Microsoft 365 Community Development, 8 Feb 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: PnP Core, PowerShell and Provisioning Engine SIG - February 8th, 2017

Description

Recording of the PnP Core, PowerShell and Provisioning SIG from February 8th of 2017. This is bi-weekly special interest group call concentrating on the mentioned topics. You can download invite from following address - https://aka.ms/sppnp-core-sig-call

A

Let's actually get moving so welcome everybody. This is the special interest group on the PMPA sites, core PowerShell and BMP provisioning engine. Like mentioned I'm here with Burt Johnson. We are sitting in the same room at treatment, and this is actually pretty rare that this happens so because, typically, we are sitting outside well we're working remotely from home.

A

So let me sorry there's a different. Let's get moving in a second there's, some people turning to the wrong setup or on call we have some issues related on the is the invites and sorry for that. Will that has been updated on the site, so that should be good, but let's actually get moving on. The right discussion is just to ongoing calls right now. So today is February 8 2017, and what we wanted to do today is is quickly have a look on on our high-level statistics from January. It might be something with not.

A

Everybody is super interested, but there's some conclusions out of that data. So we wanted to actually code that true as well before we get there quickly, explaining a shippin PMP, its own PI, SharePoint engineering. So me personally, I sit within the well I work in a team who is responsible after all, of the api's and development stories within the SharePoint, and we do own the PMP. But we work together with community MVPs and non MVPs on the PMP open source elements and components which we then provide for everybody to consume.

A

There's code samples harness documentation, x' community calls dreams are on SharePoint framework, errands craft of choices to 5 API development. This particular call is really around to be empty core component and the power shop so really concentrating those different aspects. There is a separate special interest group calls for SharePoint brain work, so if you're more into that client side, development and the SharePoint framework, which will go pretty soon into GA and that's a separate bi-weekly call as well. So for this particular call.

A

What we wanted to do is really do a quick look up on the BMP usage and conclusions, and these are actually slides which, with quencher with our second engineering senior leadership yesterday in our steering meeting to kind of a show how people are using the PMP or what how people are using samples in general in SharePoint, and we true some conclusions out of it as well, and we wanted to have full transparency with everybody who's using PMP stuff.

A

So, just working through some of those things with you as well feel free to ask any questions as we move along and then on the second part of the call we will have two demos, so there's a demo around TMP provisioning engine, refactoring bipolar powers can actually walk through some slides and explaining why we are doing this and what's the objectives of doing so, and then there's a demo on the latest PMP partial changes done by the father of the PMP PowerShell. Sorry, everyone who ninh or mr.

A

ban like we call him internally or in the team. Mr. van good and sorry I've been faking that one out as well, but let's actually move on on the on the thing. So yesterday we had a p.m. PST recruit meeting with the SharePoint engineering senior leadership, and these are the slides which we weren't sure they're. Just sharing with this open man getting it recorded as well, and so people can actually have a look on what has happened within the last quarter mainly well.

A

This is a few marks one month more than a 1/4, but obviously we are tracking, quite detailed, what's happening from a PNP and in general on a SharePoint community sign and what we have been doing, how things are improving? How much people are watching our videos? How much people are using our components across tenants? I will go through this relatively fast, so that we have time for the powershell and the engine discussions as well, but just giving you some insights on the statistics, what we are collecting on a youtube site.

A

We have released 49 videos during the past few months and overall watch time during these months is two hundred and twenty nine thousand minutes, which is pretty impressive. The SharePoint framework tutorial 1, which is explaining how to get started on SharePoint, is the number one video with 24,000 watch 10 minutes, which is critical. The numbers are pretty impressive.

A

On the MSDN sites, we have released few articles around the PM, sorry, the modern sites and list customization story. We do know that that's not necessarily sufficient for all of the business requirements what people might be having and we in SharePoint. The engineering also are improving those capabilities in the in the future, but that's really due to modern customizations modern sites, and this customization articles really boost up the BMP MSDN site usage and it went up to 44% from September on the dev tosh.com.

A

We've done, 19 new weekly webcast, 6, new blog post and 4 monthly releases, which are essentially always going back and looking what was happening in January as an example or the previous month. And then you can see what are the samples and changes, what has happened within the engine or in PowerShell or in the indifferent other solutions in the BMP.

A

The next monthly release will happen on this Friday. Most likely communications will go out on Monday. We have a monthly communique, a monthly community call on next Tuesday. The reason why the communications will most likely be postponed for Monday is that we're traveling across the pretty hectic time zones again during weekends. So we might be running out of time on the on the Skype, with business.

