[#8097] - Fix for #8034 - smart search fail on $
- Fixed in Code Base
- 8 May 2016
- Medium
- Build: staging
- # 8097
- Diff
- Fedik:highlight-empty-fix
- Success continuous-integration/travis-ci/pr The Travis CI build passed Details
- Success JTracker/HumanTestResults Human Test Results: 2 Successful 0 Failed. Details
User tests: Successful: Unsuccessful:
this fix for #8034
the problem was that $ is striped and the highlighter script tries to highlight the emptiness, that cause high resource usage by highlight.js
for testing please look the description in related issue 
Steps to reproduce the issue
- enable smart search & smart search module
- index content so it's not empty
- search for "$" (single dollar sign)
Expected result
some search results
Actual result
javascript loop resulting in massive memory leak and eventual browser crash
System information (as much as possible)
browser = firefox
OS = Win 10
Additional comments
| Status | New | ⇒ | Pending |
joomla-cms-bot
- | Labels |
Added:
?
|
||
| Title |
|
||||||
the problem was that $ is striped and the highlighter tried highlight emptiness, that cause high resource usage by highlight.js
, just noticed that for search works as expected
$ is striped not in javascript, but somewhere in Smart Search component while handle the input ... I have not searched 
I haven't checked the code, but I would think that all "punctuation" characters will be stripped. Same for the indexer. That would be expected behaviour.
| Category | ⇒ | Search |
@chrisdavenport Not so sure we can consider $ , € etc ..... a punctuation character
or at least not in all languages
so i think com_search is doing the correct behavior
what i'm missing?
For the kind of websites I normally work with I wouldn't want these characters cluttering up the index. However, I can imagine that there would be occasions where some websites might want to retain them. I would suggest that this needs to be configurable because it depends on the particular application. It's rather like how we currently have a list of "stop words" that are not indexed. I think we should have a configurable list of "stop characters" that are discarded and/or regarded as delimiters. I'd be happy to review a PR for that feature.
p.s.
i'm sorry for my shortsight, i usually work with ($, €) characters 
damned banking software work
I have tested this item 
successfully on f3f7e62
The specific issue reported by @Fedik is fixed by this PR.
The consequential issue reported by @alikon should be the subject of a separate PR along the lines I suggested.
This comment was created with the J!Tracker Application at issues.joomla.org/joomla-cms/8097.
I have tested this item successfully on f3f7e62
This specific issue is fixed. Thanks.
This comment was created with the J!Tracker Application at issues.joomla.org/joomla-cms/8097.
| Status | Pending | ⇒ | Ready to Commit |
RTC. Thanks
This comment was created with the J!Tracker Application at issues.joomla.org/joomla-cms/8097.
| Labels |
Added:
?
|
||
brianteeman
- | Milestone |
Added: |
||
| Status | Ready to Commit | ⇒ | Fixed in Code Base |
| Closed_Date | 0000-00-00 00:00:00 | ⇒ | 2016-05-08 13:46:42 |
| Closed_By | ⇒ | Kubik-Rubik |
| Labels |
Removed:
?
|
||
tested succesfully now no more loop and browser crash
not sure if is it out of scope of this pr but
something is still wrong :
wrong text $ sign is missed
wrong search result the $$$ is not matched/highlight
