Introducing “daisy” a sass/pattern-lab workflow solution using Vagrant

daisy-vagrant

I have been a big fan of Brad Frost and Dave Olsen‘s project, Pattern Lab.

Anyone that knows me heard me say “Embrace the fluidity of the web!”. Atomic Web Design is a concept that I always circle back to whenever I’m talking about web. I am also a big fan of sass, and also of Zurb Foundation.

We were able to use a lot of those ideas on a hbr.org’s new responsive site, which launched November 10. My colleagues, Kevin Newman, Fred Lalande, Matt Wagner, Kevin Davis and I just did a presentation on SASS and Pattern-Lab team workflow using Vagrant/Puppet at our Boston CSS meetup, Cascade Bos. Topic was really wide, we talked about anything from Object-Oriented CSS to Web-based Style Guide to deployment method to… everything. You can go through our slide deck here.

But for the “workflow” part, here is the gist of it.

HBR’s pattern-lab work flow

  • Instead of installing all our dependencies — such as php and apache for pattern lab, ruby and compass for css preprocessor and node for grunt and bower — into each of our individual computers, we have a Vagrant/VirtualBox virtual machine that are committed in our github repo.
  • What is really nice about Vagrant is its pass-through file system. Meaning you can edit the file on your OS (be it Mac, Windows, or Linux) using your favorite text editor locally (without going through virtual window), and it can be accessed by both host and virtual machine.
  • So we edit our www/source/scss folder, and www/source/_pattern folder right on our machines. Then, with power of grunt, you can view your changes appear on your browser, served right from your VM.
  • After we like what we see on our local machine, we make a pull request (we commit the whole thing, Vagrant and everything) in github. We have a internal agreement that you don’t merge your own request… so it gets reviewed by peers and then, eventually, gets merged to the master branch.
  • Then our hardest working member of the team, Mr. Jenkins, who watches the changes on the master branch, sees that it changed. And he pulls the latest, compile sass, compile pattern lab, and then he puts the public folder part of the pattern lab onto our internally available web server (is “intranet” still a word?). This is our living style guide. Our backend developers uses it to grab markup code out of it. Designers and editors, as well as marketing people, can access it any time they want.
  • Only part that goes directly out of our pattern lab to production is compiled CSS (it gets minified on its way to our production). But CSS HAS TO BE EDITED IN pattern lab, which forces everyone to first create a new pattern inside our pattern lab. This also ensures that our style guide is most up to date. Win-win!

That’s basically our workflow. I understand that pattern lab is not the silver bullet that works for everyone, but I am pretty proud of how it turned out for us. Another nice thing about committing your entire dev environment in the repo like this, is, say, if you spilled a full glass of red wine on your MacBook Pro by accident (Oops!). All you have to do next day is to get a loaner from IT and download your text editor and you are right back at it. If I had to install node on my machine again ON A DEADLINE, I think I am going to cry. But anyhow…

One of the thing that I wanted to do was release public version of our pattern lab github repo. One without styles in it, and one that is a stand alone virtual machine which runs Pattern Lab.

And, lo and behold, (Thanks to Matt Wagner and Kevin Davis) we were able to release it. It is available for you to download.

Introducing daisy.

Check it out. Star it, fork it. Comment it. File issues on it. Even make pull request on it!

I would absolutely love for you to try it out and let us know what you think.

Continue reading…


Hackathon Memo: How to put Foundation SASS site working with Ruby on Rails backend

Ruby on Rails is pretty cool. Pretty magical. If you need a site up for the duration of weekend hackathon, it is an amazing tool.

I had a pleasure of joining a team over the weekend on AthenaHealth’s More Disruption Please and MIT Hacking Medicine’s well-organized and well-funded hackathon. (Unfortunately my contributed github code is in a private repo… a bummer. I should have told them that open repo would be preferred… that’s the thing about some people’s approach of hackathon. But that is beside the point) Though I was only there for only the part of Saturday, I got a bit done and learned (and re-learned) a lot of things that I wanted to do, working with Ruby on Rails developers. I love working with smart people.
Here is my runnig notes. Probably doesn’t make sense, but I am hoping to revisit and edit later to make some sense, eventually make it into how to put Foundation SASS site working with Ruby on Rails backend
Continue reading…


Flexbox justify-content and Internet Explorer

A CSS layout challenge from a colleague: We have two columns, longer one on left, couple of items on the left. But the designer wants right side contents to align bottom with the left one.

Restrictions/assumptions:

  • Cross browser responsive website. Modern browsers, plus graceful fall back (“can read content”) to IE 8.
  • Right side contents, while most likely be shorter than left, can be any height. But must show the contents/not hide.
  • Do it with CSS, and (try) not use JS
  • As this will be part of template, make it as flexible/extensible as possible.