A

Obvious we've been having this by weekly calls and special interest group calls and that's pretty the participant rates, for this are increasing all the time, which is pretty good for both SharePoint framework stuff and also the p.m. decorum PowerShell on the key top. The numbers are growing quite fast as well, so we are up to thirty three thousand seven hundred and forty Forks, all together with BMP Reapers, which actually makes the BMP just BMP main repo has like two thousand seven hundred folks.

A

Well, that's a number eight in the overall repository size in Microsoft, which is pretty impressive. If we count the all of the forks in the PMP, which is the photo for 3,740, we're number two in overall list of Microsoft kid operators from a size perspective, we could just look on the Forbes, which is super cool. The PMP no get downloads is up to 66, 68 thousand or 48 percent unique visitors up to 28 percent, merge, pull request, 25 percent overall there's 247 unique contributors across the repos, and we have more than 10,000 contributions.

A

Contributions essentially means individual commits in a kit hub, but we're tracking all of this, obviously in our elevator data's way. So we know how much we can hear how much we should be investing on this area in the future and obviously will increase investments because members are looking is truncated. In January we had 1.3 billion HTTP requests going through the PMP core components and the MP powershell, and that's super impressive. So we increased that by 44% from December, so the growth is pretty huge and then.

A

Tenants usage we're in 3700 tenants on on the core components for provisioning. We missed the 1,000 objective with the two tenants, so we had 998 tenants using using our provisioning engine and then on the jobs. There's four hundred one hundred and forty four webcasts and communication presentations currently available, and then the MP part in fact has been used in 133 tenants during previous months, which is pretty cool as well good.

A

Some other kind of a conclusions out of this data before we go to the demo section so like mentioned out there relatively fast, unless there's any questions on this one, so the added model, clearly the usage of padding model and the core components and PMP PowerShell is increasing. We can see compared to August or September to January to increase this relatively big. The December always goes white laid down because December is is a holiday season and we in the PNP we're kind of over to go.

A

We are so big that the holiday season, impacting the SharePoint Online in general, is visible within our charts as well. But it's good to see their usage is growing and that's no doubt our objective related on the BMP as well.

A

Adoption across the world is crowing as well, so we can actually see that the PMP BMP power shop and the BMP core component usage is relatively big and it's across the world as well. Us is no doubt in up to number one if we would combine the Europe as a single country which, obviously it isn't, it would be pretty close one to one of close where between the US us is obviously a big country.

A

So there's more tenants, as we can see from the charts, but it's nice to see that the PMP power show and the core component is used and the BMP stuff in general used across the world widely, which is critical on premises still very relevant. We're getting this feedback all the time in the issue lists under discussions, and we can actually see that the on-premises demand also in the new care packages and downloads. So one third of the NuGet packages are our downloaded for the on-premises data, SharePoint, 2013 or Chevron 2016.

A

The checkpoint 2013 is no doubt thicker than SharePoint -16 Sharon 2016 is crowing, but we do not have obviously detailed statistics from on premises because we don't have access on on those environments, which is absolutely understandable. The previous stats are coming from SharePoint Online and because we are shape and engineering, we actually are able to get that gate out from a SharePoint Online share.

A

Pod community is truly engaged I kind of showed some of these numbers already so 247 unique contributors and 10,000 contributions, which is pretty cool and the upper chart section of the slide is showing issues at it to a repo, and we can say that there's a lot of activities and discussion is ongoing there and maybe kind of a mention on that.

A

One already that we've been doing a lot of cleaning on issue lists and we've been really kind of tackling our issues going through back historical issues, making sure that they're fixed, if they're not will reactivate the stuff and all of that so we're trying to get on top of things as fast as possible. It's a reasonable thing challenged.

A

Obviously, because none of us, like you guys or of course you just don't- have enough time to do off the jobs, but we should be able to do for requests by month and a repo it's a pretty hectic chart as well. So we can't be up to 69 pour requests for p.m. pay powershell in october, which is pretty huge or then the temporaries was relatively or much lower. So December we had six to seven probe requests overall and in January 198. But okay, as the probe requests are well below costs are kind of a challenging.

A

We can't really estimate how long a one poor request takes to resolve, because one protest can actually cause challenges or it might be a really simple fix, and obviously we want to process them in a detailed way and make sure that none of the in production deployments will be impacted by the provider. Changes huge, a few more slides before we go to the demo section as well. Side provisioning is critical functionality so, and we in obviously in SharePoint engineering, are looking into this one as well.

