Now Wave2 blogging is allowed, there are many topics that I want to share with you .. really, a lot. Time is not on my side – I’m in full preparation for Directions and NAVTechDays, so let’s see.

Today, while working on my SaaS-deployment-pipeline in DevOps with ALOps, I DID find the time to say a few words on one of the topics I’m quite excited about: Shortcutkeys in the webclient, or in other words: full keyboard shortcut support!

In my “real world”, users (yes, I do have customers, although people sometimes seem to be thinking otherwise.. ) have been asking this for quite some time .. and finally I can tell them: it works like it should!

Obviously you already knew about this, because I already announced it here (and specifically here by Microsoft) – just kidding ;-).

How it works?

Very simple: you have a “ShortCutKey” property on action, where you can set the ShortCutKey like this example:

I guess that doesn’t need too much of explanation, does it ;-)? You can simply do it the same for your actions. And I can tell you: it works! :-).

Knowing this, let’s dive a bit into what Microsoft did…

Did Microsoft apply it to the Base Application?

Absolutely. Even more, at the current version (a bit depending on localization), I found 1370 places where Microsoft added a ShortCut! I can only imagine they were able to convert that from the Base Application ;-).

This made me think though – are there THAT many shortcuts we need to start learning by heart?

I decided to dive a little bit more into it, by analysing all the shortcuts by creating a PowerShell script that loops all al files of the Base Application, and lists all ShortCutKey properties in a csv file. You can find the script here: https://github.com/waldo1001/blog.CALAnalysis/blob/master/ShortCutKeys/GetShortCutKeys.ps1

You can find the output in that same repo, but here is the overview of all shortcuts that I found are implemented:

As you see, not really THAT many, but very useful nevertheless..

Now obviously the next question:

Can we change a ShortCutKey from within an app?

And yes you can! So, if Microsoft has forgotten any, you can add them! Here is an example:

As you can see, in this case, I didn’t just add a ShortCutKey, but I even changed the “ALT+D” to something else. And it works .. . Not sure we should all be doing this – and also not sure if everyone would do this, what would happen with conflicting shortcutkeys (although, I guess the “first of last one wins”?), but now at least, you know it’s possible!

If you want to play with Wave 2, get started here:

https://freddysblog.com/2019/07/31/preview-of-dynamics-365-business-central-2019-release-wave-2/

I’m returning after a very interesting workshop on DevOps in Berlin .. . At this moment, I’m wasting my time in the very “attractive” TXL airport because of my delayed flight. And how can I better waste my time than to figure out some stuff regarding Business Central.

Figuring out indeed, because I have barely internet, a crappy seat, nearly no access to food, … so for me this is a matter of bury myself so I don’t have to pay attention to my surroundings ;-). Anyway .. It’s not all that bad .. but a delayed flight is never nice. Anyway…..

opic of today: the new systemId!

While I was converting our app to be able to publish on the Wave 2 release .. this was something that I noticed:

All the “integration id’s” are marked for removal, and – most interesting – will be replaced by “the systemID”. What is that? Google, Microsoft Docs .. none of my conventional resources helped me finding out what SystemID is .. luckily, I did come across some information on yammer by Microsoft ;-)..

RecordIDs

You probably all know RecordIDs, right? A single “value” that referred to a specific record in a specific table. We all used them in generic scenarios, right? Also Microsoft – I don’t know if you know “Record Links”? A system table that stores notes and links to specific records? Well, the link to the record is made through RecordID. We have been using it for years .. . Now, a big downside of using RecordIds was the fact when you would rename the record (one of the fields of the keys), it would change its RecordId as well .. and all of a sudden, you could lose the connection in all tables where you stored that specific ID. Long story short – not ideal for integration or generic scenarios…

Surrogate Keys

And this is where “surrogate keys” of my good friend Soren Klemmensen came into place. He came up with a design pattern (well, I don’t know if he came up with it – but he sure advocated it for a long time) that described how to implement having a dedicated unique key of one field for a record. Basically: add a field in a table, and make sure it has a unique GUID. Make it that all these surrogate keys have the same FieldNo, and you are able to generically access the value of any of the keys for any record.

This is something Microsoft actually implemented themselves. And the code is all over the place. Even still in Wave2, we have the code to fill the “Integration Ids” as they call it. Nice system, but a lot of plumbing needed to make it work. I don’t know if there was a design pattern that described what you needed to do to apply this on your own tables – I never did ;-). But definitely interesting to do for many scenarios. Thing is .. quite a lot of work.

The SystemID

Now, as you got for the first screenshot: Microsoft is abandoning this field 8000 (that so-called “integration id”) – their first implementation of the surrogate keys – and will implement “SystemId” from the platform. Meaning: whatever you do: you will ALWAYS have a key called “systemId” for your table, which is a unique GUID in that table that can identify your record, and will never be changed – even when you would rename your record.

How cool is that! Here is an example of a totally useless table I created to show you that I have the systemId in intellisense:

What can we expect from the systemId?

Well, in my understanding – and quite literally what I got from Microsoft (thanks, Nikola ):

• It exists on every record
• But not on virtual/system tables (not yet, at least)
• You can even set it in rare scenarios where you want to have the same value (e.g. copy from one table to another, upgrade…). Simply assign System Id to the record and do Insert(true,true) – 2x true
• There is a new keyword – GetBySystemId to fetch by system id
• It is unique per table, not per DB. Customers and items may have same IDs, though is hard if you are not manipulating it yourself, since guids are unique. Let’s say, they are “probably” unique – but on SQL, there is a unique key defined on the field, so only guaranteed per table.
• Integration Record is still there, however the Id of the Integration Record matches the SystemId of the main record (Microsoft has code and upgrade in place)
• You can only have simple APIs on it (no nesting, like lines). At this point, at least. It should be fixed soon, which is why the APIs are not refactored yet to use SystemId instead of Id.

A few more remarks

IF you would create a field that refers to a systemId, then it makes sense to use the DataClassification “SystemMetadata” for it. Not because I say .. just but because I noticed Microsoft does ;-).

Another not unimportant something I noticed: this is a system-generated field. So if you would need the fieldnumber, you have “recref.SystemIdNo”:

My take on it

From what I understood: there is work to do, but things are looking good:-). In fact, it is exactly what we have been asking for – and Microsoft delivers. Again! Great! I know this will see a lot of use in the (near) future! Within the Base Application, and in lots of apps.

Do know, I didn’t have any documentation about this – so all is based on some small piece of remark on yammer, and things I could see in code… So – if you have anything to add – please don’t hold back ;-). That’s why I have a comment section ;-).

Small message that I just needed to share :-).  As said – I’m prepping my sessions for Directions, which basically means: I’m spending all my free time in “Business Central” these days.  No printing, hardly any social contacts … . 

And while publishing to SaaS, I noticed this when I refreshed my page:

That’s right!  A new logo, people!  (I should say “Icon” actually, but you know what I mean).  Let’s have a closer look:

Doesn’t look bad! Not bad at all .. But it does mean I’ll have to redesign my 3D printed logo ;-).  And I will … I so will .. .  As I said on twitter earlier today: I’m so distracted now that I have to first make sure that I started a new print with a new concept before I can continue prepping my sessions :-).

Sorry for the shortness of this blog (may be you like it that way ;-)) – but just a small reminder for anybody that has been sleeping under a rock for the last couple of days:

Microsoft Dynamics 365 Business Central 2019 release Wave 2 is released!

All you need to know is simply quite well documented by Microsoft.  Let me give you a few links:

• Dynamics 365: 2019 release Wave 2 plan
• What’s new and planned for Dynamics 365 Business Central
• Business Central on Microsoft Docs
• Business Central on Microsoft Learn
• The download link for the OnPrem DVDs

And yes, that’s right: C/AL is gone, and the RTC is gone as well!  But together with that, a lot of goodies are being thrown in our lap!  If you want to know more, read the above links, or come to Directions or NAVTechDays and learn from the very people that built it!

It’s a big one – so build pipelines will break, code needs to be upgraded.  I guess it’s time for action ;-). 

More to come ..

The better half of my past week can be summarized best by this oh-so-descriptive-error-message:

Right: a message I have spent a long time on to find out what was happening – and what caused it.  Multiple days – so let me try to spare you the pain when you would encounter this error.

(tip: if you don’t care about the story, just skip to the conclusion ;-)).

History

We are rebuilding our product to Business Central – and are almost finished.  In fact, we have spent about 500 days building it – and since the recent release of Wave 2, we are fully in the process of upgrading it – because obviously, since all is extensions (we have a collection of 12 dependent extensions), that should be easy.  (think again – Wave 2 came with a lot of breaking changes… but that’s for another blogpost ;-)).

Symptoms

Our DevOps builds had been acting strange for a while – just not “very” strange .. .  In fact: when a build failed with a strange error (yep, the above one), we would just retry, and if ok, we wouldn’t care.

