Error with non-roman/UTF8? characters about tiley

[tommoor]

Answer might be to convert to romanji or fall back to something like a smiley face on a colour background. https://answers.yahoo.com/question/index?qid=20100323042314AAv36vd

[awsmsrc]

@tommoor I think 9ada943 closes this issue BUT I'm not sure on the 100% sire about the expected output

[awsmsrc]

its live anyway btw, i figured potentially incorrect values was better then broken images

[awsmsrc]

also this changes behaviour slighlty, the initials param can no longer be limited to 2 chars as jap chars are url encoded and are thus longer. I'm now just using a substring

[tommoor]

Nice, it's difficult to know what the right answer is here, ideally we'd produce real initials but I don't think that's possible without having the forename and surname separate in the DB.

Tried a couple of other character sets to see how the lib reacts:

http://tiley.herokuapp.com/avatar/11/%E6%96%B0%E6%B5%AA%E5%BE%AE%E5%8D%9A.png
http://tiley.herokuapp.com/avatar/11/%D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99%D7%AA.png
http://tiley.herokuapp.com/avatar/11/%ED%95%9C%EA%B5%AD%EC%96%B4.png

Seems to reliably produce ?? so perhaps we could catch this and replace with something more inviting... any thoughts?

[awsmsrc]

if you look at the app i have a function that parses the initials

I was thinking of maybe a middleware that takes that string and converts to a more global character set

on another note, I think we should have a first name, middle names, and last name column to the db, and lazily instantiate this and write it based on the full name
we can then change the sign up form as and when

On 9 Jun 2014, at 11:39, Tom Moor [email protected] wrote:

Nice, it's difficult to know what the right answer is here, ideally we'd produce real initials but I don't think that's possible without having the forename and surname separate in the DB.

Tried a couple of other character sets to see how the lib reacts:

http://tiley.herokuapp.com/avatar/11/%E6%96%B0%E6%B5%AA%E5%BE%AE%E5%8D%9A.png
http://tiley.herokuapp.com/avatar/11/%D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99%D7%AA.png
http://tiley.herokuapp.com/avatar/11/%ED%95%9C%EA%B5%AD%EC%96%B4.png

Seems to reliably produce ?? so perhaps we could catch this and replace with something more inviting... any thoughts?

—
Reply to this email directly or view it on GitHub.

[tommoor]

Looks like standard chars are now being returned wrongly:

http://tiley.herokuapp.com/avatar/10/TM.png

[awsmsrc]

Ill see if i can get a quick fix here

jeepers, I know why FYI this is NOT gonna be an easy problem to solve.
tileys api expects two roman alphanumeric chars, we need to think here, is this an api issue, should we be doing the formatting there?

On 9 Jun 2014, at 12:33, Tom Moor [email protected] wrote:

Looks like standard chars are now being returned wrongly:

http://tiley.herokuapp.com/avatar/10/TM.png

—
Reply to this email directly or view it on GitHub.

[tommoor]

Maybe throw some mocha around it? Spec would have caught this regression :) I'd imagine there are a few edge cases

[awsmsrc]

yeah, was thinking that. my self,
also gonna extend that hepburn lib to detect if it contains those chars and only change if necessary

On 9 Jun 2014, at 12:43, Tom Moor [email protected] wrote:

Maybe throw some mocha around it? Spec would have caught this regression :)

—
Reply to this email directly or view it on GitHub.

[tommoor]

closed by a2324bc