A

So if I switch my hat from a PMPs side to the SharePoint engineering side, the PMP provisioning engine is used currently by engineering. As the let's say, the inspiration for the future model, how the SharePoint will work related on templating, but that will take a while to actually get up and running and implemented. So the PMP provisioning engine is no doubt currently the model how to do stuff in a best way in a SharePoint online or in on-premises as well. If you want to avoid on privacy challenges in on-premises as well.

A

In general, we hit nine hundred and 8 tenants using the PMPA provisioning engine in January, 1 2 3 of the requests are for provisioning. 1/3 are for extraction, which is kind of understandable as well, and what we can see from the data which is getting collected on SharePoint Online is that the PMP provisioning engine is quite often used as a help for the migration.

A

So, quite often, when there's a new tenants coming in there's a spike on the HTTP request, usage, true PMP, and then it settles down in a much less, which indicates that there's a my creation happening for that particular tenant, which is pretty cool and no doubt the PMP provisioning engine is a great tool for you to set the slides and configurations in SharePoint Online to have the content types to lists and everything else. The PMP does not do migration.

A

So for my creation, you would have to use then a third party migration tooling, and we're not looking into extending the BMP natively to the migration itself. Do numerius reasons its sounds simple, but actually it's a super super complex scenario. So, third party migration, solving it stays the way to go a few more slides kind of additional slides. So PowerShell is a common extensibility model. This is something which we from a PMP aside.

A

We need to teach also sharepoint engineering on this one, because obviously they don't quite understand that how people are using the PMP powershell as the customization tooling toward sharepoint online, even though you you're doing scripting, but actually you are from provisioning a site you're, applying the branding you're applying to the lists and content pipes and all of that using the PMP powershell. So we hit 1000 1300 tenants with PMP PowerShell use its in January and 100 million. It's the DB record requests going through and by looking into by the way, maybe on this one.

A

Some of maybe on this one just to mention this kind of a super cool I had a few meetings yesterday with the right people here in Redmond, and we are actually actively looking into combining the native sharepoint online admin powershell with the empty powershell. No, this is super early, so it doesn't mean that it will necessarily even happen, but there's a cream lights to move forward on on thinking. How do we actually make that happen?

A

Because, right now we are still getting a lot of feedback from the community from you guys in the school that it's kind of confusing. Why do we have sharepoint online admin? Shell for some operations, that then we have the empty powershell for content level operations? Why wouldn't we actually combine these as a one super powerful, sharepoint online comic, let's set and then do that as an open source, so we're looking into practicalities of making that happen. So awesome news from that perspective.

A

And then a few conclusions, part of our popular articles and videos, so planning in sharepoint, is a big ask among customers. We absolutely get nuts, so we can see it up directly from the samples and articles and MSDN or the youtube videos as well customizing. Modern experiences guidance is that there was a high interest on that one as well, and obviously we know that that's currently slightly limited, but we are looking into enabling additional scenarios as we move along during the spring as well.

A

So there will be additional supported scenarios for modern, modern team sites as well. Sharepoint framework webcasts absolutely are pretty popular, which is understandable, not directly related on this sake, but it's nice to say that people are under kind of a catching up on hey. This is something what's coming up pretty soon, I can start using the client-side web parts relatively soon in production.

A

So it's good to catch up on on, what's happening on that side as well, and on-premises articles and videos are always highly popular, and this is a good indication for us in the SharePoint engineering that we really need to make sure that we are people on doing investments for on-premises as well, so making sure that we keep on evolving down premises.

A

Getting some of the the modern experiences like the SharePoint framework will land on is shipment 2016 environments during this calendar year, but it's a good example a reminder on the fact that hey we need to keep on doing that as well. Not all people are in SharePoint, Online and kind of a similar kind of thing is visible from our sample site. So we can actually see that on the right side, we have the most active samples between the December and January and the SharePoint UI. You are responsive packets. It's number one.

A

It's always been a number one every single month about the bad since it was actually released. So the planning is a big deal. No doubt the core component is their office. 365 Star, Trek intranet.

A

Is there as well from Frank corner MVP from Canada and I'm, not sure if Frank is actually on the call and by the way on that one there's a a new version available as part of the upcoming release already in a de France, which is enabling that scenario to work in SharePoint, 2013, SharePoint 2016, so Frank has done an awesome job on that one core event: receivers is pretty high or remote event receivers, that's kind of a surprise as well, but it's good reminder on on that requirement.