That was a mistake.

Since our move to Wave2 .. the majority of the builds from only 1 of the 12 apps failed – and even (what never happened before), the publish from VSCode failed as well with the same error message:

Insufficient stack to continue executing the program safely. This can happen from having too many functions on the call stack or function on the stack using too much stack space.

We are developing with a team of about 9 developers – so people started to NOT being able to build an environment, or compile and publish anymore.  Sometimes. 

Yes indeed: sometimes.  I had situations where I thought I had a fix, and after 10 builds or publishes – it started to fail again.

And in case you might wonder – the event log didn’t show anything either.  Not a single thing.  Except from the error above as well.

What didn’t help

I started to look at the latest commits we did.  But that was mainly due to the upgrade – stuff we HAD to do because of the breaking changes Microsoft introduced in Wave 2.

Since it failed at the “publish” step, one might think we had an install codeunit that freaked out.  Well, we have quite a few install-codeunits (whenever it makes sense for a certain module in that app) .. I disabled all of them – I even disabled the upgrade-codeunits.  To no avail.

Next, I started to look at the more complex modules in our app, and started to remove them .. Since one of the bigger modules had a huge job during install of the app – AND it publishes and raises events quite heavily, I was quite sure it was that module that caused the pain.  To test it, I removed that folder from VSCode, made the code compile .. and .. things started to work again.  But only shortly.  Moments later, it was clear in DevOps that certain builds started to fail because of the exact same error.  From victory .. back to the drawing board ;-).

Another thing we tried was playing with the memory on our build agents and docker hosts.  Again, to no avail .. that absolutely didn’t help one single byte.

…

And I tried so much more .. really.  I was so desperate that I started to take away code from our app (which we have been building for over 6 months with about 9 developers (not fulltime, don’t worry ;-)).  It’s a whole lot of code – and I don’t know if you ever tried to take away code and make the remaining code work again .. it takes time :-/.  A lot!

What did help

It took so much time, I was desperately seeking help .. and from pure frustration, I turned to Twitter.  I know .. not the best way to get help .. but afterwards, I was quite glad I did ;-).

You can find the entire thread here:

First of all: thanks so much for all of the people for their suggestions.  There were things I didn’t try yet.  There were some references to articles I didn’t find yet.  All these things gave me new inspiration (and hope) .. which was invaluable!  Translation files, recursive functions, event log, dependencies, remove all code, force sync, version numbers, …

Until phenno mentioned this:

Exactly the same error message, with a big xmlport.  It first pointed me to the wrong direction (recursive functions / xmlport) ..

But after one of our developers remembered me that from months back, we also had a big object: A 1.2Mb codeunit, auto generating all Business Central icons as data in a table, to be able to use them as icons in business logic.  Initially I didn’t think it would ever have an effect on the stability of the app (in this case – the inability to publish it) .. we wrote the damn thing more than 4 months back, for crying out loud :-/ and the code was very simple – nothing recursive, no loops, very straight forward.  Just a hellofalot of code ;-).  But .. It doesn’t hurt to try what it would do when I would remove the code .. so I tried .. and it works now!  Victory!

Conclusion

The size of a file (or object) does matter.  If you have the error above – it makes sense to list your biggest files, and see if you can make them smaller by splitting the objects in multiple (if possible.. ). 

While in our case, it was one huge object in one file.  And I don’t know what exactly was the problem: the size of the file, or the size of the object.  There is a difference.  If I wanted to have kept the functionality, I might have had to split the object in multiple codeunits, and on top of that I might have had to split them in multiple files (which – in my honest opinion – is best practice anyway..).

Also, I have the feeling that Wave 2, is a bit more sensitive to these kind of situations.. I don’t know.  It’s just – we had this file for quite a while already, and it’s just with the upgrade to Wave2 that it started to be a problem.

In any case – I hope I won’t wake up tomorrow, concluding the error is back and all the above was just one pile of crap.  Wish me luck ;-).

I got quite a week ahead of me .. .  Not only will I host one session and some workshops .. I will actually host 2 sessions, 2 workshops and an ISV session this year.  What did I get myself into?

No repeats!

If you look at my session schedule, and you have visited Directions EMEA, well, you might wonder if I’m “just” redelivering content at NAVTechDays.  Well .. No!  Totally not, actually. Without giving away anything – let me try to explain …

Development Methodologies for the future (Thursday 11:30 – 13:00)

First of all, if you attended my session at Directions, you noticed that there, I prepared actually 3 sessions, and the audience chose the topic of that particular session.  I was lucky that all three topics that I prepared were quite equally popular while the audience was voting – so I would be stupid to do just a repeat.  No, I will actually slice a completely different topic as I did on Directions EMEA.  All new content – and more ;-).  More details Thursday at 11:30 during my session ;-).

{Connect App}² (Friday 11:00 – 12:30)

My session with Vjeko at Directions was “Connected Apps” .. .  This one is “Squared” ;-).  Which means: more!  Much more! So, If you attended that one on Directions, and you thought we took it “far” – well – think again!  Just to say, also this one is not a repeat.  How could it be?  At Directions, there were only 45 minutes ;-).

Workshops

Also this year, I will be hosting workshop during the predays, which I always look forward to.  I just hope that the internet will be good, because I will be quite dependent on it ;-).  I prepare individual Azure VMs for every attendee to make it as comfortable as at all possible .. but that means: internet! ;-).  What I will be doing is something I have been doing quite a lot …

Developing Extensions for Microsoft Dynamics 365 Business Central – Introduction (Tuesday)

For the people that are putting their first steps into AL development.

Developing Extensions for Microsoft Dynamics 365 Business Central – Advanced (Wednesday)

For the people that already have put their first steps .. but still feel they need some guidance for the “stepping” to feel comfortable (if that’s an explanation at all ;-)).

ALOps (Thursday 15:40 – 15:55)

And if that’s not enough .. I’m doing yet another session .. . This one is an ISV session for the product we have been working so hard on to get to the community: ALOps.  We are a Platinum sponsor, which comes with an ISV slot – and I’m looking forward to speak to all people that are interested in “doing decent DevOps for AL easily” ;-).  We will obviously also have a booth at the EXPO – please come by!  We have stickers ;-). And we might just get your pipeline upt and running .. during the conference ;-).

In total, that means I have about 19 hours and 15 minutes of content to deliver on NAVTechDays … .  Again .. what did I get myself into :-/.

It’s over – the week I’m always looking forward to for so long passes by in a blink of an eye: NAVTechDays.  As Vjeko already shared the goodies– I will do so as well – joined with my final thoughts and some pictures ;-).

I feel myself old and repetitive by saying this conference is “something else”.  Just imagine: quality food: morning, noon and evening, quality (recorded) sessions – all available to the community, 90 minute deep-dive topics, in quality seats, with quality media equipment, and quality speakers (uhum ;-)) – all captured by a quality photographer.  Quality! That’s NAVTechDays: no dime is spared to provide THE BEST conference experience anyone could wish for.  From start to finish!  Unmatched on any level. This year, there was even a hairdresser, I kid you not ;-).

If you don’t believe me: here is the Photo Album of this year’s edition!  And if you think this is the only quality edition?  Well – think again – here are all albums of all editions!

All sessions are recorded, and already available on youtube and on mibuso.  If you have about 28.5 hours to spare – you can find all videos here.

My edition

As predicted, my NAVTechDays was a bit too busy.  So much content in only a week .. I have to admit – it’s simply too much.  I probably won’t do that again – and if I do – I’ll at least have this blogpost to hold onto to declare myself crazy .. again ;-).

One special thing I was really happy to be able to do: I got my parents into my session. That’s a special feeling, I can tell you that. You always try to explain what you do, and what impression it has on you – but they can only understand when they actually experienced it :-).

I can’t judge if my sessions were well received.  Thing is – I realize that the topics I talk about – the opinions I evangelized are not always the opinions that are shared by all of you.  Like “Code Customized AL” or “embracing dependencies” to just name a few. I know people that are passionately for code customizing AL – and who are passionately against any form of dependencies.. .  Well, I realize that it can have its effect on how the session is being received (like: complete bullshit ;-)).  All I can do, is share my experience, and what I believe makes sense to go forward .. and I still stand by 100% I have been advocating ;-). And yes – in the “real world”.

In any case … as said, you can find my sessions on mibuso and on youtube here:

NAV TechDays 2019 – Development Methodologies for the future

And the session with Vjeko about connect apps:

NAV TechDays 2019 – {Connect app}²

In the next weeks/months, these videos will also be turned into a series of blogposts.  I already planned a few – and Vjeko is already blogging is ass off as well .. .  Expect a lot, soon (or late – no promise ;-)).

All there’s left for me to say is: thank you!  Thank you for joining my session, thank you for joining my workshops, thank you Luc, for making this happen for all of us – it’s a real honor to be a small part of it!  Thank you, Vjeko, my bro, for sharing the stage with me :-).  Awesome week!

