The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2016-06-30

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 jberger ah, relative files
00:00 Grinnz_ also, sqlite:///foo is technically relative, but there was a bug in URI::db with that form so I didn't include it
00:00 Grinnz_ you can just use sqlite:foo anyway
00:01 jberger actually it looks like Mojo::URL DTRT when not using //
00:01 jberger perl -Mojo -E 'say Mojo::URL->new("file:my/path")->path'
00:01 Grinnz_ yes, it should
00:02 Grinnz_ but you're assuming already starting with a file: URL
00:02 jberger it does the same thing with 'sqlite:'
00:03 Grinnz_ with sqlite: everything would probably just work with Mojo::URL, but I didn't start with that scheme
00:03 Grinnz_ also I like the abstraction of URI::db
00:04 jberger thats fine
00:04 jberger as I said I was just curious
00:04 jberger happened to notice it
00:04 jberger I have a work side-project that is likely to use SQLite
00:04 jberger so I'm just brushing up
00:04 bpmedley jberger: Have you used SQLite much?
00:04 Grinnz_ https://metacpan.org/source/DBOOK/Mojo-SQLite-0.021/lib/Mojo/SQLite.pm#L72 one nice thing is it lets me canonicalize schemes like sqlite3:
00:04 Grinnz_ automaticlaly
00:05 jberger hmmm, is sqlite3 new/old/default?
00:05 jberger what should I be using?
00:05 jberger bpmedley: not really
00:05 Grinnz_ 3 is the current version
00:05 jberger I usually have used DBM::Deep when I need a file-backed system
00:06 Grinnz_ but I don't know if it's common to use that scheme for URIs
00:06 jberger but this project is likely to need some querying that is more than would be pleasant with DBM::Deep
00:06 bpmedley I think you will love the simple and elegant way it approaches SQL for an embeddable database.
00:06 jberger I'm getting pretty fond of Postgres so I'm worried about missing features
00:06 Grinnz_ SQLite is definitely a new way of thinking in some cases, in that columns are really typeless
00:06 jberger but we'll see
00:06 Grinnz_ or "suggested"
00:07 Grinnz_ the docs are nice in some cases, lacking in others...
00:07 jberger http://howfuckedismydatabase.com/sqlite/
00:07 Grinnz_ heh
00:08 jberger actually it is seeing things like this that scare me (and yes this has made it to the real world, I've seen it): http://stackoverflow.com/a/26241151/468327
00:09 Grinnz_ where mysql tacks on random keywords to add features, sqlite makes abominations of perfectly sensible constructions
00:09 jberger it was funny, I saw that type of query and after figuring out what it did I googled "sqlite generate sequence of numbers" and that is the first hit
00:09 jberger clearly the author did the same thing
00:09 Grinnz_ hehe
00:41 vicash sqlite is great if your data is going to stay < 2GB...
00:41 vicash sqlite is also great if you want to use in-memory databases and use sqlite's :memory: feature
00:42 Grinnz as long as you don't fork :)
00:43 vicash yes of course... i use it mostly in C/C++ world for quick solutions or use in phone apps
00:58 jberger hmmm so I'm reading about sqlite and I see there is a magical rowid column
00:58 vicash it is kind of like the oid column in postgres
00:59 jberger should I use that rather than making my own ids or should I do as normal and create an INTEGER PRIMARY KEY column
00:59 vicash if you want autoincrementing you need to create a primary key i think
00:59 jberger I've never used the oid column in postgres
00:59 jberger ok, best to stick with the standard usage
00:59 vicash yes you never know when you want to migrate the sqlite into a server based db and best to stick to normal practice
01:00 Grinnz In SQLite, a column with type INTEGER PRIMARY KEY is an alias for the ROWID (except in WITHOUT ROWID tables) which is always a 64-bit signed integer.