A

Pm PHS core mission is quite often on the list and so on so gives the feed data, and this helps us in the BMP and SharePoint engineering. Let's say: focus our resources on the right elements in the future as well.

A

Good analogue nation. Actually, as part of the community engagement, we've been doing a lot of cleaning in a kid's hub on the issue. Lists we'd be moving some issues across the Reapers and that's really to make sure that the core component issues are in a core component: repos that they're not getting lost anywhere else and then in a core component site.

A

We've been doing a lot of cleaning as well in general, and we keep on doing that in the coming months as well, because we want to get on top of things, and we want to make sure that things are whenever there's an issue for you guys. We are able to answer that in a matter of days rather than in middle of a matter of months, so we'll address on that one cute, but I think it's time to see something in practice. So we have 40 minutes.

A

I really really wanted to kind of a culture that the high level usage, which you guys to understand that hey. If I'm, considering using BMP powershell PMP a provisioning engine am I the first one. Well, no there's quite a lot of other people are using those as well and that's a great thing.

A

But the more people are using these reusable components and the more resources to BMP will get and the more higher quality the stuff will be, because we can actually help the people to address the issues which they might run into with separating and in general. Maybe on issues, even though I'm always mentioning issues, we do have automated daily unit testing with 450 real functional tests, not just unit testing. Actually so, and things are looking fine for the February February release, but I think that's it. Let's switch the power first, we're.

B

Thinking, yeah, ok, much much shorter, ok,.

A

Nobody's nervous so, let's actually switch to urban areas. Gonna do a quick update on the PMP Pavo show what has recently happened and starts sharing the screen whenever we get that.

B

There we go and.

A

Still loading and retinyl, it should be very.

B

A

Oh yeah, yeah I can imagine.

A

B

So I'm just gonna talk you through a few of the the latest editions that we have in in our approach, our library and so one of the things that was lacking actually for quite a while is the taxonomy support.

B

So we now have a couple of more command posts there to help you with term sets term groups and terms simply use is just get P and P term group press ENTER and you get all the term groups. So if you want to do a specific term set from a crew, P say, for instance, group Hey and that's completion there we go and high fived s to P and P term set I got the term sets in that group same thing you can do with terms I can also get to a specific term.

B

So if I will see all the terms in the terms have two cat, PB term term set B II intuitively names term group, the first couple bit down there. We have all the terms in that specific term set so they're, equivalent commandlets for creating terms for creating term sets and for creating term groups, removing terms, etc, etc. So they are all there now, so that makes it a bit easier.

B

The older command is to use that path, structure where you say, get speed, B taxonomy item, and then it has a term path which is B, BBB AAA, and then you get the thing in okay, I made a typo, they apparently do it in quotes, helps.

B

Then, okay for joins nothing actually, in this case, I want to look into what's happening here, but that's the idea with the olds command lists, and they didn't really work that well, they were not a very intuitive. This makes more sense. In my opinion, another thing that I want to introduce veces I think it's quite a good addition, and- and that is an issue that many people ran into with the commandlets, and that is that the collection is not initialized error message.

B

So, for instance, if I show you an example, so if I do pin P list, you get all the lists and I want to specifically in this case the tasks and I put it into a variable, get PP unleased tasks.

B

So here's my tasks list and if I want to have the views from that list, I could just do L views I get an error messages now. This is not something you did something wrong, but this has to do with the way season. Behind the scenes works, it has to do with the fact that we have to explicitly request in when we're loading that list from the server that we also want to views to be included. But we don't know upfront what you want.

B

So what we did is for a bunch of command is not for every commander, but for those were relevant we now introduce an includes, attribute or parameter, and this one is. Completion has completion, so you can use tap in there. So, if I go to the views now, this object will be loaded, including the views so now I can say doll, L use. There is my views of this specific list.

B

There are certain caveats there we go to levels deep because she can sometimes go indefinitely deep because, for instance, all this can have a parent resistor web. The web can have another lease connection, etc, etc. You can you know, there's no end to that of many level. Tp can go so we go to the two levels, deep from some of them.

B

Only one level deep, depending a bit on the object, but the includes parameter, is a very nice I would say addition to a bunch of the commandlets that there and we have been updating the help. Obviously, to tell you that the command that supports that, so there's a realize thing and that's actually what did these formats? Yes, one more commanded I want to shorten it's been written by Nicolas pencil from Norway and he's he's very much into shirts and he created a submit PMP search, query commandments.