Picture time!

For a few weeks now, people have been asking “can I have the restapp you were showing” – well, here it is: https://github.com/waldo1001/waldo.restapp

But that would be not much of a blogpost – would it ;-).

Just an example…

During NAVTechDays, I did 2 sessions:

• Development Methodologies for the future
• {Connect app}² (with Vjeko)

In both sessions, I talked about the concept of “dependencies”. Yes indeed – in my opinion, “dependencies” is an opportunity that we should embrace .. (just watch the “Development Methodologies” session if you want to know how and why). Now, during the sessions, the RESTApp was actually just an example on how we internally “embrace” the concept.

What does it do?

Well .. not much, really. At least if “making your life a lot easier” is “not much”, that is ;-).

It just “encapsulates” the complexity, the functionality, the responsibility, the “whatevery” that has to do with “REST Calls”.

I mean, did you ever try to use the httpclient-wrappers to do any kind of “decent” webservice call? Did you ever fumble with setting the “content-type”, or any kind of headers? And honestly – did you spend more than 5 minutes to make it work? Well, if all answers to these questions are “yes” .. Then you will appreciate this app ;-).

Or as a picture tells you more than an 1000 words, it turns this code:

Into:

I hope you agree that this second way of handling the same call, is a LOT easier:

• Not having to declare all these different http-types, just the RESTHelper codeunit from the RESTApp.
• Not having to care about the order of adding content type or headers or anything like that
• Not caring about error handling.
• …

The current functionality includes the following things:

• A helper codeunit for REST calls
• A helper codeunit for JSON stuff
• Logging of the request and the response message

I could have done that with a simple codeunit, dude

I hear you. “Why an App? Why implementing the complexity of dependencies for such a simple thing?”

Well, it’s not just there to make your coding easier. It’s there to make the whole lifecycle of how you do webservice calls in all your projects easier. Just think about it. You are probably going to do REST calls in many different products and/or implementations. And many of them are different, need more functionality, …

Or – you just might run into a bug on how you do it, and have to update a bunch of other implementations .. .

Or – at some point, you might figure that having a decent logging for all outgoing REST calls would be interesting (and let me tell you that: it IS interesting (and yes, it’s already included in this app))! If you have implemented a RESTApp like this, a simple update gives you this new functionality on ALL your projects. Simply release it to all your customers (long live DevOps!). You can update all you want .. as many times as you want.

Or – at some point, you need to be able to set up “safe” sandboxes, and need to overrule all webservice-calls in a sandbox to not risk “live” calls from a sandbox (guess what – this IS something to think about!)? Just update this app, deploy, done! On ALL your customers.

I can give you lots of scenarios, to be honest.. . But tell me again – how is that codeunit going to help you in any of this?

Just an example

I know, I already had this subtitle :-). But now, it’s more like a disclaimer.

Don’t see this code as being a leading app. It’s meant as an example .. and nothing more. It is not the version we are using internally (which I’m not allowed to share, as I don’t own the code). It doesn’t have “authentication helpers” or anything like that. And it probably doesn’t have all the functions that are necessary to do all kinds of REST calls.  Obviously, this is where you can come in :-). May be it’s not a “leading app” now (if that’s an expression at all) .. you can help me make it one ;-). Please feel free to do any kind of pullrequest. Anything that might help the community. Change, restructure, .. whatever!

May be, at some point, it’s mature enough to pullrequest into Microsoft’s System Application ;-). In this current state, in my opinion, it isn’t.

Ideas

I do have some ideas that I want to include in this. Like making it easier to work with different authentication-types. Or including a way for a test-url, like request-bin, that replaces all calls with this url for you to be able to track the actual request calls that are being generated.

If you have ideas, or remarks, or anything, you can always leave a comment – or use github and use the “issues” section to add ideas (or issues).

Enjoy!

Well .. no!  “Not yet”, at least ;-).

Let me be clear: this post is NOT a recommendation that you should use Docker for your OnPrem Customer’s production environments.  Not at all.  This is merely a blogpost about the fact that I wouldn’t mind Microsoft to officially support Docker as an alternative to an NST deployment.

If you don’t care about the “why” below – just upvote here ;-): https://experience.dynamics.com/ideas/idea/?ideaid=daf36183-287e-e911-80e7-0003ff689ebe

“Continuous Upgrade”

Just imagine you would be able to continuously upgrade your customers.  This has actually quite an impact on your daily life .. on anything that involves a life of a partner: from support to customer deployment to hotfixing, to release management, … . 

Let me give you a few examples – and I’ll do that with some extreme numbers: either we have 300 customers, all on a different versions – or we have 300 customers all on the same version:

Product releases

In a way, you need to be able to support all product releases on all versions of Business Central (or NAV) that you have running at your customers – it doesn’t make any sense to support a version that isn’t running at any customer, does it ;-)?  If a customer is running v13 of your product, you need to be able to hotfix it, and roll out the fixes to one or more customers with that same version. 

Even more – not only, you’d have to keep track of all the versions/releases/customers – you need to manage the hotfixes, and bump it to all versions/releases necessary (a hotfix in v13 might be necessary in 14, 15, .. as well) .

On the other hand – if everyone would be on the same (and thus latest) release: everyone can be treated the same, and hotfixing is easy, rollout is easy, administration is easy.  Simply because there is only one product release to maintain (you start to get why Microsoft is pushing us to the cloud, right? ;-)).

In order to facilitate this in Git/DevOps, one way is to create (and maintain) release-branches for all supported releases.   On top of this, you have to maintain for all these  branches a dedicated branch policy, build pipeline, artifact feed and what not .. .  Good luck doing that for 300 different versions.. .

Support

I think we can all agree that our support department would be so much relieved if they would only have to support 1 version/release, right?  All bugfixes/improvements/features/tooling/… are just there.

Bottom line

The easier we are able to upgrade a customer to the next runtime of Business Central.. the more customers WILL be on the latest Business Central and version of our product .. the easier it is to manage our product .. The easier it is to support our product .. The easier our life is.  It’s a simple fact.  No science needed here …

Upgrading an OnPrem customer

You might know my opinion on “code customizing AL” – if not, you can find it in this post.  In a way – for me – “code customizing AL is evil” ;-).  So .. In that perspective, I’m going to assume we are all on extensions/apps .. and all we have to do is manage apps at customers.

In terms of upgrading – we would upgrade apps to new version of apps, which is quite easy to do.  You can prepare all upgrades in upgrade codeunits, so in a way, when prepped correctly, upgrading is just a matter of installing a new app the right way (by triggering the upgrade routine).  I will not go into how to do this.

But that’s not all …

We also have to upgrade the platform, the runtime.  Not the client anymore (thank god ;-)), but still all the NST’s and other binaries we have installed.  At this point, it’s still quite manual: “inserting DVD and start clicking”.  I know it’s scriptable .. heck, I even created a function once to “easily” upgrade an existing installation by calling the “repair” option from the DVD (you can find the script here), but honestly, in a world with Docker …

The Docker Dream

Just imagine – all you do to install an OnPrem Business Central customer is to install a real SQL Server for the database, and use the docker images provided by Microsoft for the NST.  Why only the NST?  Well, that’s the part that needs to be upgradable like every single month. 

But when on Docker, you know how easy it is to set up a new environment, right?  How easy would it be to upgrade, to set up UAT environments in other versions, to “play” with localizations, .. .  Well, as easy as we know already by using Docker – but applying this to a production environment would really eliminate the complexity to upgrade continuously.

Honestly, I think this is the missing link to be able to implement full “continuous upgradability” for OnPrem customers. 

We already do this …

Call me nuts – but for our internal database, which is completely our own concern, we already have this running as a proof-of-concept.  And it has been running for so many months without one single problem :-).  I shouldn’t say this, but it has been making upgrading and maintaining this particular environment (with +20 apps) so much easier that we are really wondering “why not” for customers.  We won’t, obviously, but still … we dream ;-).

Vote!

If you agree with me, then you also agree with Tobias Fenster, who has created an idea on the ideas-site which you can upvote – please do!  If you don’t understand a single thing about Docker or what impact it could be for us – than just take my word for it and still upvote it here: https://experience.dynamics.com/ideas/idea/?ideaid=daf36183-287e-e911-80e7-0003ff689ebe

First of all, this is my first post of this new year, so I’d like to take this opportunity to wish you all the best and happiness for 2020! Last year, I did a post on my new hobby “3D Printing”. Well, now we’re a year later, I’m still printing almost full time 24/7, so let me wish you all the best with some 3D Printed goodies I did for Xmas ;-).

So now – “let’s get started” with ..

Why would I rename all my AL-files?

