Help with regular expressions?

faldringham · Yesterday, 05:07 AM

I want to filter out emails that contain any of the words "Trash", "Reminder", or "Landscaping" in the subject. How to do that? I have tried several variations of

(?i)(trash|reminder|landscaping)

with <Subject> <matches regular expression>

in the no-preview rules setup. Can't get it to work.
Do I need quotation marks? No quotation marks? Am I indicating case-insensitive incorrectly? Do I need to use dot-asterisk(s) to indicate that anything can precede (or follow) the matching term?

JeremyNicoll · Yesterday, 06:50 AM

You don't need quotes.
You don't have to indicate case insensitivity.
You don't need leading or trailing dot asterisk

Just: (trash|reminder|landscaping) should work.

Whn googling for help bear in mind that Sieve (which is the language that the filter rules get translated into from what you specify in the FM GUI) uses "POSIX" regexes. It's not PCRE.

You might find this useful: https://www.regular-expressions.info/posix.html

Also: https://www.regular-expressions.info/refflavors.html ... which allows you to follow a link to a sub-topic (eg character classes) & will then offer you the chance to nominate two flavours of regex (eg PCRE if you know that well) & POSIX ... & let you see to what extent those flavours of regex implement various things.

I should add that - if you look at the Sieve rules generated from the GUI - you'll see some escaping of characters - which you don't (as far as I know ) need to do yourself in your GUI rules - indeed if you do escape stuff in the GUI it'll (probably) be doubled in what's generated. The distinction matters because most of the advice on the internet for this sort of filtering (Sieve) is for people writing rules directly in Sieve (which in fact FM's system allows you to do if you need to).

If there's any chance that you'd want to write your own Sieve rules you really need to read the RFCs (technical specs for lots of internet stuff) for Sieve ... which are hard going if you've not got an academic Computing background. These include:

https://tools.ietf.org/html/rfc3028 the original Sieve RFC, made obsolete by

https://tools.ietf.org/html/rfc5228 a newer base Sieve specification

https://tools.ietf.org/html/rfc2244 describes comparators, eg for testing INTEGERS

https://tools.ietf.org/html/rfc3431 adds: counting and value-comparison tests; see rfc5231 too

https://tools.ietf.org/html/rfc5173 adds: testing of BODY contents

https://tools.ietf.org/html/rfc5229 adds: "Variables" support

https://tools.ietf.org/html/rfc5231 .... improves rfc3431

https://tools.ietf.org/html/rfc5293 adds: "editheader" support to add/delete headers

https://tools.ietf.org/html/rfc5429 adds: REJECT and extended EREJECT

https://tools.ietf.org/html/rfc6785 adds: IMAP

faldringham · Yesterday, 10:50 PM

Thanks. I will give it another try.

Yesterday, 05:07 AM	#1
faldringham Junior Member Join Date: Apr 2009 Posts: 28	Help with regular expressions? I want to filter out emails that contain any of the words "Trash", "Reminder", or "Landscaping" in the subject. How to do that? I have tried several variations of (?i)(trash\|reminder\|landscaping) with <Subject> <matches regular expression> in the no-preview rules setup. Can't get it to work. Do I need quotation marks? No quotation marks? Am I indicating case-insensitive incorrectly? Do I need to use dot-asterisk(s) to indicate that anything can precede (or follow) the matching term?

Yesterday, 06:50 AM	#2
JeremyNicoll Cornerstone of the Community Join Date: Dec 2017 Location: Scotland Posts: 850	You don't need quotes. You don't have to indicate case insensitivity. You don't need leading or trailing dot asterisk Just: (trash\|reminder\|landscaping) should work. Whn googling for help bear in mind that Sieve (which is the language that the filter rules get translated into from what you specify in the FM GUI) uses "POSIX" regexes. It's not PCRE. You might find this useful: https://www.regular-expressions.info/posix.html Also: https://www.regular-expressions.info/refflavors.html ... which allows you to follow a link to a sub-topic (eg character classes) & will then offer you the chance to nominate two flavours of regex (eg PCRE if you know that well) & POSIX ... & let you see to what extent those flavours of regex implement various things. I should add that - if you look at the Sieve rules generated from the GUI - you'll see some escaping of characters - which you don't (as far as I know ) need to do yourself in your GUI rules - indeed if you do escape stuff in the GUI it'll (probably) be doubled in what's generated. The distinction matters because most of the advice on the internet for this sort of filtering (Sieve) is for people writing rules directly in Sieve (which in fact FM's system allows you to do if you need to). If there's any chance that you'd want to write your own Sieve rules you really need to read the RFCs (technical specs for lots of internet stuff) for Sieve ... which are hard going if you've not got an academic Computing background. These include: https://tools.ietf.org/html/rfc3028 the original Sieve RFC, made obsolete by https://tools.ietf.org/html/rfc5228 a newer base Sieve specification https://tools.ietf.org/html/rfc2244 describes comparators, eg for testing INTEGERS https://tools.ietf.org/html/rfc3431 adds: counting and value-comparison tests; see rfc5231 too https://tools.ietf.org/html/rfc5173 adds: testing of BODY contents https://tools.ietf.org/html/rfc5229 adds: "Variables" support https://tools.ietf.org/html/rfc5231 .... improves rfc3431 https://tools.ietf.org/html/rfc5293 adds: "editheader" support to add/delete headers https://tools.ietf.org/html/rfc5429 adds: REJECT and extended EREJECT https://tools.ietf.org/html/rfc6785 adds: IMAP Last edited by JeremyNicoll : Yesterday at 07:50 AM.

Yesterday, 10:50 PM	#3
faldringham Junior Member Join Date: Apr 2009 Posts: 28	Thanks. I will give it another try.