B

So if you do submits PP search, query and you do a query and I write tests, it will actually just execute the search for the server and that comes to search results. Objects were there that you can use to consume and do things with, and this is a generic search, query you're executing and you can have a paste results and etc. So this is a quite a nice, like quite a nice commandlets, that's very useful to execute services straight from our shell.

B

Another last two commanders that I want to show. You is the related to side closure. So there is a get PNP side closure. Every returns to status, you can wit set, pin P side closure set P P side. Closure can I set aside to close, but there is no side policy, so you will have to M yeah and play around with these, because they look set PP side policy to give you warnings, but it allows you to close an open insight using the policies and that's it from PowerShell.

B

This will come out after the master, merge.

A

After they must dimension, so yes as part of temporary rates- yes, indeed so right now, these are not in the latest version which release with us out, but after the February must emerge which will happen on Friday weekend. Monday will pay the communications. All of this result at that point, any questions for Erin on that one.

A

Just out of curiosity, if there's anybody who who just has a question on this, this is super cool by the way so, and thanks for me Kyle that the first thing is is kind of like somebody put it in there or cholo put it in there in the I window. It's pretty sexy, it's that useful. Well, maybe there might be scenarios, but it's pretty cool.

A

No questions at this point so, let's actually move to Paulo miss if Paulo can actually switch start sharing. Paulo has few slides and then I a live demo on the engine. Refactoring. Yes,.

C

A

Have some time for Q&A.

A

C

Thank you, visa and welcome everybody. I'm gonna share my screen to show you just a few slides about the refactoring. We are applying to the way we handle provisioning templates as the realization. In this realization, then I will show you a quick demo of the refactoring in practice, and we already talked about this topic.

C

I think it was about one month ago, more or less, and we were planning to do something using a automata which is an open source library available on database like SP, MP and all the stuff we provide as PMP and now I will explain you what we decided to do, but before doing that, I prefer to introduce you the use case and why we wanna do some refactoring of the for matters in the provisioning engine.

C

So, just let you know whenever we play with the provisioning engine and whenever you apply a provisioning template internally, the engine uses a doorman model to represent the template that you will apply to the target site or that you will extract that from an existing site and every provisioning template is a complex domain model object made of a bunch of complex types like, for example, collections of property bags and trees, a collection of site fields of content, types lists and so on and so forth.

C

And, for example, the collection of lists is made of a of a an array or of a list of lists, instance, objects and every single list instance object is made of a title of a template, type and many other properties.

C

So whenever we implement a new capability, our new functionality in the provisioning engine, we create a new object handler, which will be a an object which will be able to handle a specific object like, for example, a list instance and we'll be able to create or to update a list instance based on the definition that we have in memory.

C

However, from a persistence viewpoint, whenever we want to persist a provisioning template, we can use multiple persistence techniques and formats like, for example, a JSON, serialization and XML based realization or eventually a dot BMP file, which is, as you probably know, a kind of a zip file with still an XML provisioning template inside of it.

C

So whenever we think about the dominant model in memory, we should also know or be aware of the fact that on the earth side, we will have on the persistence layer, for example, an XML file like this fake one I am showing you on this slide and, as you can see, there is a provisioning element. There is a provisioning template element with a collection of site, filled the content, type and list elements as still a child element, and so on and so forth.

C

So whenever we want have lured or save a provisioning template from persistent storage, we need to load, for example, the XML from the file or from a library in SharePoint into the in-memory representation of the domain model. And right now we do support quite a lot of different XML schemas to represent the provisioning template, because whenever we release a bunch of new capabilities, we also need to have, from an external perspective, the capability to describe to define those information within a template object and within a template XML file.

C

That's why we have so far five different schema versions. Some of them are already obsolete, but still available in the provisioning engine.

C

Maybe sooner or later we will remove them, but you should move to the new ones if you are still using one of those that are right now obsolete and each and every single version will have a different XML schema, so a different definition of the structure of the XML, which is compliant with the version of the engine for which the schema was built right now, the current version Rd 2015 12, ax and D 2016 5, and we are working on a new version which most likely will be the 2017 0 3.

C

So next month in March, we will release a new version of the schema which will be supported by the engine, probably in April, because usually we release a schema and then afterwards release the new capabilities at the engine level. But whenever we release a new version of the template or a new version of the schema, we also need to make the engine able to read and write that XML format from the persistent storage into the domain model and founded our model back to the persistent storage and just to make an example.

C

If we have again an D XML we saw before we need to have something which is called a formatter right now, which will translate, which will read and do a kind of a parsing on the XML file in order to create the provisioning template in memory which will be processed by the engine and whenever we want to save it.

