townman]

This might be cookie related but not a cookie error as such.

Empirical observations suggest this is related to session keep alive logic if you’ve ticked “stay connected” on the logon screen. Leave the session quiescent beyond the normal session time out … and after first page navigation on return to the session, the notification count becomes 101 or some other static number. Visiting the notification page corrects the count … until next time.

Users not using keep alive are (I believe) not seeing the issue.

[jab1]

Sounds very possible, @Townman . What is the 'normal' session time-out - I seem to remember getting the '101' after I had been away 4.1/2 - 5 hours on Thursday last week?

[Townman]

Hi John,

I asked that question a good few posts back.

I recall your report of seeing the issue after 4+ hours.  That would seem like a reasonable duration.  Way back when in days I used to build .asp web applications I had diagnostic tools to find out the session keep alive time (amongst other helpful stuff) but these days I would not know where to start!

[jab1]

OK, @Townman - I've been quite busy (both on this forum and elsewhere) and my memory ain't what it used to be.

Still think it's PN's 'vendors' playing silly mares, and not checking 'updates' are functioning correctly before deployment.

[Townman]

In complex software it is neigh impossible to check every logic path, particularly those which are depended on long do-nothing scenarios!  Add to that an expectation that users will log out when they have finished ... and the test lab gets nowhere near the behaviours of real users.  Such is often why user experiences are not reproducible in the test lab.

In years past, as an end user testing a new software release, I have had the head of testing come storing into the test lab for a defect report I raised reporting that I could overwrite a write protected file network file.  Simply not possible ... until I illustrated it for him.  A scenario never likely to be considered in "normal" testing - open the file on one device, close it, change it to read only on another, reopen it on the first device ... and write all over it!!!!

[Mook]

---

@Townman wrote:

In complex software it is neigh impossible to check every logic path, ...

---

Sorry @Townman I'd disagree with that there are static analysis tools that would do that for them. 

[Townman]

Would that include complex’s time dependent logic paths.

[jab1]

What ever was wrong seems to have been corrected. I have just opened the forums on both the laptops and the iPad, with no spurious notification count and no spinning dots - two problems resolved in one.

I'll still monitor it though.

[Mook]

It may be working for you @jab1 but certainly not for me, as you can see here:

[jab1]

OK, @Mook . Sorry you still have the problem, but note I'm still not 100% convinced, hence the 'I will continue to monitor'.

EDIT: Still seeing the 'dots' now I have come back.

[Mook]

You don't need to apologise, it's hardly your fault, but I have to admit I don't see the 'dots' you are referring to up by my avatar, but I do get this:

on my notification feed page.