01:01 Grinnz you can mostly count on that behavior. just don't ever do the "WITHOUT ROWID" thing. that's what I've gathered
01:01 vicash Grinnz is right because somewhere in the docs it says that sqlite uses the rowid/primary key for automatic indexing
01:01 Grinnz http://sqlite.org/autoinc.html
01:01 Grinnz lots more details there
01:11 jberger thanks
01:46 Vitrifur joined #mojo
02:10 noganex_ joined #mojo
02:16 jberger INSERT ... ON CONFLICT in sqlite
02:16 jberger has anyone else tried this?
02:18 jberger oh nevermind, it is spelled INSERT OR IGNORE rather than INSERT ... ON CONFLICT IGNORE even though the feature is called ON CONFLICT
02:18 jberger sigh
02:39 leoj joined #mojo
02:52 Grinnz "or the INSERT and UPDATE commands, the keywords "ON CONFLICT" are replaced by "OR" so that the syntax reads more naturally."
02:52 Grinnz For*
02:52 Grinnz look on the bright side, it's better than mysql's unreadable "INSERT IGNORE"
02:55 nicomen win 31
02:55 nicomen oops
02:59 disputin joined #mojo
03:18 disputin joined #mojo
03:30 ivi joined #mojo
04:18 che-quest joined #mojo
05:19 cpan_mojo Mojo-Weixin-1.1.5 by SJDY https://metacpan.org/release/SJDY/Mojo-Weixin-1.1.5
05:34 cpan_mojo Mojo-Webqq-1.8.2 by SJDY https://metacpan.org/release/SJDY/Mojo-Webqq-1.8.2
05:46 jberger Grinnz: that's true (re mysql) and yes that document about ON CONFLICT is how I figured it out
05:46 jberger But gah! Give **A** usage example here and there and I wouldn't misunderstand simple things like that
05:47 jberger Especially when the grammar diagram is right there at the top (and doesn't include OR)
05:50 Grinnz well, to be fair, the INSERT page has the correct grammar diagram :P
06:02 inokenty-w joined #mojo
06:10 ivi joined #mojo
06:15 ivi joined #mojo
06:51 inokenty joined #mojo
06:58 trone joined #mojo
07:00 dod joined #mojo
07:05 dod joined #mojo
07:06 * sri yawns
07:06 Vandal joined #mojo
07:07 sri apple is taking its time to actually give me the refund :S
07:08 AndrewIsh joined #mojo
07:25 orev joined #mojo
07:51 sri discovered something disappointing about dell too, you can only order windows laptops with the german keyboard from germany
08:11 val40 joined #mojo
08:16 nic I don't know when it started, but we have the equivalent now in the UK
08:17 nic retailers used to be able to do good deals with grey imports, but the keyboard was always US
08:17 nic now those deals are only via shady retailers.  The 'reputable' ones are restricted to UK layout
08:25 osfabibisi joined #mojo
08:27 icjs joined #mojo
08:33 leoj joined #mojo
08:59 kes joined #mojo
10:00 che-quest joined #mojo
10:14 meshl joined #mojo
10:22 leoj joined #mojo
10:23 meshl joined #mojo
10:34 kaare joined #mojo
11:04 tchaves joined #mojo
11:05 tchaves joined #mojo
12:15 val40 joined #mojo
12:22 zivester joined #mojo
12:25 cuechan joined #mojo
12:49 gizmomathboy joined #mojo
13:02 ramortegui joined #mojo
13:09 sri allright, macbook pro ordered, fingers crossed
13:10 sri feels odd to order a broadwell laptop in 2016, but i really don't want another mac with the new keyboard
13:18 bpmedley Did you get the 13"?
13:22 sri yea
13:22 sri the 15" is still haswell :o
13:23 sri curious how apple is pretty much the only one putting 28w cpus in laptops
13:33 cuechan joined #mojo
13:56 ramortegui joined #mojo
13:56 nic If you call $ua->build_tx, do you need to handle redirects yourself?