C

Back on the persistent storage, we will do the opposite path, so we will store the position template through D for matter back to the XML format and right now and so far we have five for matters one for each and every schema version that we do support so pretty clear but somehow complex to be maintained and managed in sector from a maintenance perspective whenever we want to do maintenance of these four matters.

C

First of all, we need to think about the fact that we have five different classes right now, which will become six seven and more in the future whenever we will add a new schema version, and so the current model implies to have quite a lot of files, quite a lot of classes in dotnet framework and each and every single class will have quite a lot of code inside of it just to handle the serialization in the visualisation of the XML, from XML to the domain model and back from the domain model to the XML.

C

All these four matters right now are quite complex for from a testing perspective. So if we want to do testing or automated testing or if we want to validate the behavior of the formatters, it is not that easy to be very modern through applying the testing rules against a unique big chunk of code and just to wrap up. We realize that the original idea of having a formatter for each and every different schema version is really painful from a maintenance perspective.

C

So we decided to replace this solution with something much more maintainable of saying, much more testable from a development perspective. That's why we are now introducing the idea of having a bunch of serializer objects where we have one civilizer still for each and every schema version. But, as you will see pretty soon, each civilizer is really really simple and mainly based on a base abstract class, which will have the real logic inside of it.

C

We use a lot of object-oriented programming inside the definition of these new civilizer so that we can reuse as much as we can the same code to improve maintenance and to improve code quality and, moreover, we are using a bunch of reflection to make it possible to have or to reuse the same syntax in the same code against different versions of this schemas, targeting, of course, the unique domain model that we have in memory.

C

This approach allows us and will allows us to do proper testing and to be very modern, while approaching the development and the maintenance of the serialization and deserialization of the template the very beginning. Last month we try to use the automatic or engine, but I think I can say that way that we somehow failed using and I failed, because it was my choice to try to use the ultimate library, which is totally ok.

C

I mean it is really good, but unfortunately it doesn't work very well in the context we have, because we need to do mapping between one domain model, which is the one of the provisioning engine and, on the other side, multipolar schema models which can be eventually totally different from one schema to another. And so we can't simply rely on mapping objects based on the property names.

C

But we need to do something more and that's why we decided to create a kind of our own homemade TMP outer mapping engine which will do the mapping keeping in account all the specific needs that we have in the provisioning engine of EMP. So stop talking and let me move to the visual studio environment, to show you what we did and to show you where we are so far just to be clear.

C

What we are going to see right now will not be in the February really is, but it will come out in the March release ready to support the new schema. So we don't need- and we don't have time to really is these new refactoring by the end of this week. But to be fair, you don't really need it right now. Everybody will need it as soon as there will be a new version of the schema out there, because right now we already support properly all of the schema versions that are supported by the engine.

C

Ok, so let me move to Visual Studio. Let me show you what we have done here. We have the solution of the PMP sites, core library and within the framework folder under the provisioning section. I have the providers area in which we have the JSON provider and the XML provider. All these challenges apply just and only to the XML serialization.

C

Okay, because from a distant perspective, it is much easier to handle the serialization visualization of different kinds of objects, but from a maximal perspective, we need to support different schema versions, and that's why, right now, we already have all of the SSD files for all of the schemas that we support in the engine and right now we have all of the format, as I told you before now with the new refactoring, we also have a maximal PNP schema based serializer, which is an abstract plus generic, and the generic type defines the type of the schema base.

C

The provisioning template definition. What I mean? Let's make an example, we have this schema design 16:05.

C

Well, we also have a c-sharp a-- code file in which we have a provisioning type, as well as a provisioning template type, which is exactly the counterpart from an XML perspective in the schema version, 2016 0-5 of the provisioning template type that we have in the domain model and the same patterns for every single version of the schema of the schemas that we support.

C

So whenever we want to play with the new serialization engine, we will start from a base class which will have to be inherited by a concrete class which will have to specify to provide the type of the XML provisioning template that it will be handled by that specific sterilizer from a backward compatibility perspective. This new basis civilizer, is completely backward compatible with the already existing architecture of the engine. So from an end user perspective, all of these refactorings will be completely transparent. Ok, so we will not break anything.

C

Hopefully, ok, but that's that's our that's our goal and that's why? If not that easy, it wasn't that easy to do all the refactoring. So we have a base class, which is an abstract class. We have a bunch of code inside this class which will be reused across every different version of the schemas that we want to support. For example, we can have a unique place in the base class in which we do the XML validation, so in which we do the validation of the XML file against the schema.

