I have this workgroup and I would like to make lists of interesting projects/repos contained in different subgroups. This would make it easier for us to find relevant projects.
I know you can share a project with a group if you are the owner of that project but can I, as maintainer of a group, add links to public projects without the owner of that project having to do anything?
Another option for me could be to just do a readme file with links but then I could get dead links and I would miss the project description and most importantly, when they last was updated.
This does not 100% qualifies as an answer, as i see it as a hack or misuse of a feature of GitLab. But i thought the idea is worth sharing. Additionally it will only work with projects which you can actively change - else this would need interaction of project maintainer.
GitLab supports optional topics on projects, they can be maintained by maintainers within the general settings of a project.
If you add a topic interesting. they get searchable by this tag like https://gitlab.com/explore/projects?tag=interesting
This way, you can create and maintain a list, which is always up to date and show the most recent description, name, etc. - also in connection with permissions (you will not see something you are not suppose to see).
The downside is, that this topic might not be suitable to be (ab-)used within your group, and it might add more confusion for others than it should, because you will see this topic in the project overview page.
Related
I'm trying to create a documentation Wiki for our internal software on Gitlab.
The starting point was a messy set of texts, partly from some LaTeX script but also other sources.
My biggest issue at this point is to ensure the correctness of all cross-references/links within the wiki.
I'm steadily cleaning up page by page, but it seems impossible to fix all the broken links. Especially, if some pages are renamed, or moved to another folder, every reference to this page is broken. It is also annoying to verify each link manually, i.e. by clicking on it just to see whether it links to some destination or not.
So, my question is: Are there any tools for checking the integrity of a Gitlab wiki?
I'm thinking of something like this: If I clone the entire Wiki into my machine (via git), I have a folder structure, populated with markdown files. Some tools should run recursively through all these files/folders and at least give me a list of broken links.
Of course, some re-factoring tools that would update links in other pages if I move/rename a page would be even more awesome.
Consider whether markdown-link-check might work for you. FOSDEM has a good introductory article.
I was not able to use markdown-link-check for an Azure DevOps Wiki (cloned git repo.), presumably due to the need for an implicit '.md' added to the links. Although, I did not put much effort into the attempt.
I could - and may eventually - fork the source and adapt it to Azure DevOps Wiki. You might find that it works out-of-the-box for GitLab's Wiki or, you may be able to adapt it.
Problem
I have a gitlab with a lot of old repositories. I want to mention my gitlab as a reference on my CV but I do not want all the old repositories to appear there, just the more relevant ones.
Just making the projects private is not enough as this leaves a lot of clutter in my dashboard and it is hard to see the projects I am trying to showcase.
I do not want to delete the old projects, as I want access to them in the future, I just want to hide them from other people to see that they even exist.
What I Tried
I tried archiving the old projects but they still appear on my Projects lists, just with an archived tag.
I saw mentions of playing with the "Metrics Dashboard" under the visibility settings but this is greyed out for me + I do not think this is what I need from my understanding.
Required Result
For me to be able to choose which projects appear and do not appear in my gitlab dashboard.
Thanks in advance for any help available!
EDIT
I found out that I can star and un-star projects, and that will count as activity on the project without actually changing anything. As the dashboard displays projects by when there was last activity on them then you can actually arrange your project by staring and un-staring the projects in the reverse order you want them to appear.
This somewhat does what I want, but with an ugly work around. Also it will always display 10 projects as far as I can tell, so if I want to only showcase 6 of them the best I can do is push the 4 I don't want to the bottom, but I still can't hide them completely.
This is why I am not writing this as an answer to my question. There has to be a way to just tell a project to be hidden or arrange the projects without this ugly workaround, and if there truly by design isn't a way of doing this then it will also just be good to be officially told that.
GitLab Groups do exactly what you need. See here for more info.
You can assign each project to a different group
You can move a project from one group to another as needed
You can assign different permissions and visibility for each group
You can also create subgroups
Each group/subgroup is treated as a separate namespace, meaning you access it using a different URL
So you can define a public group called yourfullname and a public subgroup called portfolio. Move the projects you want prospective employers to view to the portfolio subgroup and make sure their visibility is also public. All other groups/subgroups should be private. Then people can access your projects by visiting the following URL:
gitlab.com/yourfullname/portfolio
You can still view all of your projects in a single dashboard if you want, or you can view all projects within a group or sub-group by navigating to the desired group URL or dashboard. In the image below, archive and development are private (see the lock icon), but portfolio is public:
Currently I have an GitHub repo that I use for collaboration. I want anyone to be able to join it.
GitHub currently requires users to first find me (there is no form to request) and ask me and then they are mailed an invitation which they then have to accept.
I'm guessing there is an app out there for this but I can't find it.
I'm looking for either an integration that takes a turns a issue comment into a team add, or form the user can request an invite from.
Forking a repo remains the official way to contribute without asking. Then the contributor can make a pull request back to the original repo.
