| 30 Apr 2018 |
|  charlag changed the history visibility to "world_readable" from "invited". | 12:25:30 |
| charlag | Okay, I will post test message to test history visibility. | 12:26:10 |
| charlag changed the join rule to "invite" from "public". | 12:27:01 |
| charlag changed the join rule to "public" from "invite". | 12:27:12 |
| charlag | Okay, I've been researching how caching works in OkHttp last week.
It's basically a FIFO. It doesn't do anything smart. When it want's to cache a new request it tries to delete old ones one by one until it has enough memory.
As we use the same cache for Picasso image loading and regular requests it doesn't matter much how big our cache is: it will almost always be polluted with unrelated things.
What we could do is use separate cache for images, maybe even custom one. I think that Picasso cache API is not smart enough for this, tho, I'm not sure theres some way to encode priority. Changeing clearing algorithm is simple, saving priority somehow is hard.
It's a super low priority task but it keeps nagging me that our userpics are not cached as good as in other social networks. | 12:32:01 |
|  Tak! joined the room. | 12:33:33 |
|  tibike_m joined the room. | 12:37:09 |
| charlag changed their profile picture. | 12:37:50 |
| Tak! changed their display name from eviltak to Tak!. | 12:42:42 |
| Tak! set a profile picture. | 12:45:37 |
| Tak! | so I started looking at https://github.com/tuskyapp/Tusky/issues/573 this weekend | 12:48:25 |
| Tak! | and I noticed that SpanUtils had a lot of code smells, so I kind of went all-in on refactoring it | 12:49:46 |
|  Constantin A. joined the room. | 12:50:13 |
| Tak! | ...but now I'm not sure that I haven't gone too far - what's the feeling regarding stuff like that? | 12:50:38 |
| charlag | It's very old code - written by Vavassor and never touched since. I believe it can have a lot of smell! But can you be more specific - what troubles you in it?
I am all in for refactoring stuff as long as it doesn't break things. Maybe it should be different PRs?
I see that you're adding tests even, how can I say no to that? :D | 12:55:18 |
| Tak! | Redacted or Malformed Event | 12:55:45 |
| Tak! | tests are still super in-progress 😄 | 12:55:58 |
| charlag | Ofc, I didn't look at them, I just noticed that you intend to add them because I was "I hope it doesn't break things, we need tests for this" (because I tend to break things a lot :c ) | 12:57:04 |
| Tak! | yeah, this is kind of a perfect thing for thorough unit testing (defined inputs and outputs, the whole possibility space can be verified) | 12:58:08 |
| charlag | Not like it's impossible to write most code like that :wink | 13:00:52 |
| charlag | https://github.com/charlag/MovieDB-Knot | 13:00:54 |
| charlag | Omg, I cannot wink properly even online, lol | 13:01:45 |
| Tak! | believe me, my normal codebase is a nightmare to test 😄 | 13:02:40 |
|  tom79 joined the room. | 13:02:59 |
| charlag | I always thought that it's just codebases I have to work with but the older I get the more I believe that it's all codebases. Good code exists mostly in samples. | 13:03:24 |
| tom79 | What's the purpose of this method? https://github.com/Tak/Tusky/blob/29133bf41a5efb958578d89775af73c0246d1c64/app/src/main/java/com/keylesspalace/tusky/ComposeActivity.java#L777 | 13:18:22 |
| charlag | 🤔 | 13:19:01 |
| charlag | I guess some long time ago we were looking at what people were pasting and made actual links but we don't do it anymore? | 13:19:52 |
| tom79 | Characters left should be only 500 - textEditor.getText().length() - contentWarningEditor.getText().length() | 13:20:52 |
| tom79 | I tried to figure out this issue of 23 chars here: https://github.com/tuskyapp/Tusky/issues/573 | 13:21:23 |