“Sure, you can do that” I said, “You can use the flexbox.. Every modern browser supports that. Even IE!

Or does it?

Here is my code pen:

See the Pen Flexbox justify-content by Daigo Fujiwara (@daigofuji) on CodePen.

What I found was not pretty. But here are my findings.

Mac Chrome (Version 37.0.2062.94)
mac-chrome

Mac Firefox (Version 31.0)
mac-firefox

Mac Safari (Version 7.0.6)
mac-safari

Windows Chrome (Version 37.0.206.102)
pc-chrome

Windows Firefox (Version 31.0)
pc-firefox

and finally,

Windows IE11 (Version 11.0.9600.17239)
pc-ie-11

Ugh.

What happened there? After playing around with IE’s inspector, it seems the issue is with “min-height” rather than the flexbox itself.

So, if you give, say, a height: 1px; in the .flex-container-column, it works (then the height will be min-height). However, height of 100%, or height of auto will not. Mmmm.

So, I tried old trick featured on Chris Coyers site (Love CSS-Tricks!), but since it defaults to auto, it still doesn’t work.

Setting height would work, for example in the pen above, if you comment out line 14 in css (height: 250px;), flex would behave as expected. But since now the height is set, and not flexible, say if the right-one div had a height of 300px, right-two will disappear. We can’t have that.

Continuing the investigation


CSS Framework…

Excellent presentation/research by Harry Roberts, one of the css guy that I follow, called What Is A CSS Framework Anyway?.
My favorite quote:

What is wrong with css? [It’s] “an old, loose, leaky, globally-operating, inheritance-based language which is entirely dependent on source-order, except when you introduce its own worst feature: Specificity.”

That quote reminded me of a presentation at Boston CSS meetup, given by Bocoup’s Greg Smith: “CSS is Awful”.

Speakerdeck: What Is A CSS Framework Anyway? by Harry Roberts Published May 10, 2014 at Industry Conf

Personally, I am a big fan/advocate/apologist for Zurb Foundation. I like it because, while it is true that the code gets bloated with some unused stuff/overrides and it is “trying to be all things for all men”, they actually do a good job of keeping it simple (relatively speaking), have great/flexible “settings” file, and best of all (this is true to a lot of open sourced projects) having a community of issues, discussion/solutions and a solid documentation. Since we have a small team, but everyone is hands-on, I think it is important that we can have some place where each developers can just look up something to see if this is some issue that people are having or not to trouble shoot. Anyway, great presentation and was inspiring. I love CSS!

But Harry has a good point, why do people geeks wear jQuery t-shirts but almost never zurb foundation t-shirts? I’d wear it :)


Updating old php scripts and learning PDO (PHP Data Object).

I started making websites in 1998. Gosh, in the internet years, that’s a forever ago. Remember Perl CGI and its cgi-bin folders? I have also written a lot of website using php along the way, and LAMP stack was sort of my thing for a while. But since I started to focus every efforts on front-end stuff, it have been a while since I looked at simple backends. Some of the sites that I created long ago desparaetly needs some updates. I have, in the past, thought or looked into converting sites to ruby on rails at one point, then looked into python, then thought hard about switching to some kind of static site generator engines (like jekyll or assemble.io), and finally more recently, thought about converting them to MEAN stack. But, really, who has time for that? Lately, I am thinking about just upping my php level and “patcing” it right in php5. Or at least keep the database structure and convert the backends to be more like API or services, still using php.

This weekend, I have been sort of playing with one of my old project: Softball Stats, written by David Carlo. It is an awesome little project.

“SoftballStats is a collection of PHP4 scripts and a MySQL database that track the stats, games, players, and every play made in multiple softball/baseball seasons. SoftballStats also compiles statistics for each player.”

I could use this on my work co-ed softball team, or also use it for my son’s summer league. I started updating it on my github.

Found this excellent website: phpbestpractices.org/

Here ares some notes that I kept while updating some of scripts.

First thing is the way I have been connecting to the database.

I was doing this:

$link = mysql_connect("example.com", "user", "password");
mysql_select_db("database");
$result = mysql_query("SELECT * FROM player WHERE SeasonID = 2014");
$row = mysql_fetch_assoc($result);
echo $row['name'];

New (Well, new to me.) way to do this is using PHP Data Object

$pdo = new PDO('mysql:host=example.com;dbname=database', 'user', 'password');
$result = $pdo->query("SELECT * FROM player WHERE SeasonID = 2014");
$row = $result->fetch(PDO::FETCH_ASSOC);
echo $row['name'];

Also, way I was doing mysql_num_rows($result) will now have to changed to
$result->fetchColumn();

As with mysql_data_seek($result),

Simple enough? I was cruising along, then I encounter the error. Or more like, NOTHING. By default, error is a blank screen, as it turns out.