13:56 sri no
13:57 nic ta.  I'll do some debugging then
13:57 nic I've got max_redirects => 3 so I must be doing something bizarre
14:00 leoj joined #mojo
14:07 nic the test dir didn't have IO::Socket::SSL avail, so was ending on the transition redirect
14:18 mrEriksson joined #mojo
14:20 zivester joined #mojo
14:31 ptolemarch joined #mojo
14:35 disputin joined #mojo
14:45 punter joined #mojo
14:57 asarch joined #mojo
15:09 leoj joined #mojo
15:12 mcsnolte joined #mojo
15:14 lluad joined #mojo
15:26 dotan joined #mojo
15:47 cuechan joined #mojo
15:48 bwf joined #mojo
15:57 disputin joined #mojo
16:06 PryMar56 joined #mojo
16:22 cuechan joined #mojo
17:29 Bean joined #mojo
17:29 jberger sri: re mojo pg #22 iirc the last thing that happened was that you were trying to decide the best course of action
17:30 jberger If closing it means that you don't have any thoughts then can I at already request a clear cache method?
17:31 jberger Grrr s/at already/at least/
17:31 val40 joined #mojo
17:52 dod joined #mojo
17:59 sri jberger: didn't like any of the proposals... and when avkhozov started to propose changes with serious performance penalties i thought it would be best to end it
18:00 sri jberger: as it stands, doubt i would agree with a clear cache method
18:01 sri i'm ok with the status quo
18:01 sri so, any proposal for change would have to be really strong
18:10 sri not even Mojo::UserAgent has a reset cache method
18:20 jberger but Mojo::UserAgent doesn't have any settings that would affect connections in this way
18:21 sri sure it does
18:21 sri proxy
18:21 jberger hmmm, I suppose
18:21 jberger so the advice is "set the search_path immediately, just as you would set a proxy"
18:23 jberger anyway one thing came of this discussion
18:24 jberger the reason I was having troubles is that I was dropping the test schema before setting the search path because I assumed it would yell at you if you tried to drop your current schema
18:24 jberger but it doesn't
18:24 sri of course i'm in a very conservative mood right now... not being able to hack myself ;p
18:24 jberger (like dropping the existing database isn't allowed)
18:24 jberger yeah that sucks
18:24 sri maybe bring it up again in a few weeks
18:24 jberger k
18:25 jberger I'm doing my 50/50% vacation anyway
18:25 jberger I did work full days Mon-Weds
18:25 sri (not saying my opinion will change though)
18:26 jberger today I'm probably only going to do a half day, then long hawaiian weekend
18:26 sri that's a strange vacation
18:26 jberger its a very long one (by US standards)
18:26 jberger I don't have this much vacation time
18:28 itaipu joined #mojo
18:28 sri props to servercentral, few employers are that flexible
18:38 sri the cinematography of the last two game of thrones episodes was really something else
18:38 jberger yeah they were really good
18:39 jberger it was a point of discussion on the WhatTheFlick recap about the style of the opening of the last episode, getting ready for the "trial"
18:40 jberger I'm a big cinematography nut though so I noticed before that
18:44 jberger and yes, I do really like my employer
18:55 sri i mean, that was way ahead of any other tv show
18:57 marty joined #mojo
19:02 dod joined #mojo
19:07 itaipu joined #mojo
19:14 val40 joined #mojo
19:19 marty joined #mojo
19:51 marty joined #mojo
20:47 disputin joined #mojo
21:19 disputin joined #mojo
21:46 cpan_mojo Mojo-JSON-MaybeXS-0.012 by DBOOK https://metacpan.org/release/DBOOK/Mojo-JSON-MaybeXS-0.012
21:57 jberger West Wing had a few nice shots but of course that's a very different type of show
21:57 jberger Mostly for great cinematography you have to go to movies
21:57 jberger But that seems to be coming to the small screen slowly
21:58 jberger There used to be stigma that tv couldn't be high quality
21:58 jberger In all facets
21:58 jberger A list movie stars wouldn't do TV until very recently
22:12 val40 joined #mojo
22:19 meshl joined #mojo
22:44 leoj joined #mojo
22:55 Vitrifur joined #mojo
22:58 sri hmm, the rails 5 announcement on HN is full of "we've moved to elixir and phoenix" :o
23:10 leoj joined #mojo
23:16 meshl joined #mojo
23:24 tchaves joined #mojo
23:45 jberger Sucks to be the old hotness
23:51 zivester joined #mojo

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary