Very weird behavior when selecting “From” email address

We have a Universal Distribution Group entry with an email address – we call it

This item has two members [we call them MemberA and MemberB] which receive in their inbox emails sent to That far this works perfectly for years.

When for example MemberA wants to reply to an email sent to [which has been delivered into his Inbox] then I have to change the sender address in the From field [ is selected from the place displayed as red box]:

enter image description here

Since couple of weeks [I suspect a Microsoft update] I have this very weird behavior:
Very often, too often, the email is rejected by our Exchange with the name of the recipient and that I am not allowed to send on behalf of an other user – but we are not sending it on behalf of!

When we go to the Sent Items folder, press Forward Email, select in the From field then it works!

Here two screenshots how it appears in the Sent Items box:
1) The first try when the email gets rejected. It is in German but you can see it has been sent on behalf of.

enter image description here

And then few seconds later, with exactly the same procedure and same data the email can be sent – you see the on behalf of is missing and therefore it works.

enter image description here

Note: I replaced in the screenshots the real content with samples so it is more understandable.

What confuses me totally is, that mostly the first time the email tries to be sent on behalf of and then few seconds later not.

My question:

What can I do to disable/avoid the on behalf of mechanism?

  • Exchange 2010
  • SBS 2011 Standard
  • Outlook 2013

Enable dustRelayFee behavior on testnet to get error code 64: dust

Crafting and sending an utxo of 1 satoshi does works and get mined on testnet without error.
However on mainnet the error code 64: dust arise.

I understand the basic idea that you are not able to broadcast a transaction containing an utxo with a value so low that spending it(the low utxo) would not be enough to pay its tx fees. What is meant by Bitcoin dust?

So naturally I’m looking for setting this behavior in testnet in order to implement a proper way to handle this case but starting my testnet node with bitcoind -datadir=/Volumes/EXT_SD/bitcoin -dustrelayfee=0.00003 still allow me to send an utxo of 1 satoshi 1 sat ex, who can I setup this ?

behavior of path on unpublished/draft nodes

We have a Drupal 8 site with a press releases page. Currently the press releases page is provided by a view with a page display. The path in the page display is set to /press-releases.

I’ve been tasked with changing the design of the press releases page. We want to use paragraphs to allow the marketing department to edit the page. This means that I must first disable the view’s press releases page display (to free up the path), then create a Basic Page (which we’ve added a paragraphs field to) and set its path to /press-releases.

We use configuration export, so in my feature branch, I’ve set the “Press Releases” view page display to be disabled when the config is imported.

To speed deployment, I’ve already created the “Press Releases” Basic Page on production, along with all the paragraphs that we want on it. Of course, it’s not published (its current state is “Draft”, which in Drupal 8, I’m not entirely certain which module provides those workflow states), because we don’t want it live yet.

We use the pathauto module, so when I created the “Press Releases” Basic Page, it automatically generated the path /press-releases-0, presumably because the /press-releases path was already in use by views.

This means that, as a post-deploy step, in addition to publishing the new Press Releases node, I also have to edit the URL alias, from /press-releases-0 to /press-release (the cim command deactivates the view display).

I tried changing the URL alias on the draft node to /press-releases before it was published, but what ended up happening was that on the site, the /press-releases path returned a 404. I confirmed this by looking in the watchdog log, and, sure enough, I got an access denied at /press-releases. Finally, I published the node and cleared the cache. When I reloaded the URL, I got the Press Releases node. So it looks like the node is taking control of that path.

So my question is, is this the expected the behavior of nodes (or entities that take paths in general) in Drupal 8? And that it does this without a form validation error, or a warning or notice that an existing path was overridden? Or is this something that could be being caused by Pathauto or the Draft workflow state?

What concerns me is that any one of our editors (or even myself, as careful as I try to be) could override the URL of a view (or perhaps anything else), simply by specifying an alias for a new node.

(Also, I would prefer to be able to just publish the page in a post-deploy step. Since our deploy process involves deploying to a test and a staging environment before production, extra deploy steps are multiplied.)