C

That relates to the specific version of the XML that we wanna serialize and deserialize. We will have some code to do the processing of the input stream whenever we need to read a stream from the persistent storage and we want to convert it into a provisioning template, and so we have a bunch of code which will be reused and define in a base observe class.

C

We also have a couple of methods: the provisioning template, which is the one that starting from a stream, will give you back a provisioning template, which is the dominant model, object we have in PMP, and you will also have the to format, a template which gets as an input the provisioning template the in-memory domain model and will give you back a stream that you can store in a persistent storage. These tomatoes are those compliant with the already existing infrastructure, but inside each of them we will have the idea of this serialize and serialize a template.

C

So we have a new method called visualize template, which is a vehicle one so again with a basic implementation in the base class, which can be overridden in a inherited class as well as we will have a little bit lower AC realized template, which will be again the a vehicle matter that you can override in a specific, concrete serializer, for example, for schema 2016 0 5 come on. Ok here it is I. He married from the base of third class I, showed you and then I say this guy handles the version.

C

2016 zo 5 provisioning, temp 8, which is the one I showed you before in the C shop file automatically generated from the x SD file. So within this serializer have just a few lines of code just to define what is the XML schema to use for validation? What is the XML namespace? What is the XML namespace prefix and basically nothing more than that, and if you like, just as an example, you can do the override of the previous methods just to add some custom logic in this specific version of this réaliser.

C

So whenever you we will create a new version of the schema. We will simply have to create such kind of class, which is really easy and simple, and all the business logic will be inside the base class, which is way more better, of course, within the base class.

C

For example, let's have a look at the creation of a provisioning template in memory from a stream. We will have some basic crumbling here, for example, and then we will have the deserialized impregnated, which can be over I double as I told you well, this method right now we'll use some reflection to go through all of the objects that we can define in the engine.

C

Implementing the IP MP schema serializer interface, and this interface simply has ad serialize and deserialize method, which will allow you to get a an object which will be the representation of memory of this XML, and it will give you back the template from a dominant perspective or you can serialize a template into an XML in memory object which can be then saved onto the persistent storage.

C

So, for example, think about these template, in which we have a bunch of information, the provisioning root element, the provision template the property back entry is, and the list instances right now. For example, I have few serializer already implemented in this new engine and, for example, I have a sterilizer that will be able to handle the list instances. So whenever I will have to go through all of the list. Instance, elements of this collection of Lists I will use a dedicated serializer. They related to doing the D.

C

Serialization and deserialization of the list instance objects, so that I will have a unique place to handle and to manage whenever I wanted to testing, whenever I want to do maintenance related to the handling of least instances inside the template and whenever I declare that I have a new seller. Serializer I have some base classes to improve the code. Quality and I have an attribute that I can apply and have to apply to each and every serialization I will create so that I can declare. Whenever I will do.

C

This serialization apply this guy with a specific sequence number, which will be compared with the sequence number of all of the other serialize ur I will have same story during the serialization, and I can have different sequence numbers in the civilization and in the visualization phases.

C

This serializer target D 2016, 0 5 and the 2015 12 version of the schema. So we can use the same object to do the serialization and deserialization of two different kind of schemas and it will be used by default whenever we will do the serialization and deserialization of something related to the provisioning engine.

C

The same story applies and will apply to all of the serialize or objects that we will have and simply by writing one by one, all of the serializer that we need to handle all of the content we have in the xml, so one for the list, instances one for the content, types, one for the set fields and so on so forth. At the very end, we will have a collection of serializers which are well testable and maintainable, which will be invoked dynamically using reflection by this new civilization engine personally.

C

I think that this result it's a great improvement. If compared with the previous option, we have- and we still have right now in the released version of the engine- and this result is based on the efforts of all of the members of the core team, because we made a lot of meetings. We share the proof of concept prototypes, especially me, and there we played a lot around all the available solutions and we ended up with this idea. Of course, maybe you like it? Maybe not!

C

So that's why we are sharing with you in this special interest group. The overall outcome, in order to know about your feedbacks, to know if you like it or not, and if you have any different proposal and especially also if you have a time, availability to help us on writing all of these civilizer. That would be very, very appreciated, because, right now, we will have to stop writing about between 35 and 40 different serializers for all of the complex types that we want to handle with the provisioning engine. So.

A

You said, if you have time to implement all of this I realize that we were through asking helping us on building any of them for some yeah yeah, absolutely.

C

Absolutely sorry, it was.