Well – because of Microsoft and their new (upcoming) code analysis rule about file name conventions.  You might already have complied with their previous conventions, and … they will change.  Even better: they actually already changed .. Just read here:

https://docs.microsoft.com/en-us/dynamics365/business-central/dev-itpro/compliance/apptest-bestpracticesforalcode

In the “old” days, Microsoft wanted you to name the files including the object numbers.  But not anymore.

This doesn’t really have to be a problem .. because today, we can freely choose however we want to name our files.

But that will change (a bit) …

If you use the insider version of the AL Language Extension for VSCode, you’ll see that we will get new code analysis rules that check the file name. This is an example that I had with my waldo.restapp I recently blogged about:

It doesn’t matter what my previous naming convention was – the result is that all of a sudden, te compiler is going to have an opinion about it and will identify it as a “problem”.  It even calls it “incorrect”, not just “not following the guidelines” or whatever.. ;-).

Why would I care?

Well, if you don’t care .. then don’t!  Just put rule AA0215 in your custom coderules and move on with your life ;-).

But I care.  I want to comply with as many coderules .. as long as they don’t “hurt” me or my company.  That’s why we have a simple step in the build pipelines of our company: if any coderule would fail with an error or warning – the code is not accepted by the build.  And admittedly – this is a good naming convention, anyway.  So .. let’s comply!

Waldo’s CRS AL Language Extension

As you might know – I created a VSCode Extension that makes it possible to easily handle file naming conventions.  Because of these new guidelines (and coderules), I recently added a new option “ObjectTypeShortPascalCase” so that now it’s also possible to handle it the correct way with settings in this extension. 

BUT

Don’t rename your files by just invoking the command “Rename All Files” – DON’T:

Because:

• You might change all your code history
• You might have wrong settings
• You might not be able to rollback and rethink ..

So, let me try to give you the safest way for you to rename all your files

I’m going to assume you are source-controlling your code here.  If not .. I don’t even want to talk to you ;-). 

Just kidding – if not, just quickly initialize a git repo from the current state of your code.  You might want to revert all changes you did .. and this gives you a way to do that.  Honestly, this particular rename I did for this blog, I did 3 times .. just saying ;-).

1 – Create a new branch

Before you do any coding, you should branch out.  Same with this: just create your own snapshot of changes to work in.  You can’t be too careful ..

You created it?  Well, now you’re safe.  Any kind of mistake within the next steps – just remove the branch, create a new one, and start fumbling again ;-).

2 – Change the setup

I think this is a good setup for the new filename conventions:

"CRS.FileNamePattern": "..al",
"CRS.FileNamePatternExtensions": "..al",
"CRS.FileNamePatternPageCustomizations": "..al",
"CRS.OnSaveAlFileAction": "Rename",
"CRS.RenameWithGit": true,
"CRS.ObjectNameSuffix": " WLD",

Notice that there is no property “RemoveSuffixFromFilename” or “RemovePrefixFromFilename”.  These properties invoke an extra procedure after renaming, to make sure you have no prefix in the filename anymore.  Thing is – it would make the name not comply again with the naming convention of Microsoft’s coderule, unfortunately.

3 – No Multiroot

Also notice the “renamewithgit” setting.  This is important, but a difficult one.  If you don’t set this to “true”, any rename would result in a “delete” and “create” of your file – meaning you would lose all the history of your files.

In my world of DevOps and Source Control, this is unacceptable.  Really unacceptable.  I want my history kept at all times!  So, this setting is going to make sure you rename with git (“git mv”).

Now – in all honesty – I have spent too much time already to make this setting as stable as at all possible. And it still isn’t.  Sometimes, it just doesn’t do what it’s supposed to do.  Known issues are still:

• new files in combination with a multiroot workspace, it acts crazy (just deletes the content)
• With the “rename all”, it seems to be too fast sometimes, which fails the rename, which loses the history, which makes me angry ;-).

For the first, I have built in some fail saves, which basically will NOT rename with git of you have a multiroot workspace (not my preferred option, but for now, I have to give up on this – spent too much time already :'( ) ..

And the second – that’s why I write this blogpost ;-).  How can you still rename all your files successfully, even when it would fail on your first attempt.

But – to solve the first (you DO want to “renamewithgit”), just avoid using multiroot workspaces.  If you have multiple workspaces open, just close VSCode, and open that workspace of that one app you would like to rename all files from.

4 – Commit these settings!

You might want to revert to this very version of this set of files when any of the next steps would fail for some reason or another.  So please, commit now!

5 – Rename – All Files

Now is the time to start renaming all your files.  

Remember you set the “renamewithgit”, AND you’re not working multiroot, so it is going to try to rename all with git.  Indeed: try!  Because more than likely, “some” will succeed, and “some” will fail.  If you have failed ones, it will show the “crs-al-language” output like this:

It seems that one process is blocking another to execute decently .. and because of that, it is actually going to rename the classic way, which would mean, it doesn’t preserve your git history.

When the complete rename is done, you should also have a look at the Source Control pane.  If you have fail-messages, you should get some files in “Staged Changes” and some in “Changes”

Well: Staged are fine, and the ones in “Changes”, you’ll need to retry.  Do this by “Discard All Changes”

And “Discard All x Files”

Now, you basically only renamed the ones that were able to rename, and you staged them. The ones that failed are restored to their original state, which still has all history, but are not renamed. So ..

6 – Repeat …

it’s obvious .. this is where you repeat.  So, rename again, and discard the changes if you have failed ones.  And keep doing that, until all files ended up in your staged area (the “R” in the back indicates that you renamed the file):

7 – Commit, PullRequest …

And this means: you’re done.  Time to commit and pullrequest your changes to your master/release/dev.. .

Conclusion

I’m not proud of this workaround – I’d like to see it “just” work.  But when it comes to “renaming or moving files with git”, there seems something odd.  I mean, think about it – a simple rename in the VSCode Explorer window also ends up with a “Delete/Untracked”, meaning – you lose history.

A drag&drop to a new folder as well.

So – just my conclusion – if the VSCode wizards are not able to solve this simple “rename” issue – why would I be able to do it better than I already did ;-).

If this is useful to you! It was for me

I did a poll recently during one of my sessions – and I was surprised that about half of the people don’t regularly use “snippets” in VSCode.. .

Well, some of you probably know that I’m a big fan of snippets.  Whoever has joined one of my sessions of the last couple of years where I was talking about VSCode, or was working with VSCode .. most probably, I was talking about, showing or using snippets.

Honestly, I can’t live my life in VSCode without snippets.  It would be so much less efficient .. .  VSCode wouldn’t be much more than a notepad that is hooked up with Source Control in that case.

Back in 2017, I even spent quite some time in my session on NAVTechDays about snippets.  I’ll share the session at the end of this post…

How to manually create snippets?

I would be able to explain you how to manually create snippets – but I won’t.  Because it is actually already beautifully explained on the documentation section of VSCode here: https://code.visualstudio.com/docs/editor/userdefinedsnippets

Basically, user defined snippets …:

• … is a json-file somewhere on your roaming profile
• … have a specific syntax, which makes you give a description, a prefix, a body, … and for defining placeholders, including multi-cursor placeholders
• … can use variables, like the filename, date or your clipboard, which make it possible to create very “generic” snippets.
• …

All you need to know is on the page mentioned above, and you’ll get going with snippets in no time, I promise you!

My snippets

There is more :-).  I have been creating lots of snippets in my extension.  I admit, I copied Microsoft’s snippets and improved them – but I also have created lots of new snippets.  Ones that I use a lot in terms of “design patterns”, but also for implementing code that I’m not used to, and don’t want to forget (like the assisted setup).  If you install my “CRS AL Language Extension”, you’ll recognise my snippets with “waldo” in the suffix:

And yes, if you don’t want to work with my snippets, you can disable them by a simple setting:

"CRS.DisableCRSSnippets": true

(if you might wonder, you can disable Microsoft’s snippets as well ;-)).

Tools that can help you to create snippets

I recently was pointed to this tool: https://snippet-generator.app/ .  When you are creating your VSCode snippets, simply paste the text that you want to convert to a snippet to this tool, and you immediately get it converted to a JSON representation for a VSCode snippet.  I tremendously speeds up the creating of a snippets from minutes to seconds ;-).

On the other hand, there is another tool that you can install in VSCode: the snippet-creator.  It basically gives you a command that will convert your selected text into a user-snippet of the language of your choice:

Whatever you prefer –  both work very nice :-).

Some questions I get a lot

Where are snippets stored?

The user-defined snippets that you create,  are stored here:
%USERPROFILE%\AppData\Roaming\Code\User\snippets.

The snippets that come from an extension, are stored here:
%USERPROFILE%\.vscode\extensions\<extensionname>\snippets

Can I disable snippets?

Well, no.  You can’t in any decent way (that I know of) disable snippets.  I know I was talking about a setting in my extension, and yes, that’s a way, but it’s not a decent way ;-). 