The goal is to "manage" (through PR review) the flow of contribution.
The other alternative would be to add several people owner of an organization team: that way, you would be the only one people would have to ask in order to be collaborators.
If this is an organization that you're trying to add members to, there is already some automation around that.
JazzBand allows anyone to join the organization. Their website uses the same mechanisms as add-to-org to add people to an organization.
Looking at their source code, it appears both use the GitHub API to add members to an organization.
PUT /orgs/:org/memberships/:username
That said, if this is a personal repository, you'll instead want to follow the API to add a collaborator
PUT /repos/:owner/:repo/collaborators/:username
It's likely you could modify either of those projects to fit this need. Cheers!
So, maybe I'm a bit old-school, but when we created websites in the past, we'd develop the site on a development server, then publish or promote the pages and files to the production server. This has always seemed to be a good way to go so that users didn't see messed up pages or (God forbid) a downed server because one of us screwed up.
But it doesn't seem that Microsoft had this idea in mind when they created SharePoint...at least, I haven't been able to find a way to do this in the infrastructure as it's defined.
Does anyone know if there's a management strategy for SharePoint development? I've read online that we can make a backup of the development environment and restore to the production server. That might work the first time, but any updates to the production server can't do that without risking data loss on the production server. I've seen some tools out there for migrating list contents, pages and documents from one server to another--although, admittedly, I've not yet investigated them.
But, another concern of mine is custom content types. It seems that once a list is using a content type, you can't update it without deleting the items from the list, disassociating the content type, and reassociating the content type. Shouldn't there be some way to UPGRADE a content type?
Anyway, if you have any suggestions for any of these current dilemas, I would LOVE to hear from you.
Thanks in advance,
Dan
Thank you for your quick reply.
We already have several features created for our site and a solution package bundling features directed at the fundamentals (content types, columns, etc), and another solution for features having to do with branding (page layouts, master pages, etc.)
But it seems like this is a one-time-shot...basically, it gets our server set up, right? Once people have started using the production environment, we're going to have documents, pages, list items all existing in our content database, and it'll be impossible to update things like content types, columns.
Features you have to deactivate and uninstall before you can install and activate the new feature, right? I've seen a Version property on the feature definition, but as near as I can tell, this doesn't do anything. Solutions seem like they can be upgrade by incrementing the version number, but it doesn't seem to modify things like content types and columns--especially if they're in use. Plus, I'm not sure how extensive the upgrade with solutions is.
There's precious-little documentation out there for this sort of thing. It seems like everything I'm reading is how to get your SharePoint server set up initially...not managing it long term.
Do you have any advice or suggestions?
Thank you all for your suggestions.
But we've been working on this site for over a year now. I'm pretty confident that we're already setup according to what most of you are recommending. We already have several features that install things like content types, columns, master pages, page layouts, and workflows. Most of these features are contained within solution packages. We have all of our development environments set up as VPC servers.
So, I have the initial deployment pretty much set. What I'm REALLY hoping to find out is how I can upgrade things like content types and columns and stuff down the road. Is it possible to change content types once they're in use? Because it doesn't seem, based on my initial testing, that this is possible. I'm not to worried about the assemblies because it looks like they swap out just fine, but the only way I've gotten a content type updated is by deleting any items referencing them (i.e. all the pages in my pages library), removing the content type, then re-adding it.
Do any of you know if there's a way to update a content type AFTER the initial deployment? ...when users have already created items based on the content types we've already deployed?
(The other part of my question was actually moving existing pages from the development server to production, but I can live without that. My major worry is the content types.)
The best way to go is developing with features. Once the features are done, you ca deploy them with Solution package (called WSP).
The only thing left to do is to reactivate those features. That way, you can progressively roll-out new features without having to do everything in production.
WSPBuilder is an application that helps you build WSP.
For automating all of this... good luck. There is a lot of work involved.
UPDATE:
Deploying Content Types and Columns are tricky. Once the website has been created, you can't update them anymore through features. You need to go through the code and recursively go through all the sites and modify the specific content type that match the name.
We've tried and it's not possible to do that normally with features. This need to go through something I call "deploying with code".
You really really need to define your content types using a feature because that way each content type will have a set GUID and will be stored in the database using the same name. This becomes important when running CAML queries over the site and there are a few other little gotchas when content types are created "will nilly" if you will.
I prefer STSDev for rolling out solutions using custom content types.
There are two ways to edit pages on the server. You can define the page library to have major and minor versions. This allows editors to edit the page and a defined publisher to publish them. This is good on an internal site, but is not recommended for a public facing site.
For a public facing site you will need to use Content Deployment
I cannot stress enough that before going ahead with a production release you make sure you have features for the content types.
As mentioned here, Chris O'Brian has a post saying that you should not use features unless necessary. One of his reasons is that it slows developement.
I disagree with this. Developement is slower if you are unfamiliar with features, but once a level of knowledge is reached, it is not a major factor.
Do listen to him about the backup and restore method of moving the content.
