Things I can’t remember: More GREP goodness

When cleaning up mailing lists, I often need to remove the zip+4 info. It is usually at the end of a line so this grep works in BBEdit to find them.


-[0-9]{4}$

Look for a dash followed by the numbers 0-9 four times and then an end of line.

Working on a word list, we needed to find and delete all of the words with 1, 2, or 3 letters. So we use this code switching out the 1 for 2 and 3. Note the ^ and $. This means that we start at the beginning of the line look for the pattern and nothing else until the end of the line.


^[a-zA-Z]{1}$

We could use the same technique to find all of the words with more than 5 letters, but then we’d have to do a bunch of searches. Instead we used this.


[a-zA-Z]{5}[a-zA-Z]

Here we look for pattern of five letters and then look for one more. Not to belabor the point, but suppose together is in the word list. It has five letters, toget and one more h. It also happens to have some more letters after that, but we don’t particularly care. We just care that there are more than 5.

ON DUPLICATE KEY UPDATE

When processing addresses for postcard mailings we get a lot of addresses that don’t pass the VistaPrint address validation process. Fortunately, VistaPrint has a button that lets you download the addresses. What makes it easy for me to mark the bad addresses in my database is that I don’t use a salutation and instead put the customer number in the salutation field. I strip out all of the address information and keep just the customer ID. I then massage the IDs so that they look like this.


INSERT INTO `customers` (id, BadVistaPrint) VALUES 
(15771, 1), 
...
(16248, 1), 
(16249, 1)
ON DUPLICATE KEY UPDATE BadVistaPrint=VALUES(BadVistaPrint)

What this code does is look for customer ID and then update the BadVistaPrint value. Not sure why it uses the ON DUPLICATE KEY syntax. Note that the last row before the ON DUPLICATE KEY does not have a comma.

I paste this into the SQL section of phpMyAdmin and I my database is now updated with bad addresses. I can manually look for easy to fix addresses e.g. missing numbers, zip codes, etc. and look them up. There are lots of addresses that are mailable but not in the VistaPrint database e.g. University departments and rural route addresses and I mail them separately.

I also use the USPS zip code finder to validate addresses.

Improving legibility of OSX Yosemite

The latest update to OSX incorporates design elements from iOS7. If you hate iOS7, you’re going to really hate Yosemite.

However, there are a couple of things you can do to restore screen legibility. Open up the Accessibility preference pane in System Preferences. Click on the checkbox labelled increase contrast. It automatically checks the box for Reduce Transparency. This does two things. First, it stops background colors from bleeding through menu bars and option panels. Second, it darkens the text so it is more legible.

That still isn’t enough for my non-retina displays, so I also move the Display Contrast up to the first tick mark. Finally, on the physical monitor controls I reduced the brightness a bit. I’ve been using it all morning and the text is readable. Unfortunately, there is no way to fix the contrast on buttons so that they are more readily identifiable, but I suppose I’ll get used to it.

Safari:
I like to see the entire URL when maintaining my websites, so in the Advanced Safari preferences I check, “Show full website address”. In the security section I turned off “Autofill web forms using: Credit cards”. Then under the View menu item, I turned on “Show Tab Bar” and “Show Status Bar”.

Dock:
In the Dock preference pane, I disable “Double-click a window’s title bar to minimize”. Prevents you from accidentally disappearing windows. This isn’t a new feature, but I just found out how to disable it.

My usual method for pinning the dock the bottom RHS corner doesn’t work, so if anyone figures out how to do it, pls let me know.

Finder:
In the General preference pane, change “Sidebar icon size” to large. Improves legibility.

Lots of other annoying things I’ll probably get used to, though it would be nice if someone figures out how to change the folder color in the Finder from the garish blue. The good news is that everything I’ve tried so far works just as before—including all the software I wrote.

Validating apps so they will run on original iPads

Most of my apps were built with iOS4 or iOS5. They don’t require any of the new features of iOS6,7,8 so they will run fine on older devices that can’t upgrade. (I do have conditional code for iOS7s look and feel, and of course they run on the new iPhone sizes.) The original iPads will run iOS 5.1.1 but not any version that is newer. Unfortunately, the dropdown for choosing the minimum deployment target only lists iOS 5.1, which won’t validate.

Deployment dropdown

So how do you change this,
Deployment 5.1
to this?
Deployment 5.1.1

In my case I happen to be lucky in that Xcode 5 asked me if I wanted to fix the error and fixed it for me. So my project default is 5.1.1 and my problem is just to get it into the build settings. After much experimentation, I discovered that bringing up the deployment dropdown in the build phases column, closing it, then hitting the delete key fills in the box with the project default. I don’t know how to get the project default to contain 5.1.1 if it doesn’t already.

Processing remove requests and bad email addresses in our customer list.

Update 2016-05-25. I found a much easier way.

Remove Requests

We don’t like to bother our customers with lots of emails, but from time to time we let them know about new products and sales. Some of them use the remove link on the bottom of the email to unsubscribe from the list. I have a rule in Apple Mail that automatically routes the remove response to a folder. We usually get a couple of remove requests each time we send out a mailing and we usually process them manually. I wanted to automate the process a bit and create a master list of email addresses that do not want our mailings. That way if someone orders from us under a different name or address, we won’t be sending them email if they opted out earlier.

After looking around in the Library/Mail folder it looks like our Remove folder is located at


 ~/Library/Mail/V2/Mailboxes/Remove.mbox/

I CD to that folder and then redirect the results of a recursive grep command to a file on the desktop. Since they are responding to our email, I look for the From: portion of the email. Note the period after the search term. I means to look for all files in the current directory. The -r says to look in all files in directories below this one too.


grep -r "From: " . > ~/Desktop/Remove.txt

Then I open the file in BBEdit and remove extra lines, e.g. anything with our company name.

I only want the email addresses, so I can use this grep line to remove everything before the address.


.*<

I can remove everything after the address with this.

>.*

Sort the lines and process duplicates and you’re done. I then import the email addresses into my MySQL database.

Bad Domain

Our mail server will try to find missing domains for a few days and generates warning messages. When it finally gives up it generates a message with the subject “Returned mail: see transcript for details”. The nice thing about these messages is that the failed address is easy to find. It looks like this:


The following address(es) failed:

 abby612@earthink.net
   retry timeout exceeded

In this case it probably couldn’t find the server because it was looking for earthink but the email was probably to earthlink. When I pull them out of the failure message, I correct obvious mis-typing before adding them to the bad email database.
The code to find the addresses is:


grep -r "^ .*@.*$" . > ~/Desktop/Failed.txt

The code looks for lines starting with a space, then an email address, then the end of the line. There are surprisingly few false positives e.g. lines that contain other things than just an email address.

Bad Addresses

It’s much harder to remove the bad email addresses since the format from different email providers varies tremendously. I ended up looking for lines that contain an “@” and doing a lot of manual cleanup. I’ll see if I can figure out a better way next month.