00:00:06<@JAA>I.e. it isn't.
00:00:51dm4v quits [Read error: Connection reset by peer]
00:01:14<pabs>aha
00:04:39dm4v joins
00:04:41dm4v quits [Changing host]
00:04:41dm4v (dm4v) joins
00:16:45dfgffgd joins
00:37:16BlueMaxima joins
01:02:41dm4v quits [Read error: Connection reset by peer]
01:03:34dm4v joins
01:03:36dm4v quits [Changing host]
01:03:36dm4v (dm4v) joins
01:22:56duce1337 quits [Client Quit]
02:40:20duce1337 joins
03:28:41qw3rty_ joins
03:32:19qw3rty__ quits [Ping timeout: 258 seconds]
05:29:20<systwi>JAA: I saw you earlier mention using zstd before. How would you say it compares to LZMA4? I've always gone with 7-Zip's highest compression using LZMA4 and, in general, the files have tended to be the smallest. I know it ultimately depends on the input complexity, but is zstd generally better?
05:30:12<systwi>I personally care more about a smaller output rather than the time it takes to compress/decompress.
05:55:17<pabs>how deep does ArchiveBot go when downloading a site? does it also archive pages on external sites linked from the site?
06:01:08<systwi>pabs: It will go one level outside the hostname specified. E.g., if you run "!a http://example.com/" and it has a link to "https://eff.org/", it will grab the latter, too, but only that page and its assets. It won't start grabbing all of eff.org, too.
06:23:57grawity quits [Remote host closed the connection]
06:24:56grawity (grawity) joins
06:25:36<pabs>and it won't grab pages that are linked to on the second domain?
07:21:40<systwi>No, it will skip them.
07:23:43<pabs>ok, thanks
07:24:12<systwi>You're welcome
07:26:32Matthww86 joins
07:27:32Matthww8 quits [Ping timeout: 250 seconds]
07:27:32Matthww86 is now known as Matthww8
07:38:22Matthww8 quits [Ping timeout: 250 seconds]
07:43:24BlueMaxima quits [Client Quit]
07:52:27Matthww8 joins
08:09:56godane (godane) joins
09:03:44HP_Archivist quits [Ping timeout: 250 seconds]
10:00:31duce1337 quits [Changing host]
10:00:31duce1337 (duce1337) joins
10:20:27spirit joins
12:03:13yano quits [Read error: Connection reset by peer]
12:03:57yano (yano) joins
12:04:00psy quits [Ping timeout: 250 seconds]
12:04:15noteness quits [Read error: Connection reset by peer]
12:04:32noteness (noteness) joins
12:06:44psy (psy) joins
12:44:31Matthww88 joins
12:45:05bsmith093 quits [Ping timeout: 258 seconds]
12:45:28Matthww8 quits [Ping timeout: 258 seconds]
12:45:28Matthww88 is now known as Matthww8
12:45:34bsmith093 joins
13:00:59Iki joins
13:33:13dfgffgd leaves
13:38:02Matthww8 quits [Ping timeout: 250 seconds]
13:42:06Matthww8 joins
13:54:56lunik1 quits [Quit: :x]
13:55:27lunik1 joins
13:59:18lunik1 quits [Client Quit]
13:59:44lunik1 joins
14:21:53lunik1 quits [Client Quit]
14:22:52lunik1 joins
14:24:22Stiletto quits [Ping timeout: 258 seconds]
15:17:39Arcorann_ quits [Ping timeout: 258 seconds]
16:02:53thuban quits [Ping timeout: 258 seconds]
16:19:00<@JAA>systwi: I haven't experimented much with LZMA, but I believe it generally (with the right settings) achieves a higher compression ratio with a much longer runtime than zstd. But ultimately, you'll have to try it with the specific data you want to compress.
17:06:24<spirit>generally as a rule of thumb: zstd is like zip but usually way faster, lzma is a bit advanced compression ratio wise but super slow
17:06:32<spirit>zstd allows training with a dictionary though iirc
17:07:31<spirit>in other news, why the flying fuck are raspberry 4s still murdering SD cards :(
17:10:26<@JAA>Uh, no, zstd isn't like ZIP at all. It's purely a compression algorithm and format like Brotli, Deflate, LZ4, LZMA, and what have you. ZIP is an archive file format and more comparable to e.g. tar, but it has direct support for compression with a number of different algorithms.
17:10:47<spirit>sorry i meant it is pretty close to DEFLATE afaik
17:10:58<@JAA>Yeah, that's the most common one. ZIP also supports LZMA though.
17:11:23<spirit>huh
17:12:48<@JAA>But comparing zstd to ZIP with LZMA is still not fair. You'd have to put the zstd stream(s) into a ZIP (without another compression layer, obviously) for that.
17:14:45<@JAA>Also, Lempel-Ziv and all its derived algorithms (including LZMA and zstd) are all dictionary-based as far as I know. Not sure if any common implementations of the others allow you to train and specify a custom dict though.
17:15:10Stiletto joins
17:15:48<@JAA>But you can use a custom dict with zlib in Python, for example.
17:16:13spirit backs away into a hedge and pretends he never said anything
17:17:57<@JAA>I mean, your rule of thumb is a good one with s/zip/deflate/. :-)
18:22:36DogsRNice (Webuser299) joins
18:24:55HP_Archivist (HP_Archivist) joins
18:25:27<lunik1>LZMA is slower to compress and decompress, but you can generally get slightly better ratios out of it
18:26:49<lunik1>I don't think there's much point in using zip or bzip2 over zstd at this point, other than compatibility
18:29:06<lunik1>*using gzip
19:47:24thuban joins
20:07:18spirit quits [Client Quit]
21:04:22Barto quits [Ping timeout: 250 seconds]
21:04:43Barto (Barto) joins
23:05:08Arcorann_ joins
23:17:24HP_Archivist quits [Ping timeout: 250 seconds]