See bug 234464 for the context of this bug. This is about how ordinary search terms (e.g. words) are handled by the Search tool. Google behaviour (see http://www.google.com/help/basics.html ): =============================================================== Google finds occurrences of the _word_ anywhere in the document. * Automatic "and" Queries Google returns pages that include all of the search terms (words). There is no need to include "and" between terms. To restrict a search further, just include more terms. * Capitalization Google searches are not case sensitive. * Word Variations (Stemming) Google now uses stemming technology. Thus, when appropriate, it will search not only for your search terms, but also for words that are similar to some or all of those terms. Any variants of your terms that were searched for will be highlighted in the snippet of text accompanying each result. * Automatic Exclusion of Common Words Google ignores common words and characters, as well as certain single digits and single letters. In this case, Google will indicate it on the results page. Words can be forced to be included by prefixing them with "+", or with a phrase search, i.e. putting quotation marks around two or more words. Quicksearch (see http://www.squarefree.com/bugzilla/quicksearch-help.html ): ============================================================================ Quicksearch finds occurrences of the given string as a _substring_ in any of the following fields of a bug: summary, status whiteboard, product, component, and in some cases also url. Description & comments are not searched by default. * Automatic "and" Queries same as Google * Capitalization same as Google * Word Variations (Stemming) Not directly applicable due to the use of substring matching. If you do the stemming yourself before entering the search terms, you should get all the results you get with Google-technology, and some more. * Automatic Exclusion of Common Words This not done directly because it does not seem to be necessary. There are however some conditions to exclude product / component fields from the search: - if a word is too short, i.e. less than three characters - if it is mentioned on a special blacklist defined by the bugzilla maintainer product/component can still be searched by using a prefix (e.g. product:).
If we use MySQL's fulltext search feature we're limited to the features it supports, but fortunately it does everything mentioned in this bug, although stemming requires boolean-mode searches, which we haven't enabled yet because they only work in MySQL 4.0, and we're still supporting 3.23.
One last thing that seems to belong here: Quicksearch uses #foo to express the condition that a summary "contains the word" foo instead of just the substring. I suspect that such #foo terms are handled similar as in a fulltext search limited to summaries, but I'm not sure.
Assignee: ui → query-and-buglist
Component: User Interface → Query/Bug List
You need to log in before you can comment on or make changes to this bug.