spring-attic / sagan Goto Github PK

Per http://invis.io/JKHLAVZE.

It's useful to be able to point users to a subset of the guides and tutorials based on a search one has prepared already.

So it would be perfect to allow something like https://spring.io/guides?filter=web produce the same output as entering web into the filter box on the page.

• ready to implement Dararith's design laid out in #134 (comment)
• make sure to sort results by publication date

As requested by @phumphrey on behalf of a user tweeting to @SpringCentral, it would be best if the date of a blog post shows up in its search results, e.g.

@trisberg wrote:

Seems something is missing after the springsource.org Drupal conversion - this link http://www.springsource.org/node/22687 gets a 404 and is used in several places like the sticky post on the forum.

Should that redirect to the new blog site https://spring.io/blog? I don't see it there either - should have been posted around Sep 6th this year.

e.g., from http://spring.io/projects:

This has been reported confusing by users on several occasions.

As reported by @olivergierke, @philwebb, @phumphrey and others.

guide.js forms (what appears to be) an incorrect html id here: the index is placed outside the double quotes. This:

var button = $('<button class="copy-button snippet" id="copy-button-"' + index + ' data-clipboard-target="' // ...

Should probably be this (moved the double quote after copy-button- after index):

var button = $('<button class="copy-button snippet" id="copy-button-' + index + '" data-clipboard-target="' // ...

Trevor, I know we’ve tried this a few times, but I’d really like to get it worked out. I’ve followed the instructions at https://confluence.atlassian.com/display/BAMBOO/Integrating+Bamboo+with+HipChat for a new build plan I’m working on for #150, and no luck.

Can you look into this? Thanks.

@trevormarshall, per comments in HipChat, we’re seeing unexplained transient build failures in the newly-created Sagan CI build plan. The workaround is to simply manually run the build again, which tends to work as expected.

Any ideas? I won’t have much time to look at this before I take off, and this is going to be a real nuisance to folks like Greg and Dave (and anyone paying attention to HipChat), since this build plan is now the blessed way to get changes into production.

/cc @gregturn

Bundle all .js sources into a single file for efficient resource management.

Blocked until #110 is complete. We should also wait to start this issue until we've finalized exactly what we want to do with @rstoyanchev and @bclozel on resource transformations in the core framework.

• Resolve cram.js issues regarding thymeleaf templates and use case.

Could someone check to see why this commit:

171d0a9

isn't yet active here: http://projects.spring.io/spring-hadoop/

• @nebhale to update re getting CLA integration online by March 24th / 25th (3/15)
• @dussab to update re DB provider per #121 (comment) (2/4)
• @dussab to update with status / next steps, per #121 (comment) (2/3)
• @dussab and @cbeams to discuss resourcing
• When this work is complete, update CONTRIBUTING.md to reflect the new process.

@cbeams wrote to @nebhale:

Brian told me a while back that you'd been working on an equivalent to CLAHub for us?

I'd really like to be up and running against that when we open source Sagan. Should have thought to ask you about it last week.

Let me know and we can set something up to discuss.

@nebhale responded:

Defintely, let's do it. I've had about zero input from anyone outside of me so I'm interested in having a second set of eyes looking at it. It needs to be re-written as a boot application to stay on message, and isn't a fancy single-page app, but code is here, and a running instance is here. At the moment, it only takes care of the signing side of the work-flow, and doesn't have GH integration yet (everything stopped for S1). Ping me whenever you've got 30 minutes and we can chat about where to take it from here.

I know, we left those links out when doing the first release, but people are asking for it. So it would be great if we could find a way to get the links to md5 and sha checksums back on the STS/GGTS download page.

I think the best place would be a very small link in the download context menu (the one that appears when you click on "Based on Eclipse 4.3.1", for example). At the moment it gives you four options. Each of those options should get an md5 and a sha checksum download link right below the "dmg 379MB" entry (for example).

The link to "#initial" is always "Create a resource representation class", as it is hard-coded in sections.md.

This should either be worded more generically or there should be some mechanism for the tool to look ahead and whatever header that #initial is sitting on and use that as the label for the link.

On Oct 3, 2013, at 1:33 AM, @trevormarshall wrote:

I would like to get a playbook together to go over when dealing with outages. Basically a line by line guide ordering items such as fetching the logs, notifying hipchat room, restarting, unbinding / binding db, escalating, things of this nature.

@briancavalier wrote in #132

In addition to taking a different approach to sending the searchable metadata, we should also consider changing the client's filtering strategy. Currently, it just wraps the typed string in a RegExp. So, for example. typing two words that both appear title or subtitle but that aren't adjacent to one another will yield zero matches (e.g. try filtering or "learn RabbitMQ"). This seems unintuitive to me.

