html_sourcelink_suffix don't allow anymore to make search result user friendly

[pybride]

Subject: html_sourcelink_suffix don't allow anymore to make search result user friendly

Problem

• To make the search result user friendly (see HTML Search results aren't reader friendly #1618), the only solution for the moment is to run again sphinx, generating text output, which is filtered and then placed in the generated HTML _sources folder.
  The sphinx-pretty-searchresults plugin makes the filtering better, but it is still a text output from sphinx.

Therefore, if the documentation source has a something.rst file, a something.txt would be put in the _sources folder.

But since sphinx 1.5 and #2454, the behavior is to add the html_sourcelink_suffix to the source file name: the javascript search will look for a file something.rst.txt

Unfortunately, this new behavior cannot be disabled (not like the previous #2399).

I see three possible solutions:

• add text output option text_file_suffix to allow the second run to generate file like something.rst.txt
• put again the html option html_sourcelink_keep_suffix, so that when set to false, the javascript will look for something.txt
• correct the whole html search (HTML Search results aren't reader friendly #1618) by integrating the sphinx-pretty-searchresults

Environment info

• Python version: 2.7
• Sphinx version: 1.5.x

[pybride]

As workaround, in searchtools.js_t, replace the line
$.ajax({url: DOCUMENTATION_OPTIONS.URL_ROOT + '_sources/' + item[5] + (item[5].slice(-suffix.length) === suffix ? '' : suffix),
with
$.ajax({url: DOCUMENTATION_OPTIONS.URL_ROOT + '_sources/' + item[5].slice(0, -4) + '.txt',

[ekpgh]

For what it's worth, sphinxprettysearchresults was updated this week to v0.2.0, which works with Sphinx 1.5.

[tk0miya]

@pybride Is this still needed?

[pybride]

@tk0miya I just returned from a long trip, so I didn't tested the new sphinxprettysearchresults, but apparently it isn't needed anymore. I will close it.

[tk0miya]

:-)