In fact, what I do in that extension, is simply rename the “snippets” folder to “snippets-disabled”.  That way, the extension is not able to find the snippets, and won’t show them anymore.  The downside of this is that it will give errors in the background because it’s not able to find the snippets anymore, like:

It’s not really noticeable, but they are there… .

Can I change snippets?

Well, no again.  To be fair: you CAN change a snippet in the extension folder, but do know that when the extension is updated, it basically is going to overwrite the snippets … and you lost your modification.  So in my opinion, that’s not an option.

That was it!  Hope you’re into snippets and this blogpost was completely useless.  If not, at least I hope this got you triggered a bit ;-)!  The only thing left for me is to share the NAVTechDays session I was talking about earlier:

I recently blogged about how to rename files with my VSCode Extension to comply with the new filename conventions from Microsoft.  Somewhat related to that, we need to talk about …

Pre- or Suffixing

It’s been known for a while: pre-or suffixing is necessary to “live in the cloud”.  It prevents duplicate names in various places: duplicate table names, duplicate field names .. things like that.  It is actually very well described on Microsoft Docs here: https://docs.microsoft.com/en-us/dynamics365/business-central/dev-itpro/compliance/apptest-prefix-suffix.

In short: you need to pre-or suffix, especially for apps on AppSource, but actually on anything you do! 

AppSourceCop can help (let’s say “force”) you

You are probably familiar with  : the codecop with code analysis rules that are specifically for AppSource.  Well .. when you enable this, it can help you.  And in the next release: it will force you to remember to set an affix (which means: a suffix or a prefix).

You can simply enable the AppSourceCop by:

In terms of “affixes”, the AppSourceCop needs to know which you use.  If you don’t do that – in the next release (v16) – the analyzer will tell you that you HAVE to tell him:

What you need to do 

From the documentation, it’s quite clear what you need to do:

• Create a file AppSourceCop.json in the root of your workspace (next to the app.json)
• Fill it with the property “MandatoryAffixes”, and provide all affixes that you intend to use in the app.  Here is an example of the content of that file:

You’ll see that IntelliSense will help you to complete this file… .  As you can see, I also provided the supported countries as also that will be mandatory (by a coderule) in release 16.  But let’s focus on the affixes in this post ;-).  What I told the AppSourceCop in this case, is that I will use only one affix, being “WLD”. 

Done?

Well, you’re not done yet.  All you’ve done now, is told the AppSourceCop on which affixes it needs to check the objects and controls.  That way, it can “remind” you that you need to pay attention to it, like here:

Obviously, it’s your job now to  take this into account, and haven decent names for your objects, fields, .. .

Wait .. it’s not going to provide that affix for me??

No!  But don’t worry – that’s where I come in ;-).   You might be familiar with my “CRS AL Language Extension” in VSCode that can handle renames of files.  Well, during the rename, it can provide a prefix or suffix.   So by simply setting some extra settings for my extension (and yes – by providing the suffix again), you’ll be able to get this going for you in an automated way:

In short, this setting will automatically provide the WLD-suffix where it needs to when you save a file.

There is a disconnect though …

The AppSourceCop allows multiple affixes .. and my CRS AL Language Extension does not.  So whenever you intend to use multiple affixes, I’m afraid you’ll have to disable the automatic rename by my extension, because it will only apply one.  Simply remove the settings about the suffix, and provide it manually (AppSourceCop will remind you where ;-)).

Even more – if you enable the CodeCop in the next release, you’ll even have another rule to worry about: file name conventions.  So, if you enable codecop:

This is something you could get in v16:

I actually already blogged about that very recently, so I’m not going to repeat that.  BUT – this coderule is also going to assume than when you used this “MandatoryAffixes” (which you have to in the next release), that you DON’T use the affix in the filename (read the error above carefully, and you’ll notice it doesn’t want my affix in it).

Well, I had foreseen this setting for it:

But that also is only useful for only one suffix or prefix at this moment. 

In short – auto renaming can become a challenge … :-/ … and that sucks…

Stay tuned…

Headaches that I intend to solve for you in the near future.  This is what I intend to do in short term:

• I’m going to read and take the “MandatoryAffixes” setting from the AppSourceCop.json file into account
• When you have enabled “RemoveSuffixFromFilename”, I will loop all affixes, and remove it when it’s used as prefix or suffix. The danger there is that I’ll remove too much, or make the wrong assumptions, so I’ll make it case sensitive, and probably stop removing after I find one affix – just to minimize the risk. Future will tell how well it works. But this will hopefully solve the fact that the CodeCop wants us to disregard a multitude of affixes.
• I will keep the “ObjectNameSuffix” and “ObjectNamePrefix” setting. It will act as a “current” setting. Meaning that when I find that one of the affixes in the AppSourceCop.json is already applied, I’m not going to apply anything anymore. If not, I’ll apply the one from the settings.. . I’ll try to make this case sensitive as well.

Or that’s at least how I have it in my head now … .  I also realise that simplicity is key – I shouldn’t make it too complicated to use this.  So .. yeah .. this did give me quite some headaches already ;-).  So, if you have any idea, feedback, tip, … please share ;-).  I’m open for all ideas.

As you probably already picked up in the media – Microsoft has released its plans for Wave 1 release (in April) for this year.  You can find all the information here: https://aka.ms/Dynamics365ReleasePlan.

The document contains all the information on all we want to know regarding “what’s next” for all related applications, like Marketing, Customer Service, Field Service, …. and also our beloved “Business Central”.  And since this is a blog about Business Central – let’s focus on …

Microsoft Dynamics 365 Business Central

The Microsoft Docs endpoint that has all information on Business Central is here: https://docs.microsoft.com/en-us/dynamics365-release-plan/2020wave1/dynamics365-business-central/planned-features .  You can completely disregard this blog, and just read that link, and have all the (updated) information ;-).

But – I’m still going to talk about at least a few of these points.  I did that before, and last time, I remember I did rant about one thing (Code Customized AL) – let’s see if there is something to rant about in this release ;-).

What am I looking forward to?

You’ll see I’m focusing most on Tech.  and less on functionality.  That’s just me – sorry ;-).

AL Interfaces

I don’t know exactly yet what to expect from this and in which way it will be used in default Microsoft Apps for Business Central, but it is at least a new interesting  capability in terms of code architecture. 

Look up events and insert event subscriber in code

Yes!  Sounds perfect!  By using Shift+Alt+E, you’ll get a list of all available event that you can easily search through:

Tip: it seems you can already use it – just use the vsix from the insider-docker-image if you have access ;-).  Just be careful though – it seems to add all parameters, even if you don’t need them.. (which in my book is bad practice).

Camera/location AL API available in the browser

When I read this, my immediate reaction was: “uhm, ok, whatever”.  I wasn’t really waiting for this, however, I do think this can really come in handy.  Let’s see…

Obsolete tag and URL properties

I can’t say I needed this, to be honest, but in the future, when real refactoring is going to be necessary, I really think this is very useful to clearly explain the ins and outs of certain functionality that will have to change, and how to deal with that if you depend on it.

But then again, in my company, we’re rebuilding the product from scratch.  If you’re converting/migrating, I can imagine that “refactoring” will come rather sooner than later ;-).

Multiple variable declarations of the same type in the same line

You can see that the AL Language is steadily growing into a more mature dev environment, isn’t it.  This is something that is so “normal” to have in any other language, and finally coming to AL.  Though, I must say, there are other fish to fry, I guess ;-).

Ability to refactor a field from a table to a table extension

It seems – but I’m not sure – this is an answer on “how to transfer data from a customization in the base app, to an extension”.  How this moves to BC online, I don’t know – but definitely worth digging into!

Migrations to Business Central online

This is obviously really important and need no further explanation!  I just hope it’s going to be a painless process, because upgrading to 15 .. Has.not.been.painless.at.all!  But a migration is not an upgrade either, is it? ;-).

Business Central integration with Common Data Service– can I get a “finally” here, please? :-).  It’s not just “bla bla” anymore – there will actually be some “boom boom” as well – most probably SaaS only.

What am I NOT looking forward to?

Well, I guess this “Application version for aliasing base application” thing.  I’m not saying it’s useless, but it looks like this is only added to facilitate “Embed Apps” – which is nothing more than “Code Customized AL” (in my opinion) – and you know how I think about that ;-).

Added to this, the “propagateDependencies” has been a change that will be introduced in the next CU Update – and it has brought me nothing but pain – downloading symbols in a build pipeline just became more complex (luckily, navcontainerhelper eases the pain ;-)).

Is that all that is new?