For me it seems to be the old logo.
http://docs.spring.io/spring-boot/docs/0.5.0.M5/api/

• @trevormarshall to set up varnish in front of schema.spring.io
• @trevormarshall to provide springsource.com rewrite rules (done, see here)
• @bclozel to port springsource.com rewrite rules to rewrite app
• @trevormarshall to provide springsource.org rewrite rules
• @bclozel to port springframework.org rewrite rules to rewrite app
• @bclozel to deploy redirect app and smoke test with local /etc/hosts mappings
• @trevormarshall and @bclozel to set up page rules at cloud flare for www.springframework.org/schema* and /dtd*, per trial run with foo. as below:
  
• @bclozel and @trevormarshall to switch DNS for (www.)springframework.(org|com) to CNAME for redirect.spring.io
• double check that our Pingdom alerts for schema URLs monitor everything we need, especially the legacy URLs that are in most xml files. These are of the form: http://www.springframework.org/schema/…
• @bclozel to check with @trevormarshall that the old cluster is not serving traffic anymore

When all of the above are complete, @trevormarshall should be OK to shut down the old Apache cluster.

OSX + Firefox

Click "forum"
Hit the back button

The forum text disappears

Repro:

1. Go to http://spring.io/docs
2. Reduce browser width until nav goes responsive
3. Click on "Spring Framework"
4. Attempt to click on any "Reference" or "API" link
5. Notice that the mouse cursor never changes into link-clicking mode
6. Notice that if one does click, the pop-up window simply disappears
7. Now stretch the window back out
8. Repeat steps 3 and 4
9. Notice that links work as expected.

Currently, all Getting Started Guides (GSGs) assume an implementation path based on Gradle. Better would be to afford the user the option to use Maven, Gradle or STS (possibly IDEA as well).

The Maven POMs are already in place, but there are implementation challenges in how to present the option to the user. The approach needs to be discussed, but the feature is definitely desired and has been requested several times.

body text.

second line.

The UI for search is really nice but can be a little confusing when you first use it. It is easy to miss the flashing cursor and it is not obvious that the popup is a text entry box.

It would help if:

• There were some visual clues that this is an edit box (perhaps just a simple '>' at the start to guide you, or maybe just a classic text box).
• When you clicked on the search box, the grey text disappeared, this would make it a little more obvious that you can replace the text.

Originally requested by @dsyer. When searching spring.io for ApplicationContext, it would be ideal if a link to the actual source for ApplicationContext were presented to the user, e.g.:

https://github.com/spring-projects/spring-framework/blob/master/spring-context/src/main/java/org/springframework/context/ApplicationContext.java

@jacksoncvm @briancavalier @unscriptable @gregturn

As you may have noticed based on GitHub issues notifications coming your way, I'm currently in the process of evaluating a pure GH-based issue management workflow, i.e. taking Tracker out of the picture.

I'm a big Tracker fan, and it was invaluable in getting the site over the finish line, but there are several issues weighing against its continued use in the management of this project:

• It's open source on GitHub, meaning the most natural way to manage issues from end users (internal but particularly external) is via GH Issues
• Tracker is simply too disconnected from the project by comparison. This has led to a process where duplicate issues get created, one for the original story added in Tracker, one for the PR that the assignee eventually submits. (Note that I am aware of deeper integrations between GitHub Issues and Tracker's API, but this requires us to set up a server to help manage this and it's just generally more moving parts I'd prefer not to deal with.)
• User management in Tracker is a problem; first, getting users to know where it is, second, getting them rights to add issues, etc if they need to. With GH it's much simpler, can work with our existing GH Team management, etc.
• Tracker's otherwise useful story estimation system is basically meaningless without a dedicated regular 8-hour-a-day-team
• Which brings me to the most important point: for good and practical reasons, we are not managing this project in typical Pivotal fashion. You all have a number of responsibilities, and it's not reasonable for me to hold daily stand-ups and weekly IPMs and retros with each of you, nor to expect 8 hour work days from everyone all the time. Furthermore, the shape of the team changes over time, and will continue to—this is a good thing!

For these reasons, I think we need to take a different approach:

First, let's embrace GH Issues completely. It's the lowest common denominator anyway, i.e. we need to support it for our external users. Let's make it the one and only issue tracking system.

Second, make use of a service like waffle.io or HuBoard to provide higher-level management and prioritization of issues. I've used both, and currently prefer waffle for a number of reasons I won't go into just now for brevity, but am happy to discuss. Check out the dashboard here: https://waffle.io/spring-io/sagan. We can meet and discuss how it all works later, but feel free to explore for now.

So: those @-mentioned above, please weigh in with your thoughts, questions, +1's etc, since you're the ones actively collaborating on the site right now. I'll make an announcement to the dev list once that discussion is complete and things are switched over.

