Here's a challenge for folks - something someone may already have written perhaps? How to do a reasonable (doesn't have to be 100% perfect) pattern match on a string to determine whether or not it's a properly-structured email address. I DON'T mean is
the address valid in as far as it exists as a registered email address. Just whether or not the string of characters is potentially OK as an email address

Rob

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Since top level domains are constantly being added, instead of a hard-wired check, why not just call out to a service like https://validateemailaddress.org/ (or one of those listed at https://duckduckgo.com/?t=ffab&q=validate+e-mail+address&atb=v144-1&ia=
web)?

Regards
– Bhaskar

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 09.11.20 19:50, rtweed wrote:

How to do a reasonable (doesn't have to be 100% perfect) pattern match on a string to determine whether or not it's a properly-structured email address.

Hi Rob,

I have found this one in one of my old M-utilities (it's Cache-dialect,
so you have to adopt it to your needs):


// Check E-Mail address format
// mx=1: check the MX record too
//
ChkEml(eml,mx=0) Public
{
s ok=0 i $l(eml,"@")=2, $zcvt(eml,"O","UTF8")=eml s
eml=$zcvt(eml,"l"), ok=1
s:ok ok=$tr($p(eml,"@"),"+-_","AAa")_"@"_$tr($p(eml,"@",2),"-","A")?1ln.an.(1"."1.an)1"@"1.(1ln.an1".")2.25a
s:ok&mx ok=$$ChkMX($p(eml,"@",2),2)
q ok
}

// Check existence of a MX-Record
//
ChkMX(dom,tim=2) Public
{
s dom=$zcvt(dom,"l"),$t=0 d OsCmd("nslookup -query=mx "_dom,.data,tim)
f i=1:1:data s line=$tr($zcvt(data(i),"l"),$c(9)," ") i $p(line,"
")=dom, line["mail exchanger" q
q $t
}

// OS-Command
//
OsCmd(cmd,&rsp,tim=1) Public
{
k rsp s rsp=0, io=$i o cmd:"QRI":5 q:'$t
while $t { u cmd r ans:tim u io s rsp($i(rsp))=ans }
c cmd u:io]"" io
}


Regards,
Julius

--
An old Windows has old security holes.
A new Windows has new security holes.
Another OS has other security holes.
For safety you must care yourself.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

For my purposes I just want a pattern matching algorithm to determine if an email address is correctly formatted according to the rules. I found this which outlines those rules:

https://help.returnpath.com/hc/en-us/articles/220560587-What-are-the-rules-for-email-address-syntax-

If these are correct, then it means a simple use of the ? pattern match is probably not going to be sufficient. Here's my somewhat verbose logic with some tests. Any suggestions/improvements/corrections??

replace(string,substr,to) ;
i $zv["GT.M" QUIT $$^%MPIECE(string,substr,to)
QUIT $replace(string,substr,to)
;
isValidEmail(email) ;
;
n chk,domain,dupFound,i,name,specialChars
;
; just a single @ ?
i $l(email,"@")'=2 QUIT 0
;
s name=$p(email,"@",1)
; missing name?
i name="" QUIT 0
; starts or ends with a letter or number?
i $e(name,1)'?1AN QUIT 0
i $e(name,$l(name))'?1AN QUIT 0
; duplicated special characters?
s specialChars=".!#$%&'*+-/=?^_`{|"
s dupFound=0
f i=1:1:$l(specialChars) d q:dupFound
. n s,ss
. s s=$e(specialChars,i)
. s ss=s_s
. s chk=$$replace(name,ss,s)
. i chk'=name s dupFound=1
i dupFound QUIT 0
;
; any other character than alphas, numbers or special characters?
s chk=$tr(name,specialChars,"")
i chk'?1AN.AN QUIT 0
;
s domain=$p(email,"@",2)
i $e(domain,1)'?1AN QUIT 0
i $e(domain,$l(domain))'?1AN QUIT 0
; missing top-level domain?
i $l(domain,".")<2 QUIT 0
; missing domain name
i $p(domain,".",1)="" QUIT 0
; missing intermediate domain parts
i $$replace(domain,"..",".")'=domain QUIT 0
; invalid characters in domain name?
s chk=$tr(domain,".-","")
i chk'?1AN.AN QUIT 0
; looks like it's formatted OK
QUIT 1
;
incorrectTest(email)
i $$isValidEmail(email) d
. w email_" was invalid but passed",!
e d
. w email_" was correctly rejected",!
QUIT
;
correctTest(email)
i $$isValidEmail(email) d
. w email_" passed OK",!
e d
. w email_" was OK but failed",!
QUIT
;
emailTests ;
;
d correctTest("john.doe@gmail.com")
d correctTest("john.doe43@domainsample.co.uk")
d incorrectTest(".doe@gmail.com")
d incorrectTest("@domainsample.com")
d incorrectTest("johndoedomainsample.com")
d incorrectTest("john.doe@.net")
d incorrectTest("john.doe43@domainsample")
d correctTest("john.do%e43@domainsample.com")
d incorrectTest("john.do%%e43@domainsample.com")
d incorrectTest("john.do%e43@domainsample..com")
d incorrectTest("john.do%e43@domainsample.com.")
d incorrectTest("john.do(e43@domainsample.com")
d incorrectTest("john.doe43@domai%nsample.com")
;
QUIT
;

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Looks like Google Groups mashes up the test email addresses. Here they are with the @ replaced with _at_ to bypass that

d correctTest("john.doe_at_gmail.com")
d correctTest("john.doe43_at_domainsample.co.uk")
d incorrectTest(".doe_at_gmail.com")
d incorrectTest("_at_domainsample.com")
d incorrectTest("johndoedomainsample.com")
d incorrectTest("john.doe_at_.net")
d incorrectTest("john.doe43_at_domainsample")
d correctTest("john.do%e43_at_domainsample.com")
d incorrectTest("john.do%%e43_at_domainsample.com")
d incorrectTest("john.do%e43_at_domainsample..com")
d incorrectTest("john.do%e43_at_domainsample.com.")
d incorrectTest("john.do(e43_at_domainsample.com")
d incorrectTest("john.doe43_at_domai%nsample.com")

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)