openzim / youtube

Create a ZIM file from a Youtube channel/username/playlist
GNU General Public License v3.0
43 stars 26 forks source link

kiwix-serve search is broken when video title contains an apostrophe, with TED-Ed ZIM file #143

Closed holta closed 3 years ago

holta commented 3 years ago

Summary:

Example: https://youtu.be/dcZ0BXJYlUA is a video that is a part of the latest Kiwix TED-Ed ZIM file (teded_en_all_2021-01.zim, published 2021-01-16) as can be seen here: http://iiab.me/kiwix/teded_en_all_2021-01/A/building_the_world_s_largest_and_most_controversial_power_plant_alex_gendler.html

Recap — users cannot search for the above video, because even after it appears in the search box's autocomplete, as follows:

Building the world's largest (and most controversial) power plant - Alex Gendler

But then when you try to click it, the apostophe is expanded as follows:

Building the world's largest (and most controversial) power plant - Alex Gendler

FYI/Context: the very latest kiwix-serve 3.1.2-4 is in use on the above-mentioned server: http://iiab.me/kiwix/teded_en_all_2021-01/

PS please move this ticket to another repo if it belongs elsewhere. Thank you.

kelson42 commented 3 years ago

@holta Might that be a duplicate of https://github.com/kiwix/kiwix-lib/issues/398 ?

holta commented 3 years ago

@kelson42 that's really great news if PR https://github.com/kiwix/kiwix-lib/pull/458 might have fixed this just 2 days ago.

To allow for testing: will a new build of kiwix-tools (incl kiwix-serve) be published to http://download.kiwix.org/release/kiwix-tools/ including this change?

kelson42 commented 3 years ago

@holta Would you be able to give a try to http://download.kiwix.org/nightly/2021-03-06/ ? At some point in the next weeks a new build/release will be published.

holta commented 3 years ago

It's likely fixed by http://download.kiwix.org/nightly/2021-03-06/ that I tested using http://iiab.me/kiwix/teded_en_all_2021-01/

Caveat: occasionally the search box still displayed a few of the following: (after clearing Chrome browser's cache, before relaunching the browser)

Building the world's largest (and most controversial) power plant - Alex Gendler
containing 'STRING-YOU-JUST-TYPED'...

(Possibly the Chrome browser is not fully clearing its cache for pages still residual in memory.)

holta commented 3 years ago

At some point in the next weeks a new build/release will be published.

Great news. FYI the new release of kiwix-tools would appear to be 45% larger (100MB instead of 69MB) which is fine, but what are the reasons if you have a moment?

Or if you don't have time, ideally its new functionality/changes can be summarized here? https://github.com/kiwix/kiwix-tools/blob/master/Changelog

PS Feel free to close this issue as fixed and resolved. As the above-mentioned intermittent glitches appear to have been transient as a result of the Chrome browser retaining stuff in memory (clearing a browser's cache is not quite enough!)