Of course not!  I just didn’t want to “just” copy/paste all items just for the fun of it.  All Tech things, ok, but there is so much more that is on the plate:

• Application enhancements (also read Erik Hougaard’s blog about this)
• Country and regional availability (Colombia, Croation, Czech, Hungary, Latvia, Slovenia and Peru)
• API to AppSource to be able to hook up your DevOps for your AppSource Apps (really important)
• Modern clients– always interesting to see what the new client will bring.  What really stands out to me is:
  • Navigate trees more easily– treeview is almost back, I guess
  • Import profiles and UI customizations– very necessary to be able to use this more efficiently (by consultants and admins at customers).  I just hope it’s going to be stable, because I have been removing all customizations too often because of client crashes lately :-/.
  • Saving the URL as a bookmark, will include filters now – interesting :-).
  • Improvements to data entry is something everyone likes to read in regards of a web client.  But it’s not clear what it actually includes (to me) yet.. .
  • Enable non-interactive printing with advanced printer selection– looks like a good step into a decent printing experience.. .
• Service fundamentals

What am I missing?

Monetization for AppSource apps – for sure.  I might have misread or missed it, but I still don’t see it on the roadmap.  Why?  Why oh why?  Do we really want 100ds of apps on there with all a different implementation of monetization?  That doesn’t make any sense.. .

That said – James Crowter wrote a nice article on what HE would have liked to see as prios for 2020. You can find it here: https://msdynamicsworld.com/story/my-dynamics-365-business-central-priorities-microsoft-2020 .

You can see he has much more experience with an actual “load” of Customers on BC SaaS.  And I can only agree with his list and argumentation.. .

Ideas

Anyway – a big “driver” of this content is the “Ideas” website where they gather ideas from the community.  You can easily access it by: aka.ms/bcideas.  If you miss out on anything, the first question you should ask is: “did I ask for it?” ;-).

Conclusion

Nothing much to conclude besides quite the same as Erik : nothing groundbreaking, but definitely improving and a further evolution.

NAVTechDays is already over for a while .. and yes, I already blogged about it.  But I recently had to refer to a part of my session on “Development Methodologies”, and I noticed that someone named “Marcus Nordlund” actually put quite some time to completely “menutize” the video in the comment section of the video :-). 

Awesome effort that I needed to share! Thanks, Marcus!

Here it is:

• Future 3:08
• Development methodologies 3:52
  • Don’tNet 4:04
  • Embracing Depencies 4:43
  • Strict Guidelines 5:26
  • No Test, no approval 6:01
  • Extendable patterns 7:13
  • Avoid Breaking changes 8:23
  • When i’ts hard in al, it doesent belong i AL 9:02
• Embracing Depencies when moving to AL 9:27
  • Moving to AL
    • Migrate to Code Customized AL 10:03
    • Migrate to a Monolith extension 12:18
    • Rebuild monolith from scratch 12:34
    • Rebuild a set of apps with dependencies 12:53
  • Dependencies 13:16
    • Dependencies – Why? 13:52
    • Dependencies – How? 15:04
    • Dependencies – Schema example 15:16
    • Code Example of schema 15:59
    • Dependencies – Real schema example 19:27
    • Please Think things through 20:13
    • How do I think things through when moving to AL 20:53
    • Dependecny Analysis based on c/al 21:07
      • waldo.Model.Tools 22:13
      • Flow 25:15
        • Get All objects from C/AL & Talg all 25:34
        • Manually correcting/Ignoring modules 30:00
        • Get Where Used per object 30:14
        • Analyze dependencies per module 32:33
        • What is a circular dependency? 32:27
        • Solve circular dependecies 35:57
        • Change design 37:15
        • Manually create App-layer 39:27
        • Analyze dependencies for app 39:48
        • Solve Circual dependencies again 40:27
        • Tools we used are availabile for you 41:13
    • Dependecy-stuff a bit further 41:41
      • How to architect? 45:54
      • To be albe to set up demo environments 49:17
      • How it looks in the app 50:29
      • How it looks in code 52:09
      • Rest 57:02
  • Don’tNet 58:25
  • Some other more hands-on develoment methodologies 59:09
  • Where do we put business logic 59:18
    • Method codeunit 59:58
      • Decouple method from another extension 1:02:59
  • One App, One Repository 1:06:00
  • Test Driven Development 1:08:11
    • Workspace with different repos 1:09:02
    • Integration tests 1:10:14
    • How to create integration test 1:11:25
    • Automated Testing in Microsoft Dynamics 365 BC 1:13:22
  • Breaking Changes 1:13:56
  • Number Series 1:15:12
  • Translations 1:16:48
• Summary Development methodologies 1:18:08
• Questions 1:18:20Move code from tables to codeunit 1:18:55
  • Test Codunit – test XML ports 1:19:40
  • Tips for connection between BC and Azure DevOps 1:20:38
  • Where do you do the builds of apps? 1:22:26
  • Multiple apps, extend a table will be many separate tables 1:23:29
  • Upgrading testdriven development when converting to apps 1:25:02
  • When has it been to hard to do it in AL 1:25:44

Here is the complete video:

Sidenote:

You might have figured – I’m a VSCode fanboy.  One of many.  You might remember the session I did on NAVTechDays 2017 (Rock ‘n Roll with VSCode), where I dove quite a bit into the possibilities this great tool comes with.  But I didn’t talk about the concept of “Multi-root Workspaces”: an ability of VSCode for you to work on multiple “projects” at the same time, in one environment.

2 years later, last NAVTechDays, I talked about Dependencies quite a lot, and when you think of it – in terms of dependencies in AL for Business Central, these “Multi-root Workspaces” might make a lot of sense, because when you have all these apps, you might have to work on multiple apps at the same time. 

Even more, in that same video, you’ll see that I ALWAYS have at least 2 apps: an app, and its test-app.  So in terms of “Test Driven Development” (a methodology I believe is indispensable), you will always have an app, and a dependent app, and you will always work on both apps at the same time.  So – “Multi Root” to the rescue!

What

Well, the concept of “multi-root workspaces” is actually most simply explained by: opening multiple projects (workspaces) at the same time, to be able to work on multiple pieces of software at the same time.  Or in terms of Business Central: to work (compile, publish, develop, …) on multiple apps at the same time.

The concept is quite well explained on the docs-page for VSCode: https://code.visualstudio.com/docs/editor/multi-root-workspaces

VSCode Extensions

The one downside of this concept is that not every VSCode extension might be ready in a Multi-root environment.  In fact, it took a while to get my own extension (the CRS AL Language Extension) ready for Multi-root workspaces (and there still might be issues with it ;-)).

Same for Microsoft’s AL Language Extension.  It took a while – but just imagine:

• When you’re working on extension A, from which Extension B is dependent – Extension B will recognize the symbols of Ext A without even having to create (compile) an app file – just because Ext B is in your Multi Root Workspace
• When you’re debugging, you’re able to go to definition into al-files of other apps you’re debugging – not just the only-one-level-dal-files.  Just because you have the dependent app in your workspace.
• Whenever you compile a main app, the symbol files of the dependent apps are updated with the new version of the main app.  Just because it’s in your multi-root workspace.

All this is already possible!

And that’s what a real “Multi-root” experience can give you – and why I think we should always look into this.  It makes all the sense in the world – in a world with “lots” of apps and dependencies, to work on them simultaneously.  And I’m sure – if you’re not yet doing it – it can speed up your development process even more!

In Practice

May be a few screenshots that can show you how it could look like.

In this case, I’m actually working on 7 apps at the  same time, with all of them having a test-app as well (which we use for unit-testing).  The last app is the integration-test-app.

How does it determine what app I’m working in?

Well, simply by the active editor.  If I would want to compile and publish the BASE-App, I would open one of the files of that app, and simply press F5. 

Symbols constantly updated over all workspaces

In my case, the BASE App is a library-app for all other apps.  And if I would simply start coding in that BASE-App, like this useless codeunit:

I would – without even compiling my app – be able to code against this in any app that is dependent from that BASE-App:

Basically meaning: I can code against all apps at the same time, without even downloading/updating symbols, or compiling or publishing.

Updating Symbols

Even, when you hit compile, you’ll notice that the symbol files in the .alpackages-folder are updated for all apps that depend from the app that you’re compiling.  Very cool!

Can I control this a bit?

Well – you do have the “dependencyPublishingOption” in the launch.json, which has 3 options:

• Default: set dependency publishing will be applied
• Ignore: Only the current project is published, dependencies are ignored.
• Strict: Publishing will fail if the project has apps that not part of the workspace and depend on it

I like the default setting – something it takes a little longer to include all dependencies in the compile – but at least you’re checking all dependencies .. So that must be good ;-).  But, in some cases – especially in case of big apps with many objects – you might want to avoid unnecessary recompiles of dependent apps.. .

Does this mean everything needs to be in one (Git) repository?

No!  VSCode is smart enough to handle multiple workspaces at the same time.  In case of the screenshot above – and as mentioned in a recent webcast I did about handling dependencies in DevOps– all my apps are in separate repos, together with their Test-apps.  So in case of the screenshot, it would be a collection of about 8 GIT repositories.  When you switch to the “Source Control” window, VSCode clearly shows all states of the repositories, like you see here in the screenshot:

You can simply see when repos need attention (new/modified/delete files), what branches, and what the synchronization state is.  From this window, you can obviously also change branches, sync, and so on.. . 