My php.ini was in /Applications/MAMP/bin/php/php5.5.10/conf/php.ini (Note that it is not MAMP/conf as you expect, it’s in /bin. you can find this infor using . and display_errors was in line 277.

Actually, you need to put those two lines in the php file:
error_reporting(E_ALL);
ini_set('display_errors', '1');

Got some things fixed, but broke more things. Ugh. Still working on it. Ah, time. I wish I had more.
Continue reading…


Conflict resolution on git

cant-merge

Git is just amazing. Learning command line git is one of the best thing that has happen to me. I’ve been lucky to work with some talented guys from bocoup, and I’ve gotten to know quite a bit about it lately and just wanted to document things that I do when I encounter that dreaded (well, not THAT bad) “We can’t automatically merge this pull request.” message.

Also if you have not seen this Cheat Sheet do yourself a favor and take a look.

So, pull request can’t be merged. It’s because what you worked on (likely your branch) have some conflicts with what you are trying to merge, likely origin/master. this stack overflow answer was useful in understanding what I am doing.

My instinct was to git pull. But that is a bad idea (Quick! git reset origin/your-working-branch --hard to get what you have pushed to github!). When you encounter “can’t merge” message, git’s help suggest this:

Step 1: From your project repository, bring in the changes and test.

git fetch origin
git checkout -b your-working-branch origin/your-working-branch
git merge master

fetch then merge is much better than pull. Pull does too much under the hood (as my co-worker Tim says “too much magic”). But still, chance is that you may end up with WAAAAY too many conflicts and therefore very hard to resolve them. Perhaps there is even better idea:

git fetch --all (You may not have to do –all, but I am paranoid and this is my feels safe blanket, if you know what I mean. It’ll ensure you have the latest remotes.)
then git rebase origin/master

Ah, rebase. I had sort of hard time understanding concept of rebase, but perhaps git’s message explained it to me the best. It says “First, rewinding head to replay your work on top of it…” All your commits are “replayed” after changes that are already in master. People used words like “flattening the history” when explaining rebase, which was confusing to me. but “replay your work on top” explanation worked best for me.

When you rebase, or “replay your work on top,” you resolve conflict one by one, as it encounters conflict.

You may get a message like this:

Applying: My commit 1.
Using index info to reconstruct a base tree...
M	source/js/file.js
M	source/js/folder/file.js
Falling back to patching base and 3-way merge...
Auto-merging source/js/folder/file.js
CONFLICT (content): Merge conflict in source/js/folder/file.js
Auto-merging source/js/file.js
CONFLICT (content): Merge conflict in source/js/file.js
Failed to merge in the changes.
Patch failed at 0002 My commit 1.

When you have resolved this problem, run "git rebase --continue".
If you prefer to skip this patch, run "git rebase --skip" instead.
To check out the original branch and stop rebasing, run "git rebase --abort".

Nice to have that –abort option. Unlike merge or pull, you can just forget about the whole thing. Continue till you are done.

But because rebase changes the history, you may get message like this.

Your branch and 'origin/your-working-branch' have diverged,
and have 95 and 2 different commits each, respectively.
 (use "git pull" to merge the remote branch into yours)

If you push force that message will go away
git push origin your-working-branch --force

$ git push origin your-working-branch -f
Counting objects: 51, done.
Delta compression using up to 8 threads.
Compressing objects: 100% (12/12), done.
Writing objects: 100% (13/13), 1.56 KiB | 0 bytes/s, done.
Total 13 (delta 8), reused 0 (delta 0)
To https://github.com/daigofuji/my-repo.git
 + 368da98...4266772 your-working-branch -> your-working-branch (forced update)

When you push force you’ll see the previous sha and the new one, so if you mess up and want to revert, simply reset hard on that previous sha and push force again
they look like previous_sha…new_sha

so in this case
git reset 368da98 --hard

In case of “Oh Crap, I made a spelling mistake and I just pushed that.”

You could git reset HEAD~1 -- soft to undo last commit and re do it, OR,
change the file, then git commit --amend
Either way, since you already “Pushed” you need to push –force to make it work.

If your working directory is dirty (meaning you have conflicts) and you don’t have any work you need to save:

# Cleans the working directory.
git reset HEAD –hard

# Reset the branch to the remote (in this case your-working-branch.
git reset origin/your-working-branch –hard

Continue reading…


My talk from css meetup: Vector-based Responsive Web Design

css-talk

I had a great time talking at April 22, 2014 Cascade Bos meetup. My talk was titled: Embedded web fonts, Icon fonts and SVG – Vector-based Responsive Web Design using CSS.

Coincidentally, the same day, Chris Coyer of CSS-Tricks have published a blog post titled: Inline SVG vs Icon Fonts [CAGEMATCH].