A

C

Few of them we can do team working and if each and everybody in the community who has time to contribute one average just once you're either it could be really really helpful and can speed up the process of releasing the new version of the serializers, which.

A

Will then speed up the time to be able to release the new version of the schema, which will then have additional capabilities and.

C

Sorry visit and as I wrote on the slide, if you think that you don't have enough time to write the serialize of no worries, we will take care of them, but it would be really really helpful for someone also helping us to write the test methods for this realizers, because we would like to completely automate the testing of those civilize errs and include those tests into the automated tests that we already have for PMP.

A

Absolutely and one thing, but I wanted to do a pollination this quite a few times already during the presentation. So if you want to use the BMP provisioning engine, you actually don't need to understand anything of what we went through. So this is say, deep dive in the how the engine actually works and how we're planning to implement some changes in the core of the engine. Just if you consume the engine, it's you don't need to actually vary about this one.

A

So just just to in face that, yes, it is long yeah, but if there's any any input on this one seems like a level 500 presentation.

A

If there's any any feedback on this one and really what we're looking for this refactoring is too simple: to simplify the maintenance for the future and make it more flexible. We kind of a Crone to this current status with with the old design, but but it's getting more well, it's not as optimal as it should be.

C

There are some questions in the chat area: I don't know if I can handle them. If you want to help me on answering them, these are but, for example, I just saw the latest one one about JSON instead of XML, why not using just JSON instead of XML?

C

Well, we we think that adding also the capability to support XML is great, because most of the SharePoint developers are used to use XML and we try to make our best to make it possible to copy paste and sometime slightly change the XML coming from the old-school feature framework and having the same capability in the new provisioning engine. So having the support for XML, I, think I would say.

C

We think if I can say that that, if still a killer feature of the engine, because it makes it easier for SharePoint developers to migrate from the old school development to the new model, yeah.

B

Versus four fields, we don't have our own schema. We just use the existing field, schema from chef. Went there exactly.

A

And also in general, when we are looking into the template, if it's serialized, the JSON or XML XML actually gives you the details, what does it actually contain? It's more structural, but it's it's a fair point that, yes, you are already right. Now you can serialize stuff in JSON as well, so yeah.

B

But I also I wanna, actually I've never looked into that. But if you can like validated chasin according to a schema, never losing to that, but you can read XML, so we can actually make sure that what we're parsing is valid. Xml is a valid template.

C

I'm looking to the chat area just to see if there are any more questions with Royce.

C

No I don't see ya.

A

Bill is pointing out the validation of XML as well and by the way bill. It's nice to see that your font is changing for every single message. That's cool, never mind.

A

But anyway, so we'll keep on working on this refactoring of the engine, we'll get it we're killing moving. Here's some examples, we're moving on there on the spring time, and this is really for making the engine more flexible for the future as well, because, yes in the SharePoint engineering, we are looking into creating similar kind of a recipe model for sites in the future, but it will take some time to actually get it up and running and for the engineering, the ketchups, what they have within the PIAT the engine. So that'll take a long time.

A

So it's cute, but if we don't have any questions from this one we can flip may be back on the on the chin or a Q&A. If there's any random questions on the on this core component, the PowerShell or in general on SharePoint developments more than happy to help on those from the audience, and if you want you can absolutely flip to audio as well. If there's any input what you work in, we have six minutes, not a massive amount of time, but still.

A

No issues with anybody, that's cool.

A

And in general, let's see if we get any questions on the iron window or in Oriya, but like mentioned this, for the February must emerge will happen on Friday in Akita Preppers, most likely to communications look out on Monday, because it's surprisingly painful thing to get collected and then on next Tuesday we have the monthly community call where the kur true everything was. This happens within the past month.

A

So whatever the videos were released for at AMC and articles released and we'll have some nice live demos over that net there as well, more simplified demos, not a deep dive, 500 level, demos and they're looking into what are they kind of showed that in-depth call still during this week? But if there's no kind of urgent questions or feedback I think we can close up to call a few minutes earlier. That's absolutely fine, and if you have any issues, please use the github issue list.

A

If you have questions we would prefer maybe to use to take a community for those, but especially if there's box or issues or you can't get the engine to work. Just let us know there was a one thousand tenants using the BMP provisioning engine last month and in general, core and PowerShell is used in in more than almost 4,000 times in a month. So I think we have a momentum, just a lot of usage, which is great and it's great to get to input from the community as well. So thank you, everybody for joining.

A

Thank you for contributing again on this month and let's keep up with a let's. We were together for making things happen, so thank you. Everybody Cheers, bye,.