There is more …

And I have more for you.  To make working in multiple workspaces more easy, I’m working on a set of scripts, which you find on my PowerShell repository: https://github.com/waldo1001/Cloud.Ready.Software.PowerShell

I will maintain a set of scripts in the folder ./PSScripts/NAV Extensions v2/MultiRootWorkspaces/ .  Today, I have foreseen:

• LaunchJson_CopyToAll.ps1: If you change docker image or whatever – this script will simply copy one launch.json to all other workspaces so that you will easily publish other dependent apps to the same sandbox

• Symbol_Cleanup.ps1: This script removes all symbol files – just imagine when you want to publish against another version (localization, insider, ..), you can simply cleanupt all files and even invoke the next script, which will …

• Symbol_Download.ps1: This script will download as many symbols as possible for all workspaces that are available in your multiroot workspace.

I’m also working on a “CompileAll” script – would be nice to have one script that figures our the dependency tree (I already got that one), and invoke start compiling all the apps – and – may be – if possible – even starts publishing them to the server instance ;-).  Let’s see where we end up with that one ;-).

There might be issues with the scripts – they are brand new – but please, use, abuse .. and contribute ;-).

Enjoy going Multi-root!

You probably know about the twitter hashtag “bcalhelp” .. A way for you to ask help on twitter about anything AL for Microsoft Dynamics 365 Business Central.

Well – yesterday, it made me smile … there was someone that was trying to find something on Vjeko’s blog.  Pretty clever to use the bcalhelp for it :-), because he immediately got response.  Here is the twitter thread (if you can call a question/answer a thread ;-)):

You see that “trickynamics” got immediate response from “Marton Sagi” .. And if you don’t know Marton Sagi – then you probably realize that you actually do, because it’s the guy behind the AL Object Designer– one you should add to your VSCode Extensions ;-).

Anyway – not being able to search Vjeko’s blog is something that needs to be fixed.  It’s unthinkable.  So I decided to contribute to that – and I would encourage all the social-engaged people in the community to blog or share this one on all levels ;-).

How to search Vjeko’s Blog:

Well, it’s a wordpress blog, and every wordpress blog can be searched by adding “/?s=querystring“.  Just try to search my blog on the top right, and see what URL you get.  A similar thing you can do with vjeko’s blog.

So, a few examples:

• If you want to search for waldo: http://vjeko.com/?s=waldo
• If you want to search for multiple words, like “waldo’s blog”: http://vjeko.com/?s=waldo%27s%20blog
• Or actually something useful: http://vjeko.com/?s=business+central
• Or: http://vjeko.com/?s=Azure+AD+Business+Central

You get the picture

How come the search is removed?  Doesn’t Vjeko want us to search his blog anymore?

Well, I talked to him, and he is actually in the process of moving his blog to his own server – and moving a wordpress apparently comes with some challenges .. so let’s give him some time, I guess ;-).

Disclaimer

Yes, with all the heavy content out there, I thought it was time for a lightweight article – which is mainly meant as a joke to put a smile on some faces ;-). Enjoy…

Not too long ago, I did a webinar for ALOps.  The idea of these webinars is simple: how can you get started with DevOps for Microsoft Dynamics 365 Business Central.  And I’m not going to lie: I will focus on doing that with ALOps.  But that’s not the only focus of these videos – I will touch lots of stuff that has nothing to do with ALOps, but more like strategies, simple “how to’s” and so on – which should be interesting to anyone that is trying to set up DevOps in any way.  The webinars will end up on YouTube, and in the description of the video, I will try to give an overview of the handled topics, and create direct links to the exact place in the video.  Here is an example of the description-section of the first video:

Now, In that webinar, I explained just in a few minutes how to setup a build agent for DevOps.  And actually, there is more to say about that – so I’d like to extend on that a little in this blogpost.

What is a DevOps agent?

Just think of it: whatever you define as a build, whatever you will define as a release – scripts need to be executed on some kind of environment by “some service on some server”.  A DevOps build agent is this “some service”: a service that is installed on “some server” that will execute the steps that you define in a pipeline.

How you do that?

Before we go into that, let’s first talk about WHAT you will need.  And that is “Docker”.  I strongly believe that if you don’t build with Docker, you’re building the wrong way.  It would mean there is some kind of database/serverinstance that is waiting for you to build and such – which would have had to been cleaned every start of a new pipeline (take schema-changes into account and such..).  While building a docker container every start of a pipeline would mean you have a clean isolated environment every single build.  You can’t get any more stable/isolated/encapsulated than that.

So .. long story short – let’s assume we all use docker for our build pipelines .. so in that perspective .. our server with DevOps agent needs to be able to:

• Use docker
• Setup containers that have a certain localization and version of Business Central

Well – now we know that – let’s review our options that we have for creating a DevOps agent:

• Microsoft-Hosted agents: Not really my favourite.  Microsoft has agents ready for you to use.  First, it seems very interesting and secure, but there are challenges:
  • You have docker, and in many pipeline runs, the same image can be used.  But since the Microsoft Hosted Agent (VM) is discarded after one use (which is secure and all, sure), your next run will again have to pull that image from the docker repo .. No way to reuse docker images.
  • If something goes wrong, there is no way for you to fix / investigate / replicate the problem on the agent (no way to remotely log in)
  • Slow: not because of resource (that’s quite ok), but because of the fact that you’ll have to download the entire docker image every single time
  • Security limitations: it could be that you run into PowerShell/Security limitations that you don’t have under control, like local file access, the ability to download license in a secure way, .. these things.. .  It’s hard to specify which exactly, but if it happens, it is difficult to work around – and usually it means you simply have to accept that that specific step is not going to work on a Microsoft-hosted agent. 
• Self-hosted agents: that’s where I will refer to that section of the video I was talking about before ;-).  It’s my favourite option, because this can be fast, cheap, redundant, flexible, debugable, … however you want.  And as you can see in the video – it really doesn’t have to be difficult to set up a new agent. 
• AzureVM: Microsoft has foreseen a nice and easy way for you to use a Azure VM as your build agent.  Simply use the template: aka.ms/getbuildagent, fill in the parameters, and everything is done for you – a few minutes later, an agent will pop up in your agent pool.  It can’t be done easier, in my opinion.  But ..

So, what is your preferred way, waldo?

Well, as said, the Microsoft-hosted agents definitely not.  fast running pipelines are important in my opinion, and this is certainly not a way to have it running fast.  Just look at this build pipeline on a public project– the “start docker” will always pull the image and then run the build.  Don’t get me wrong – the agent itself seems to be pretty fast (faster than some of my other self-hosted agents), but the fact that it needs to download the docker image every single time .. doesn’t make sense.  If anyone knows a way to prevent this – I’m all ears.

For the same reason, AzureVM isn’t my favourite either.   It’s great to showcase, and as a backup scenario (should you quickly need an extra agent).  But in terms of DevOps build agents, in my opinion, Azure is slow and expensive.  I would never use AzureVMs as a long term solution for my DevOps Build Agents.

So – I guess you know my preference:

Self-hosted agents

In essence, in this case, you have complete freedom on:

• The hardware: You want faster running pipelines?  Invest in CPU power (Ghz, not number of cores).  You want multiple agents on one machine?  You can do it!
• Where it is hosted: your own data center? Under your desk?  Some kind of naked motherboard with some memory and a CPU on top of your server rack (and yes, we do have this ;-))? 
• What it is installed: you want docker or not?  If docker, may be pre-load all necessary images at night?

So, the way I think about DevOps agents:

• They need to be easy to set up
• They need to be cheap
• They need to be as fast as possible
• I don’t care about redundancy of one server (I set up a pool of servers which make it redundant)
• I don’t care about failing hardware (it’s easy to set up, and again: there is a pool)

So, we have been investigating cheap cloud solutions, like hetzner.de, which offer cheap and fast cloud servers.  I think these are ideal if you insist to have a cloud server as a DevOps agent.

And we have been investigating some OnPrem configurations.  This picture is a POC in our serverroom: 3 motherboards with different configs (fast Ghz / lots of cores / lots of RAM / fast M.2, …).

We found out that there is one actual really important parameter: that the CPU speed (not number of cores, not RAM, not Disk speed).  So, now you know that, go get yourself the cheapest 5Ghz PC, and have some insanely fast builds ;-).

