This proposal is a larger suggestion that is out of scope for the Community Tech team. Participants are welcome to vote on it, but please note that regardless of popularity, there is no guarantee this proposal will be implemented. Supporting the idea helps communicate its urgency to the broader movement.
Allow for searching for the start and end of a string
Problem: If one searches for a regex term in Wikipedia, such as intitle:/[a-z]{5} [a-z]{6}/ , they get junk like Topographic prominence.
Proposed solution: Adding a start and an end function to this term would remove the junk.
Who would benefit: People who want to search pages by their enumeration. This is especially true with Wiktionary. Also, this would help reduce load on the Wikimedia servers, as they would start need to search pages that started/ended with the start/end query.
More comments: You can use the standard regex start and end functions, ^ and $.
Indeed, the underlying regular expression engine we do use does not support such syntax. Adding support for this (while not completely impossible and without going into the details) would require a non negligible effort to do properly (adapt/rewrite the regex parser, possibly use reserved unicode characters as start/end markups in the index, ...). DCausse (WMF) (talk) 10:55, 30 January 2023 (UTC)[reply]
Indeed! I haven't thought about this possibility but this does seem to be a valid workaround, for searching for pages that end with test one can search for intitle:/test/ -intitle:/test./, see it in action on wiktionary. @CitationsFreak: would documenting this workaround be good-enough? DCausse (WMF) (talk) 14:59, 3 February 2023 (UTC)[reply]
Actually after discussing with my colleagues, this workaround does not quite work because it will ignore words that have the suffix you want elsewhere in the word, e.g. intitle:/ed$/ is not equivalent to intitle:/ed/ -intitle:/ed./ as the later will exclude educated but actually should match with the former. DCausse (WMF) (talk) 18:19, 8 February 2023 (UTC)[reply]
Voting
Support if possible, but it would probably mean introducing proper regexp support (e.g. PCRE) which would have many other major benefits. Certes (talk) 21:47, 10 February 2023 (UTC)[reply]