/cc @phumphrey, @dararith

In 2 places, main.css uses -wekit instead of -webkit.

• see if we should use spring-ws or marketo's jar API
• get credentials (username, secret key, SOAP endpoint)
• get specific SOAP args for spring.io when creating a lead (lead source, munchkin ids...)
• integrate HTML+CSS for newsletter form and remove iframe code
• create a SOAP client
• create a REST endpoint

• @jacksoncvm to respond to #117 (comment)
• @bclozel to update the issue with new TODO items (marketo's API)

Our current HSTS configuration forces https everywhere on the site, resulting in internal users copying and pasting https urls instead of http. This is problematic because there are https-related bugs in a couple areas on the site, and external users shouldn't be exposed to these.

See original story at https://www.pivotaltracker.com/story/show/56568046 for complete background.

Bottom line, however, is that our current font families (Google fonts) do not render properly in Chrome on Windows, and generally do not render in-browser as well as we would have liked with in comparison to Photoshop.

• @dararith to choose new fonts as time permits
• @jacksoncvm to implement

http://spring.io/guides/gs/not-a-guide should be a 404 I would say (but currently it's a 500).

https://github.com/springframework-meta/springframework.org/blob/master/src/main/java/org/springframework/site/web/configuration/DefaultViewControllerConfiguration.java#L71

Should now be supported with embedded tomcat

Some apps are currently under spring-operator, others (localhost) under Dave's account.

spring-logo and spring-icon.zip were inadvertently added to git history early in the project. This is unfortunate because of their large size (~17MB) and means that cloning the repository takes much longer than it should.

Proposed is that, prior to open-sourcing the repo, we rewrite history in order to remove these files completely.

https://twitter.com/rmannibucau/status/382566464580255744

@briancavalier wrote:

Start with

• results.html and
• search.js

This is really the simplest possible place to start. It is a huge problem right now to trace dependencies between html, css, and JS on every page. Starting here will let us get a bit of engineering infrastructure in place, that will hopefully make it easier to apply DI to other pages. search.js is also the only bit of JS that is possible to unit test currently.

Convert the following, existing js files to CommonJS modules. Use curl as a loader.

• guide.js
• team.js
• ...

Currently the GitHub Application name (used for admin authentication) is springframework.org (production). This should be changed to something more accurate (maybe spring.io Administration) and drop the (production) label.

I'm not sure if this related to the migration or some other aspect, but it seems that some migrated blogs have missing images

http://spring.io/blog/2012/10/15/spring-data-release-train-arrived

descr

Trying to update the project page for spring-hadoop. Made a small change and got this email from GitHub:

The page build failed with the following error:
page build failed

Here is the change: spring-attic/spring-hadoop@9d831ca

Dave Syer seems to think that there is a conflict between the first two commits for this projects gh-pages setup. Could someone take a look at this and figure it out. I haven't been able to get my Ruby environment up and running quite yet so I can't trouble shoot this ATM.

• commit comment formatting
• resolving issues by #issue mentions
• attaching pull requests to existing issues with hub pull-request -i
• branch naming practices
• code formatting
• import optimization
• cover guidelines for asking questions via GH issues, e.g. what topics are suitable for questions (basically everything under http://spring.io and http://redirect.spring.io, and what is not (guide content, forum issues, etc).

Our current "talk to us on Twitter" approach is fine for a start, but some blog authors have already weighed in saying that this is not enough. Explore linking posts to dedicated comment threads in the forum, and reconsider the possibility of Disqus-based comments.

The forum link on http://projects.spring.io/spring-security/ points to http://forum.spring.io/spring-projects/security, but it should be http://forum.spring.io/forum/spring-projects/security

When editing an already published blog post it appears the save + preview unpublishes the existing post. It would be nice if the site would allow preview without unpublish.

• filter.js
• team.js
• guide.js
• …

Once these tests have been written we'll be dependent on #113

• Selenium server 2.35.0
• windows 7
• windows 8
• all versions of IE >= 9
• linux (your choice of distro - Ubuntu latest would be fine)
• OS X (latest is fine)
• Safari 5
• Safari 6

• Refactor codebase to use Lombok to improve readability and maintainability of the site.
• Run delombok against master via CI and commit changes to to “delombok” branch, in order to provide
  • a comparison to master, demonstrating the benefit (i.e. removal of boilerplate, lines of code saved)
  • a simple way of understanding exactly what any given Lombok annotation is doing.
  • an escape hatch for users who want to import the code into their IDE but do not wish to or cannot for some reason install the Lombok IDE plugins.
• Update wiki documentation to instruct users to install IDEA / Eclipse Lombok plugins or to use the delombok branch

You might want to consider BDD mockito syntax:

given(...).will(...)

rather than

when(...).then...

as we use that style for Spring Framework tests.