Disable Windows Updates

I know this sounds weird.  But think about it.  Windows updates can have a huge impact on the stability of whatever Docker image you are using for your build pipelines.  Just take these blog posts from Freddy into account:

• https://freddysblog.com/2020/02/14/hyperv-isolation-to-the-rescue/
• https://freddysblog.com/2020/02/26/the-world-after-february-18th/

May be I don’t care about the hardware too much in terms of redundancy – but I do care about my pipelines running stable builds .. and clearly a simple windows update can mess up my build – all of a sudden, builds will start to fail.  So my advice for a DevOps Agent (and actually any kind of NAV or BC installation) would be:

• Disable Windows Updates
• Update Windows in a controlled manner, like:
  • Take a snapshot
  • Update
  • Run tests
  • If tests fail, roll back snapshot

You don’t want your development to stop being able to build just because Microsoft changed how they are handling 32-bit applications in a February update.. .

DOAAAS – DevOps Agent As A Service

With ALOps, we’re thinking about providing a service for anyone that needs a fast DevOps Agent fast.  A cloud service sort of speak, that people can use, where their code can be built fast.  Speed is key.  And cost should be minimum.  And focus on AL – meaning:

• Pre-installed Docker
• Pre-loaded Docker images
• Best practices to optimize speed and stability for AL builds
• Controlled windows update
• Your license and code is secure
• …

What do you think – interesting?  Or not worth the investment?  Always nice to have feedback on this ;-).

Recently, I came across this post by Jack Mallender.  An interesting idea on how to efficiently find AL Objects among your files. It basically comes down to using regex in combination with the global search functionality in VSCode, like (yep, I’m stealing this from Jack’s post – sorry, Jack ;-)):

It immediately convinced me that this would be very useful for everyone, so I was thinking – why not making it part of “waldo’s CRS AL Language Extension“?  It didn’t seem too difficult for the experienced TypeScript developer – so for a noob like me, it should be do-able as well ;-).  A few hours later – after a lot of googling – I found the 9  lines of code that made this easily possible .. I’m not joking ;-).

So – I present to you – a new command as part of the extension: Search Object Names.  Simply call the command, provide the searchstring, and look at the result in the search window:

Now, I made it so that when you are on a word, or selected a word in the active editor, it’s going to take that word as the default Searchstring.  Just imagen, you’d like to go to the source of the code of a variable you’re on, simply make sure your cursor is on the word, and invoke the command:

Settings

May be it’s a bit overdone, but yes, there is a setting as well (CRS.SearchObjectNamesRegexPattern) , because you might want to search differently than I do .. .  We were discussing that on twitter, and I just decided to not decide for you on how you want to search, but let you set it up if you would like to search differently as me.  Let me give you a few options on what would be interesting settings…

Find the source object (default)

Pattern: 

'^\w+ (\d* )?"*'

Setting in VSCode:

"CRS.SearchObjectNamesRegexPattern": "^\\w+ (\\d* )?\"*"

// Mind the escape characters

This is the default pattern, which means you don’t have to set anything up for this behaviour.  Basically this pattern will search any occasion where it starts with a word, than optionally a number, and then your search string.. .  In other words: the exact source object of the object name you’re searching for.. .

Find all references

Pattern: 

'\w+ (\d* )?"*'

Setting in VSCode:

"CRS.SearchObjectNamesRegexPattern": "\\w+ (\\d* )?\"*"

// I just removed the "^" from the default setting, which indicates "search anywhere"

This pattern will search any occasion in code – which means: also the variable declarations.  Let’s say it’s an alternative “where used”.  I won’t set it up like this as a default setting, but I might just change it ad hoc in the search by simply removing that character.. .

Find anywhere in the name

Pattern:

'^\w+ (\d* )?"*(\w+ *)*'

Setting in VSCode:

"CRS.SearchObjectNamesRegexPattern": "^\\w+ (\\d* )?\"*(\\w+ *)*"

// basically added that there could be multiple words before the searchstring.

This pattern is somewhat more complicated, but if you would not rely on your search term being the beginning of the object, but rather “somewhere” in the object name, you could use this one.

Enjoy!

Indeed – I didn’t see any official statement yet – but it’s obvious, v16 is the current latest MS release .. and if you don’t believe me – just check Docker (the latest “current” release is already v16 – docker image “mcr.microsoft.com/businesscentral/onprem”)…

I’m not going to bore you with what is already online – I will simply point  you to the resources I could find today:

• Dynamics 365: plan voor wave 1 van 2020
• What’s new and planned for 365 Business Central
• Business Central on Docs
• Business Central on Learn
• The download link for Wave 1 (still under construction when writing this post)

And – of course – the post I wrote about it earlier this year also contains some information and links ;-).

Expect work to be done!  At this point, I’m moving our 18 apps to v16, and want to comply with the bunch of extra coderules and new concepts Microsoft has foreseen in this post … .  I promise you – a LOT of work :(.  But more about that in a next blogpost ..

You remember this post?  I tried to warn you that when v16 comes out, there will be a new code rule that will check your filenames – and you’ll have to (if you don’t disable it) comply with the file name convention of Microsoft.  If you don’t automate your file naming, then you’re in for some .. uhm .. challenges.  I just made sure that the automation of the filenames complied with Microsoft’s rules .. .

I need to correct my store in that post though.  I had been working on this “RenameWithGit” setting, which didn’t work with multiroot workspaces and had some other stability problems.  Only after my post – thanks to a reaction from James Pearson on twitter – I learned there is a much simpler way to do this.

First of all …

Forget about the “RenameWithGit” setting

Indeed – just forget I ever built it.  I’m actually thinking of taking it away in the near future.  I already removed it from all my workspaces, and I strongly recommend you to do the same.  It doesn’t work like it’s supposed to work .. and I’m embarrassed enough about it ;-).

There is only one word you need to remember ..

Staging!

All you have to do after you renamed all objects is to stage the changes.  That’s it.  And then you’ll see that actually everything is just fine.. .  I found some kind of description about this ability here: https://stackoverflow.com/questions/29706086/how-to-stage-a-rename-without-subsequent-edits-in-git. 

If you’d rename a file, you will get a delete and a new untracked file in Git, like you can see here:

When you stage the file in vscode, you get the rename:

What it actually does in a stage is it will compare the files, and when more than 50% is the same, it will indicate it as a “rename” it in stead of deleting the old name, and creating a new file.  That’s smart!

And yes, indeed .. I have been immensely wasting my time on the “RenameWithGit” setting :(.

Can I make sure everyone always stages before commit?

Well .. It’s actually good practice to always “intentionally” stage.  You must have seen this message already:

In VSCode, it’s called “smartcommit”.  But honestly, in my opinion, the smartest commit is an intentional commit.  I don’t like this message, and I switch it off by setting this up in my user settings:

“git.suggestSmartCommit”: false

I’m not forcing you to do so .. but this way, you can easily check in VSCode if the rename of the file,  was actually a rename of the file – and not deleted and new files.  Like it was intended. 

So, what is now the safest workflow to rename my files?

Quite the same as I mentioned in my previous post about this – but a bit different. 

1. Create a new branch

 Yep, I still recommend to do the entire process in a separate branch.  Of course.  It will give you a way out if you messed up ;-).

2. Change the setup

The setup is actually very similar as in my previous post, only now with the “RenameWithGit” = false.  To match the file name convention of Microsoft .. this is what I would use:

"CRS.FileNamePattern": "<ObjectNameShort>.<ObjectTypeShortPascalCase>.al",
    "CRS.FileNamePatternExtensions": "<ObjectNameShort>.<ObjectTypeShortPascalCase>.al",
    "CRS.FileNamePatternPageCustomizations": "<ObjectNameShort>.<ObjectTypeShortPascalCase>.al",
    "CRS.OnSaveAlFileAction": "Rename",
    "CRS.RenameWithGit": false,
    "CRS.ObjectNameSuffix": " WLD",

Alternatively, you could add the “RemovePrefixFromFilename” or “RemoveSuffixFromFilename” – but make sure you set up the mandatoryAffixes-setting in the AppSourceCop.json as well, for the CodeCop to accept the removal of the prefix or suffix.

3. Commit

This commit is there because you might want to revert after a failed rename-attempt.

4. Rename all

This is the same “Rename All” function I talked about in my previous post:

It will rename all files of the current active workspace (the active workspace is the workspace of the currently activated document (file)) – not all workspaces.  So you probably have to do this for all workspaces separately.  Thanks to the “RenameWithGit” to false, I expect no mistakes.  I was able to apply this to +6000 files today .. so it’s tested, I guess ;-).

5. Stage

This is THE step I was talking about earlier.  Here you can check if the rename was successful – all renamed files  should indicate an “R”, like:

When you see that – you’re good to go and …

6. Commit

Just commit, push and create a pullrequest to the necessary branch .. and you should be done!

Wait .. So I can do this in a multiroot workspace as well?

Yes indeed – this flow does work in a multiroot workspace.  Do execute it for all workspaces separately though, like mentioned before.  That’s how I implemented it.. . 

Conclusion

It’s a piece of cake.  Really.  So just do it, comply with the naming convention, and don’t feel like you’re cheating on Microsoft ;-).