His conclusion is very similar to my conclusion but better!

If you can go IE 9+ / Android 3+, inline SVG is better at pretty much everything than icon fonts. If you need the deeper browser support, I feel like an inline SVG fallback would be too big of a pain to be worth it.

Definitely worth a read if you are interested in this topic.

Continue reading…


Foundation connecting my (front-end) world

foundation5
I don’t even remember how I first found Zurb Foundation, but it have almost become my obsession. It’s everywhere.

One of the smartest guy I know and a good firend in Tokyo, Keiichi, contacted me and told me that he is going to bootstrap this StartUp. I have not heard from him in a while, And I was really glad and I told him that I will help create his website. Immediately, I started thinking I will create something cool with Foundation and Jekyll. I need to update my repo

Lucky for me, Kianosh Pourian and Dale Sande’s excellent Sass in Depth book came out about the time when I was getting into Foudation and its beautifully crafted SASS, it has been a great help and timing. I took Advanced Foundation class last year, I’ve converted my blog to Reverie-based one and I have successfully integrated Pattern Lab (another of my obsession, you could say — and I know, I know, Brad Frost says “do the sh*t yourself…”) with Foundation for my work, on vagrant box, no less!

So perhaps it was just fitting that I made my first Open Source pull request on github today. On Foundation. Yay, Open Source!

Anyway, perhaps foundation js and require js integraton next? Who knows!

I am a huge fan of Zurb Foundation and I hope they will stop in Boston for their World Tour and perhaps organize a Boston Foundation Users Meetup? Not sure if there are enough interest? Maybe we can host Cascade Bos meetup?


Vagrant+Puppet is AMAZING!

vagrant-up

At my work on my team, we have five main developers. Two of us are on Mac, while the other three are on Windows. I, as a front end guy in the group, am advocating for SASS/Compass be used prominently on the redesign, like in depth use (great book btw). There is one little problem — workflow. To be able to developed locally, SCSS require pre-compiling, and it uses ruby. Mac comes with ruby, but Windows does not. Grunt could do it, but then you need node.js/npm. If I want everyone on my team to have it, I need to install all of those (ruby/sass/compass/node/npm/grunt)… that would be just way too much customization/potential places to break.

Enter Vagrant and puppet. It is just amazingly awesome. You can create a linux machine on VirtualBox – which is an open source VM. I am running latest puppet-enabled ubuntu box available from puppetlab, which is Ubuntu 13.10 (saucy Salamander).

Dealing with Vagrant really tickled my inner linux nerd-ness. Once upon a time, I was involved with running a group called BJLUG, Boston Japanese Linux Users Group — this was way back when, when you still had to pay for Apple’s CJK Language Pack to have Japanese installed on your machine, and I figured out that if you use linux, it was FREE to write email back to Japan etc. There were certainly a lot of LUGs back then. Not sure if they are around any more.

I learned how to install puppets, thanks to my colleague Kevin D. who patiently walked me through it, and was able to create box with apache and the likes.

Now we all need to download, and then just run vagrant up and vagrant provision. And your local scss changes, on your filesystem, using your favorite text editor on your OS, will be watched and css is compiled without any additional work. This is beautiful.

The Vagrant CLI documentation was useful for me, who have not used Vagrant before.

Some useful command:

  • vagrant up – Start the VM
  • vagrant provision – Provisions the VM
  • vagrant ssh – Logs into the VM with ssh
  • vagrant reload --provision – Restarts and provisions the VM
  • vagrant suspend and vagrat resume – If you don’t want fully shut the VM down
  • vagrant halt – Stops the VM
  • vagrant destroy – Deletes the VM

Actually, first time I saw use of vagrant was WordPress VIP team’s quick start. It includes all their server hosting environment. They do good work there over at VIP. And also there are ways to have local WordPress environment in vagrant such as Varying Vagrant Vagrants (VVV) environment. I am still using MAMP for my local wordpress theme/plugin development (when I have “free time”) but perhaps I should reconsider it. For new version of Daigo.org theme?
Continue reading…


Font awesome, and php split() to explode()

Couple of update notes

Noticed couple of things that I am doing over and over, and everytime I am doing it, i had to look it up (Thank you, Google for making us dumber and dumber…) thought to put it in a place for my future self…

Going from font-awesome 3.2.1 to the current 4.0.3

It used to be
<i class="icon-btc"></i>
but now you have to make it
<i class="fa fa-btc"></i>

PHP Split

split() is deprecated as of PHP 5.3.0. So if you want to convert something like 1974-05-25 into an array, an was using it, you can use explode() insted, like so:

list($Year,$Month,$Day) = explode('-',$info[dob]);
$stampeddob = mktime(12,0,0,$Month,$Day,$Year);
echo date("M. j, Y",$stampeddob);