| Time |
Nickname |
Message |
|
00:58
🔗
|
GitHub114 |
13terroroftinytown/06master 140865e89 15Christopher Foo: Bump client min version to 4... |
|
00:58
🔗
|
GitHub114 |
13terroroftinytown/06master 14dc7feb8 15Christopher Foo: Use unicode-escape encoding for JSON on the done API... |
|
00:58
🔗
|
GitHub114 |
13terroroftinytown/06master 14ff2c42e 15Christopher Foo: services: Results must return body encoding or None.... |
|
00:58
🔗
|
GitHub114 |
[13terroroftinytown] 15chfoo pushed 3 new commits to 06master: 02 https://github.com/ArchiveTeam/terroroftinytown/compare/a6cc2d550efa...0865e891f512 |
|
01:02
🔗
|
GitHub114 |
[13terroroftinytown-client-grab] 15chfoo pushed 1 new commit to 06master: 02 https://github.com/ArchiveTeam/terroroftinytown-client-grab/commit/1adc311feb9b30c4cac6173ecdbe31f4d0571a45 |
|
01:02
🔗
|
GitHub114 |
13terroroftinytown-client-grab/06master 141adc311 15Christopher Foo: Bump submodule. |
|
01:06
🔗
|
chfoo |
i reenabled the 3 shortener queues |
|
01:12
🔗
|
aaaaaaaaa |
I have a suggestion, can you increase the time out on updatesubmodule for the tasks that aren't doing it from 10 seconds? |
|
01:12
🔗
|
aaaaaaaaa |
It looks like it is erroring out when new items pop up every 10 seconds. |
|
01:13
🔗
|
chfoo |
it's supposed to only error every hour |
|
01:15
🔗
|
aaaaaaaaa |
I mean when I click on "work on this project" it pops up two items. The one does the updatesubmodule and the other dies and then sleeps for 10 seconds. Then a new item pops up and does the same until the other item finishes updating. |
|
01:15
🔗
|
chfoo |
oh, ok |
|
01:15
🔗
|
aaaaaaaaa |
looks too much like it is erroring out and broken |
|
01:16
🔗
|
chfoo |
the problem is that the version of git in the warrior vm is really old so it doesn't support updating to the latest |
|
01:17
🔗
|
chfoo |
so i'm trying the new command, and then trying the old command |
|
01:17
🔗
|
aaaaaaaaa |
could you make the other item sleep longer? |
|
01:18
🔗
|
aaaaaaaaa |
that way there isn't a build up of red "FAILED" |
|
01:18
🔗
|
chfoo |
i'm not sure how. seesaw wasn't really meant for updating submodules though |
|
01:19
🔗
|
aaaaaaaaa |
Thats ok, just a suggestion. |
|
01:20
🔗
|
chfoo |
it's something that needs to be fix though |
|
01:21
🔗
|
chfoo |
i realized that i forgot to change the nickname in my warrior |
|
01:22
🔗
|
aaaaaaaaa |
easy to spot |
|
01:27
🔗
|
chfoo |
i keep confusing which warrior vm is being used for dev or for real :) |
|
01:28
🔗
|
aaaaaaaaa |
I organized mine into groups. The blue border around it helps. |
|
01:33
🔗
|
aaaaaaaaa |
https://github.com/ArchiveTeam/terroroftinytown-client-grab/blob/master/pipeline.py line 83 |
|
01:34
🔗
|
aaaaaaaaa |
module is spelled "moule" |
|
01:39
🔗
|
chfoo |
thanks, will fix |
|
01:40
🔗
|
chfoo |
oh, there's a retry delay. that should reduce the time it errors |
|
01:44
🔗
|
GitHub169 |
13terroroftinytown-client-grab/06master 1419d9e71 15Christopher Foo: UpdateSubmodule: Fix typo, decrease retry delay, increase max tries. Bump ver. |
|
01:44
🔗
|
GitHub169 |
[13terroroftinytown-client-grab] 15chfoo pushed 1 new commit to 06master: 02 https://github.com/ArchiveTeam/terroroftinytown-client-grab/commit/19d9e71c814c9de0d666be9a09c7d6c57ecc17c7 |
|
01:46
🔗
|
chfoo |
the tracker isn't assigning shorteners randomly. anyone know of a good way? |
|
01:49
🔗
|
aaaaaaaaa |
much better |
|
01:49
🔗
|
aaaaaaaaa |
is the tracker terror of tinytown without the -grab? |
|
01:50
🔗
|
chfoo |
yes, https://github.com/ArchiveTeam/terroroftinytown |
|
01:56
🔗
|
chfoo |
maybe order by random since i don't envision having thousands of shorteners in the tracker |
|
02:04
🔗
|
GitHub115 |
13terroroftinytown/06master 14eff0f9e 15Christopher Foo: tracker: Security: Add @authenticated to delete project. |
|
02:04
🔗
|
GitHub115 |
[13terroroftinytown] 15chfoo pushed 1 new commit to 06master: 02 https://github.com/ArchiveTeam/terroroftinytown/commit/eff0f9e9e1ded2e6f112ab767af44a75bf96aa4c |
|
02:10
🔗
|
GitHub26 |
13terroroftinytown/06master 1469353fa 15Christopher Foo: tracker: Use order_by random() to assign random projects.... |
|
02:10
🔗
|
GitHub26 |
[13terroroftinytown] 15chfoo pushed 1 new commit to 06master: 02 https://github.com/ArchiveTeam/terroroftinytown/commit/69353fa7a3c02e142ea75e40c09ab3d9c1ccc080 |
|
15:24
🔗
|
aaaaaaaaa |
I see vantec found this project. |
|
16:35
🔗
|
vantec |
shhhhh aaaaaaaaa |
|
16:36
🔗
|
vantec |
sweet, looks like I can do 1M urls a day |
|
17:13
🔗
|
chfoo |
anyone know a good algorithm to turn a number into a randomly looking shortcode? it has to be one-to-one mapping. it seems like using using bytes from md5 digest doesn't seem idea because we'll never know when it cycles. |
|
17:18
🔗
|
aaaaaaaaa |
base 36 conversion or not random looking enough? |
|
17:19
🔗
|
aaaaaaaaa |
or heck a different, less used, base would look odder. |
|
17:23
🔗
|
aaaaaaaaa |
A coworker suggested using some simple cipher. |
|
17:42
🔗
|
aaaaaaaaa |
heck, maybe just xor it with a big enough key |
|
17:45
🔗
|
Smiley |
rot13 ? |
|
18:21
🔗
|
chfoo |
a combination of those might work. the reason is that the auto queue works only by incrementing a integer and it doesn't work well for those shorteners with random, fixed-length shortcodes |
|
18:23
🔗
|
chfoo |
so i figured if there was a function that converted the auto queue integer into a random number, it would give a better distribution of random chance of hitting a url |
|
18:23
🔗
|
chfoo |
rather than starting from 000000 and slowly going to ZZZZZ for example |
|
18:24
🔗
|
xmc |
gray code |