G'Day all... I have a strange issue with cfmail and email attachments... If the name of an email attachment contains enough diacritical characters then when the email is delivered the attachment filename will be corrupted. For example 'ThêQuickBrownFoxJúmpsOverTheDog.docx' is fine but 'ThêQuickBrøwnFoxJúmpsOverTheDog.docx' results in a file name resembling '=UTF-8QTh=C3=AAQuickBr=C3=B8wnFoxJ=C3=BAmpsOverTheDog.docx='. I'm guessing this is an issue with the length of the filename and handling of utf-8 characters because even a relatively short name like 'ThêQúickƁrøwñƑøx.docx' is corrupted into '=UTF-8QTh=C3=AAQ=C3=BAick=C6=81r=C3=B8w=C3=B1=C6=91=C3=B8x.docx='. The same files sent via Gmail to the same recipient are delivered with the filenames as you would expect. I'm thinking this is a Lucee bug but thought I would check here in case someone has seen this before (I couldn't find it in the backlog). Tested on 5.3.10.79-RC and 5.3.9.141 - and yes, I have the charset attribute set to 'utf-8'

https://github.com/lucee/Lucee/blob/5.3/core/src/main/java/lucee/runtime/net/smtp/SMTPClient.java#L739 ?

So? It isn't a Lucee issue?

dunno!

ThêQuickBrownFoxJúmpsOverTheDog.docx

ThêQuickBrøwnFoxJúmpsOverTheDog.docx

so the

ø

is triggering a different encoding?

I don't think so - the result '=UTF-8QTh=C3=AAQuickBr=C3=B8wnFoxJ=C3=BAmpsOverTheDog.docx=' suggests all the diacritic characters have been encoded. And if I replace the 'o' and change the 'u' to 'ú' instead the name is still corrupted.

the encoding is triggered by any character value gt 127 in the

StringUtil.isAscii

check

fair enough - so it must be the length of the encoded string that is causing the issue?

this ticket made that change https://luceeserver.atlassian.net/browse/LDEV-422

let me set those up - I was playing with them earlier but our mail send logic is built around handing the errors directly so sending 'async' is not possible without a lot of refactoring and regression testing - but I can plug a listener into my test code, no problem...

hmmm, it's passing in null for encoding

fileName=MimeUtility.encodeText(fileName, "UTF-8", null);

https://github.com/lucee/Lucee/commit/aeb37d0b080797a2aaea82aa9dca3847dbd2bcca https://stackoverflow.com/questions/21574745/java-encode-string-in-quoted-printable which defaults to Q encoding https://docs.oracle.com/javaee/1.3/api/javax/mail/internet/MimeUtility.html#encodeText%28java.lang.String,%20java.lang.String,%20java.lang.String%29 but Lucee currently doesn't use quoted-printable encoding (yet) we have a pending patch for that https://luceeserver.atlassian.net/browse/LDEV-4039

ok... thanks for digging into the issue! Do you want me to quote any of the above in the ticket - or is it better/easier for you to raise the ticket?

have a go 🙂

Should I post this on https://dev.lucee.org/ first?

nah, that banner in jira is from me, we have triaged this enough to file a ticket

ok

sure, I will create a testcase for this along with the ticket

Cool, thanks @pothys-mitrahsoft

Does this mean my quoted-printable patch might finally get some traction? 😁

like clock work.... 🙂

@zackster - nope, sorry, it actually got worse...