![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
This week(ish) in coding
If anyone missed it, we're going to be doing more update/Q&A posts from now on, and the first one went up on Monday: on LJ | on DW | on the blog. Feel free to leave any questions you might have in the comments on any of those sites!
Seriously, last weekend was like the attack of Murphy's Law - things were going along pretty swimmingly, and then, boom, about a dozen weird bugs in a row. We did deploy version 1329, though - samvara has release notes! Since then, I worked on a couple of small bugs: I realized caching needed to account for show/hide warnings preferences, the filters weren't generating properly on search, and I helped out with making html visible in summaries again.
Weird issues to resolve:
-I thought the filters were working, but I didn't take into account that people might tag their stories with multiple versions of the same tag (like Alternate Universe, AU) which messes with the results due to reasons redacted for tl;dr. Filter Supernatural fic by the Alternate Universe tag and you'll get some random results mixed in. Need a new approach on that one.I have sworn to resolve this by Saturday! Fixed and testing.
-Pagination goes a bit wacky sometimes - filter Buffy stories for Angel crossovers and the first page will claim to return the number of works tagged with BtVS *or* AtS instead of BtVS *and* AtS. What's weird is that it actually gets the right results, and the second page shows the accurate count. IDK, pagination gem weirdness (count vs. group_by in the automatically generated queries, I think).Odds are good that changing the underlying filter code per the first issue will fix this one too. Fixed and testing.
-You can't get rid of an OpenID. I fix this, I test it locally, it works, I deploy it, it doesn't work. Small changes don't seem to be cutting it, so I want to really go through the OpenID plugin and gem and figure out how they work and add some code to simplify how we're verifying these and see if there are other issues I can address while I'm in there.
-Translation caching - changes to the default text aren't visible. And I can't figure out how the caching is even set up, and it's causing problems with another plugin that would be handy to use. I'm a little dubious about the plugin we started with, and I'm also trying to think of ways to make the whole process more efficient and easier for the translators. Hopefully some ideas will percolate while I'm working on the other stuff.
-You can use tables in your stories now, but there are some extra spaces before them, and if you import a story with tables from google docs or word, there are a LOT of extra spaces before them. I leave that one to the rich text/regex experts!
-Caching of story blurbs/cast tags has to take into account that cast tags are affected by tag wrangling even if the work itself hasn't been updated. Ie, say you post a Smallville fic with the pairing tag "Clark Kent/Lex Luthor" and the character tags "Clark Kent" and "Lex Luthor". If the tags haven't been sorted by a tag wrangler, cast tags will be "Clark Kent/Lex Luthor, Clark Kent, Lex Luthor". Once the tag wrangler sets up the relationship between the characters and the pairing, the cast tags will condense down to just "Clark Kent/Lex Luthor".
(Am I missing things? I'm probably missing things.)
Other stuff:
-We've been discussing how to improve chapter navigation - we talked about it at the last AD&T meeting, and I want to try to put together a working mock-up to see what people think.
-We've been using/talking about using search for things other than works (users, tags, etc.) and getting some designs for more complex work searches, so I have a note to self to look deeper into the underpinnings of our search plugin (Thinking Sphinx) to get a better idea of what the best approaches are.
-There are new proposals and designs for admin features, invitations, challenges & collections, open beta requirements, the front page of the site, etc. that I should finish reading for the committee meeting tomorrow.
-I have some ideas about speeding up the fandoms page - in fact, I wrote all the code, saved it to a flash drive and then... lost the flash drive. I do find that when I have to rewrite code, it tends to come out better the second time around, but it's still a bit inconvenient. Ah well.
-Needs more caching! Although it's handy to do it incrementally so I can see where the pitfalls are with our particular setup.
-I want to write some more tests for work filtering and co-authorship.
Meanwhile, back at the ranch:
More awesome work on training/recruitment and colocation! Colocation is down to the details now - it's very exciting. We're going to be adding some javascript counters to show you how many more characters you can add in form fields with character limits. Lots of ongoing work with the rich text editor and figuring out what html tags to allow where and how to format it all properly. And other assorted awesomeness!
Code Revisions
Committed: 1333
Testing: 1329
Deployed: 1329
We also just opened issue 666. Just for the record, ahem.
