0
I Use This!
Inactive

Commits : Listings

Analyzed about 6 hours ago. based on code collected about 7 hours ago.
May 31, 2023 — May 31, 2024
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
some simple caching. After implementing it's unclear if there is a lot of benefit here, though. More... about 12 years ago
a bit of efficiency stuff. Actually get the most recently published work to display. More... about 12 years ago
update to reflect caching work More... about 12 years ago
Another refactor, in a sense rotating how the expirations are done. I think this is clearer, because all the code specific to a particular kind of observed object is now isolated. More... about 12 years ago
fix another bug - if a publication's name changes then we potentially have to expire the cache corresponding to both its new _and_ old sort_name. More... about 12 years ago
Refactored. We have a general method that can take the record passed to the sweeper and call a lot of methods on it to determine if something has happened that should require expiration. We then have another method that uses that and a supplied block to figure out what ids to expire and return as an array. This simplifies the code in the actual sweeper. More... about 12 years ago
Some refactoring on determine expiration. More... about 12 years ago
fix bug - when an object is destroyed then nothing about it will be marked as changed (i.e. obj.changed? = false), but we still want to act. So check also if it is destroyed. I'll refactor this presently. More... about 12 years ago
fix small potential bug - we need to cycle the row class outside of the row caching. This way if a publication is added or removed, causing a rebuild of the page cache, the rows will get rebuilt with the right class. More... about 12 years ago
minor refactor More... about 12 years ago
Also cache the footer. More... about 12 years ago
Cache this. More... about 12 years ago
As if to demonstrate the inherent complexity in caching/sweeping, fix bug - publication index table also depends on state of contributorship(s) attached to a work. Take this into account. More... about 12 years ago
Add more caching. More... about 12 years ago
Don't instantiate the publishers in the controller. More... about 12 years ago
Trying one-fetch for publisher data - generating this view is still really slow on production. More... about 12 years ago
remove some unnecessary code More... about 12 years ago
Start of attachment_fu -> paperclip conversions. Migrations to fix up database and move old content to where paperclip expects it. Also removed what seem to be unused columns in attachments table. Changes to Attachment and subclasses to use paperclip Changes to import to be able to find imported file Fixed controller for import. More... about 12 years ago
Use make_resourceful as a gem instead of a plugin Update gems. More... about 12 years ago
Get the work counts for the index_table more efficiently. Since the publisher is used a few different ways it may be just as well to leave that part as it is. More... about 12 years ago
remove unused method More... about 12 years ago
Lookup most recent years and publication counts for the table en-masse. This improves the speed quite a bit. More... about 12 years ago
Cache rendering of shared/sherpa_icon in index views and make other appropriate changes to related views/partials. Some other efficiency measures - some of the displayed data we can get pretty easily from the database without instantiating objects. It may be possible to streamline this further by doing all the necessary work in a single query, but even doing many queries speeds things up significantly compared to instantiating the objects. More... about 12 years ago
enable caching in order to do some experiments with it More... about 12 years ago
note change to paperclip More... about 12 years ago
Add a rescue statement to image_url method. While this really should never be needed in practice, I think that there's a chance that someone trying to run the migration for stopwords while having this slightly later code may get an error without this here. On the other hand, I think (and hope) that if so this is the only change needed to prevent that, and it's unobtrusive so I'm willing to try it. More... about 12 years ago
The batch csv people upload also uses shared/_upload_form, so a change or two needed to be propagated there. More... about 12 years ago
I think this is the rest of what is needed to convert to paperclip. More... about 12 years ago
add more commentary More... about 12 years ago
Added backports gem, which enables use of make_resourceful gem with older rubies. More... about 12 years ago