Can I combine multiple “no results behavior” to avoid repetition?

I have a Search API view that has 3 attachments. This is so I can display different content type results with unique layouts.

If the user searches something too specific, they’ll see the No Results Behavior up to 4 times in a row. This looks bad.

How can I only display the No Results Behavior (custom textarea) when there are no results in the view AND its 3 attachments?

Is Behavior Driven Development Worth It and Plausible

Test-driven and behavior-driven development seem to appeal to me. Perhaps it is because I like to have things organized and neatly done. I also like to be confident in the conclusions I make, which spills over to the products I put out as well. Do not confuse this with OCD.

However, I am wondering if it is actually plausible to use BDD or TDD for some rapid development. Suppose I would like to build a site. I am aware of some features it must have in the beginning. However, I am also aware that it will grow. New pages and functionality will be added. On the one hand, it seems prudent to write some functional and unit tests. This way, I will feel more confident in the product since I am currently the only one working on it. It will get to a point where it is impossible to notice all consequences of your changes without automated tests.

On the other hand, it appears to be an overkill to do this for a relatively basic app. I know plenty of people who would just take a few weeks and use their nights and weekends to build it quickly without tests. What happens after that in this case I do not know because I have not been there. My experience with TDD has been in places where the software is so complex that one would not even think of making commit without testing or writing a unit-test first.

As a more general question, is BDD a utopia or does it actually make one (or organization) more productive?

What has your experience been? What is the line one should stride and how does he find it?

Much appreciated.

Update products from CSV using import behavior Update only not add new

Can we update products from CSV but skip new products just update products that already exist. Like I have a CSV which has data with skus some already exist some are new but I want to update products not to add a new ones when I use import behavior add/update

enter image description here

it updates products but also add new products from new skus.

Why does C# type pattern matching use a different variable scoping behavior than traditional switch blocks?

Traditional switch blocks have one scope, so the following throws a compiler error “A local variable or function named ‘message’ is already defined in this scope”:

switch(value) {     case 1:         string message = "Val: 1";         break;     case 2:          string message = "Val: 2";         break; } 

As Eric Lippert states:

A reasonable question is “why is this not legal?” A reasonable answer is “well, why should it be”? You can have it one of two ways. Either this is legal:

switch(y) {     case 1:  int x = 123; ... break;     case 2:  int x = 456; ... break; } 

or this is legal:

switch(y) {     case 1:  int x = 123; ... break;     case 2:  x = 456; ... break; } 

but you can’t have it both ways. The designers of C# chose the second way as seeming to be the more natural way to do it.

There are other good explanations too, like this one:

I think a good reason is that in every other case, the scope of a “normal” local variable is a block delimited by braces ({}).

So then why does scoping behave differently with a type pattern matching switch block?

Animal p = new Dog();  switch(p) {     case Dog a:         break;     case Cat a: // Why is this legal?                    break; } 

How to fix Thinkpad FnLock LED behavior?

On my A485 running 18.04, I noticed that when I press Fn+Esc, it correctly sets the alternate function key behavior but the FnLock indicator LED on the Esc key doesn’t turn on. How can I set the FnLock LED to light up on Fn+Esc?

I noticed the intended behavior when enabling the BIOS setting for “Fn Sticky keys”, in which the LED correctly turns on when it should.

I tried acpi_listen, but it doesn’t show any events happening when pressing Fn or Fn+Esc (with/without Fn sticky keys), even if it does for the other function keys.

Is there a way to get it to light up without having to use Fn Sticky keys?

Writing both behavior tests and unit tests? [on hold]

Most I’ve read said that behavior tests should not be testing for exceptions because that’s the job of unit test. I am currently learning Cucumber. But the feature tests in Cucumber doesn’t seem to allow for testing exceptions, and I was thinking whether I should add regular unit tests in Junit to cover the exception cases that can’t be tested in Cucumber, this would mean I have a file for behavior tests and a file for unit tests.

Is this a good idea? Or should I still test exceptions in Cucumber step definition class by catching the exception and then failing the test manually?