Seriously, last weekend was like the attack of Murphy's Law - things were going along pretty swimmingly, and then, boom, about a dozen weird bugs in a row. We did deploy version 1329, though - samvara has release notes! Since then, I worked on a couple of small bugs: I realized caching needed to account for show/hide warnings preferences, the filters weren't generating properly on search, and I helped out with making html visible in summaries again.
Weird issues to resolve:
-I thought the filters were working, but I didn't take into account that people might tag their stories with multiple versions of the same tag (like Alternate Universe, AU) which messes with the results due to reasons redacted for tl;dr. Filter Supernatural fic by the Alternate Universe tag and you'll get some random results mixed in. Need a new approach on that one.
-Pagination goes a bit wacky sometimes - filter Buffy stories for Angel crossovers and the first page will claim to return the number of works tagged with BtVS *or* AtS instead of BtVS *and* AtS. What's weird is that it actually gets the right results, and the second page shows the accurate count. IDK, pagination gem weirdness (count vs. group_by in the automatically generated queries, I think).
-You can't get rid of an OpenID. I fix this, I test it locally, it works, I deploy it, it doesn't work. Small changes don't seem to be cutting it, so I want to really go through the OpenID plugin and gem and figure out how they work and add some code to simplify how we're verifying these and see if there are other issues I can address while I'm in there.
-Translation caching - changes to the default text aren't visible. And I can't figure out how the caching is even set up, and it's causing problems with another plugin that would be handy to use. I'm a little dubious about the plugin we started with, and I'm also trying to think of ways to make the whole process more efficient and easier for the translators. Hopefully some ideas will percolate while I'm working on the other stuff.
-You can use tables in your stories now, but there are some extra spaces before them, and if you import a story with tables from google docs or word, there are a LOT of extra spaces before them. I leave that one to the rich text/regex experts!
-Caching of story blurbs/cast tags has to take into account that cast tags are affected by tag wrangling even if the work itself hasn't been updated. Ie, say you post a Smallville fic with the pairing tag "Clark Kent/Lex Luthor" and the character tags "Clark Kent" and "Lex Luthor". If the tags haven't been sorted by a tag wrangler, cast tags will be "Clark Kent/Lex Luthor, Clark Kent, Lex Luthor". Once the tag wrangler sets up the relationship between the characters and the pairing, the cast tags will condense down to just "Clark Kent/Lex Luthor".
(Am I missing things? I'm probably missing things.)
Other stuff:
-We've been discussing how to improve chapter navigation - we talked about it at the last AD&T meeting, and I want to try to put together a working mock-up to see what people think.
-We've been using/talking about using search for things other than works (users, tags, etc.) and getting some designs for more complex work searches, so I have a note to self to look deeper into the underpinnings of our search plugin (Thinking Sphinx) to get a better idea of what the best approaches are.
-There are new proposals and designs for admin features, invitations, challenges & collections, open beta requirements, the front page of the site, etc. that I should finish reading for the committee meeting tomorrow.
-I have some ideas about speeding up the fandoms page - in fact, I wrote all the code, saved it to a flash drive and then... lost the flash drive. I do find that when I have to rewrite code, it tends to come out better the second time around, but it's still a bit inconvenient. Ah well.
-Needs more caching! Although it's handy to do it incrementally so I can see where the pitfalls are with our particular setup.
-I want to write some more tests for work filtering and co-authorship.
Meanwhile, back at the ranch:
More awesome work on training/recruitment and colocation! Colocation is down to the details now - it's very exciting. We're going to be adding some javascript counters to show you how many more characters you can add in form fields with character limits. Lots of ongoing work with the rich text editor and figuring out what html tags to allow where and how to format it all properly. And other assorted awesomeness!
Code Revisions
Committed: 1333
Testing: 1329
Deployed: 1329
We also just opened issue 666. Just for the record, ahem.
Knocks on wood and crosses fingers and gets out the holy water.
Re: Knocks on wood and crosses fingers and gets out the holy water.
Come over to the dark side! We're always looking for new people, you know. ;)
no subject
*loves on issue 666*
So sorry to hear about the flashdrive :(