If you do that, all mess in the content types and fields and webs you may have created during developement (for me that is always quite a bit) will be moved to your production site.
Instead of having a nice clean site where everything is consistent, you will end up with little bugs and some areas of the site behaving differently to others simply because of old development cruft.
I recommend taking a look at Chris O'Briens most recent post, and his great Content Deployment Wizard: it's not all about Features!
Maxim is right in that most items should be deployed via features that are wrapped in solutions (WSP files). Your strategy should be to make sure your solutions and assemblies are broken into related bits of functionality. This is also beneficial in that features can be isolated at certain levels like sites and webs. Feature activation code, deactivation code and feature stapling should be used when updating any content updates. Content deployment can also make sense.
Once thing to remember is that if the updates are only in code then the assemblies can be updated without requiring the feature to be reactivated or the solution retracted and redeployed. All that is required is the Application Pool to be reset.
Microsoft has a couple articles on Dev environments and you can Google many others who recommend environments. We do development on virtual machines and deploy most items to an virtual integration server. Once we smoke test it we then deploy our solutions to QA so on and so forth. The benefit i sthat features and solutions are easy to retract. Once it goes out to production it should be thouroughly tested.
Developing in SharePoint has it's issues, that goes without saying, but so far I have found that the benefits outweight the problems.
Team-Based Development in Microsoft Office SharePoint Server 2007
We developed a custom solution which would update content types and fields for a Site Collection. Underneath the covers, through code, SharePoint allows us to modify the Fields as well as values in the Fields and Site/List Content types.
For moving the actual content from QA to Prod we use Echo
What is the general feeling amongst developers regarding the changing of files in the 12 hive.
For example if you were asked to remove the sign is a different user menu item, you would need to modify the relevent user control on the filesystem. Now if you just go and modify it via notepad or copy over and then if you go and bring a new server into the farm you will need to remember to do the same on the new server.
Obvouisly you could deploy the changed file as a solution and have that done automatically, but I'm just wondering if people are hesitant to make changes to the default installed files?
I have done a bit of SharePoint development, and I must tell you that messing with the 12-hive is a ticket to a world of pain if you ever want to move the app.
I'd rather hack up some javascript to hide it, at least that can be bound to the master page, which is much more portable.
And remember, you never know when the next service pack comes around and nukes your changes :)
I agree with Lars. Sometimes you will not be able to avoid it, depending on your needs. But, in general the best policy is to avoid modification if at all possible.
I know that some of the other menu items in the current user menu (change login, my settings, etc) can be changed by removing permissions from the user. Under Users and Groups there is an option for permissions. I can't remember the exact setting (develop at work, not at home), but there are reasonable descriptions next to each of the 30+ permissions. Remove it and you start hiding menu options. No modifications to the 12-hive needed.
There is a very simple rule: if you want to keep official support from Microsoft, don't change any of the files in the 12 hive that are installed by SharePoint.
I've never encountered a situation where the only solution was to change such a file. For example if you want to change an out-of-the-box user control of SharePoint, you can do so by making use of the DelegateControl, and overriding it in a feature.
More info:
http://msdn.microsoft.com/en-us/library/ms463169.aspx
http://www.devx.com/enterprise/Article/36628
I know it's tempting to quickly change a file, and I have to admit sometimes I just do that on a DEV box, but don't go there on a production server!
Not sure if there is much use pitching in, as everyone else pretty much has it covered, but I would also say don't do it. As tempting as it is, its just impossible to know the full impact of that little change you have made.
From a support perspective you will make it difficult for Microsoft support (patches/hotfixes).
From a maintenance perspective you are also opening yourself up to long term costs.
Go the javascript route.
The way to go about it is to use a Sharepoint Solution (WSP) file.
To change the user control, create a new Sharepoint feature with the new functionality.
Include this feature in your solution.
Deploy the solution either using the stsadm command line, or through Central Site Admin.
This will then get automatically deployed to all the servers in your farm, and it avoids you overwriting anything default sharepoint files.
For more info, check out Sharepoint Nuts and Bolts blog on http://www.sharepointnutsandbolts.com/ which give an introduction to WSP and Sharepoint Features.
I've done this many times and I will speak from experience: Never ever touch the onet.xml files within the 12 hive under any circumstance. Any error that you make in there, and to make the CAML even more complex the file is largely whitespace sensitive, will have an impact on every part of SharePoint.
You should also consider that aside from the substantial risk to the installation, you may well be building in dependencies upon your changes that are then over-written in a future patch or service pack.
Most of the time, you can accomplish everything you want to using features and solution packages without modifying the files. However, there are a few (rather annoying) rare cases where your only option would be to modify a file on the system. I have used it for two particular cases so far. One was to add the PDF iFilter to the docicon.xml file, and the other was to add a theme to the themes.xml file. In both cases, it seemed to be the only way to achieve the goal. Still, we used a solution package to write those files out to all the servers in the farm.