00:00
marciol has joined #jruby
00:05
marciol has quit [Ping timeout: 260 seconds]
00:09
prasun has joined #jruby
00:15
alex0ptr_ has quit [Remote host closed the connection]
00:15
alex0ptr has joined #jruby
00:15
alex0ptr has quit [Remote host closed the connection]
00:16
alex0ptr has joined #jruby
00:20
alex0ptr has quit [Ping timeout: 240 seconds]
00:21
snowp has quit [Ping timeout: 240 seconds]
00:33
snowp has joined #jruby
00:54
hone has quit [Ping timeout: 260 seconds]
00:55
Guest55967 has quit [Remote host closed the connection]
01:16
bbrowning is now known as bbrowning_away
01:45
<
GitHub123 >
jruby/master 46ca986 Charles Oliver Nutter: Use original accessor to access unreified instance vars. Fix #4455...
01:49
marciol has joined #jruby
01:54
marciol has quit [Ping timeout: 240 seconds]
01:56
camlow325 has quit [Quit: WeeChat 1.5]
02:17
enebo has quit [Quit: enebo]
02:22
enebo1 has quit [Ping timeout: 240 seconds]
03:08
prasun has quit [Ping timeout: 260 seconds]
03:38
marciol has joined #jruby
03:43
marciol has quit [Ping timeout: 260 seconds]
04:38
ankitr has joined #jruby
05:27
marciol has joined #jruby
05:32
marciol has quit [Ping timeout: 255 seconds]
05:46
ankitr_ has joined #jruby
05:48
ankitr has quit [Ping timeout: 255 seconds]
06:07
matthewd has quit [Ping timeout: 260 seconds]
06:08
matthewd has joined #jruby
06:21
msg has joined #jruby
06:22
msg is now known as Guest24425
06:24
bbrowning_away has quit [Remote host closed the connection]
06:25
bbrowning_away has joined #jruby
06:27
thedarkone2 has quit [Quit: thedarkone2]
06:36
Specialist has joined #jruby
07:16
marciol has joined #jruby
07:21
marciol has quit [Ping timeout: 260 seconds]
07:45
Specialist has quit [Ping timeout: 240 seconds]
08:30
blaxter has joined #jruby
08:38
Specialist has joined #jruby
09:05
marciol has joined #jruby
09:10
marciol has quit [Ping timeout: 255 seconds]
09:10
blaxter has quit [Quit: foo]
09:51
vtunka has joined #jruby
10:02
drbobbeaty has joined #jruby
10:02
shellac_ has joined #jruby
10:19
Specialist has quit [Ping timeout: 260 seconds]
10:23
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
10:37
vtunka has quit [Quit: Leaving]
10:41
shellac has quit [Remote host closed the connection]
10:42
shellac_ is now known as shellac
10:54
marciol has joined #jruby
10:58
marciol has quit [Ping timeout: 255 seconds]
11:19
Specialist has joined #jruby
11:21
neoice has quit [Ping timeout: 268 seconds]
11:21
neoice has joined #jruby
11:29
ankitr_ has quit [Ping timeout: 260 seconds]
11:53
Specialist has quit [Ping timeout: 240 seconds]
12:03
bbrowning_away is now known as bbrownning
12:03
bbrownning is now known as bbrowning
12:16
drbobbeaty has joined #jruby
12:17
Specialist has joined #jruby
12:42
marciol has joined #jruby
12:43
shellac has quit [Quit: Computer has gone to sleep.]
12:47
marciol has quit [Ping timeout: 240 seconds]
12:51
vtunka has joined #jruby
13:23
prasun has joined #jruby
13:26
alex0ptr has joined #jruby
13:58
prasun has quit [Ping timeout: 260 seconds]
13:59
prasun has joined #jruby
13:59
shellac has joined #jruby
14:00
tcrawley-away is now known as tcrawley
14:02
Specialist has quit [Ping timeout: 268 seconds]
14:03
ankitr_ has joined #jruby
14:05
Specialist has joined #jruby
14:12
marciol has joined #jruby
14:16
ankitr_ has quit [Ping timeout: 260 seconds]
14:16
Specialist has quit [Ping timeout: 240 seconds]
14:17
Specialist has joined #jruby
14:21
ankitr has joined #jruby
14:24
prasun has quit [Ping timeout: 268 seconds]
14:26
ankitr has quit [Ping timeout: 260 seconds]
14:50
enebo has joined #jruby
14:54
enebo_ has joined #jruby
14:59
ankitr has joined #jruby
15:06
Specialist has quit [Ping timeout: 260 seconds]
15:07
cremes has quit [Ping timeout: 268 seconds]
15:08
cremes has joined #jruby
15:17
marciol has quit [Remote host closed the connection]
15:17
marciol has joined #jruby
15:21
alex0ptr has quit [Remote host closed the connection]
15:22
marciol has quit [Ping timeout: 240 seconds]
15:25
alex0ptr has joined #jruby
15:28
prasun has joined #jruby
15:38
<
GitHub166 >
[jruby] headius closed issue #4492: YAML: Can't load files with characters in the supplementary plane
https://git.io/vD12d
15:40
<
GitHub156 >
[jruby] headius reopened issue #4492: YAML: Can't load files with characters in the supplementary plane
https://git.io/vD12d
15:48
<
GitHub32 >
jruby/master b0487b4 Charles Oliver Nutter: Bump psych version to 2.2.3. Fixes #4492.
15:48
<
GitHub5 >
[jruby] headius closed issue #4492: YAML: Can't load files with characters in the supplementary plane
https://git.io/vD12d
15:55
enebo_ has quit [Quit: enebo_]
15:56
bbrowning is now known as bbrowning_away
16:02
camlow325 has joined #jruby
16:06
Specialist has joined #jruby
16:08
<
headius >
nirvdrum: I'm find with fixes to jcodings that help AOT but not if they eager-load things we lazily load right now
16:08
<
headius >
there's way too much data in there to eagerly load everything
16:14
alex0ptr has quit [Ping timeout: 240 seconds]
16:17
<
lopex >
headius: also bad for ruboto
16:18
tcrawley is now known as tcrawley-away
16:18
<
lopex >
headius: we could generate a switch and fallback to newInstance maybe ?
16:19
<
headius >
I don't see how that would help them wanting to statically compile everything
16:19
Specialist has quit [Ping timeout: 255 seconds]
16:19
Specialist_ has joined #jruby
16:19
<
enebo >
hmmm we need macros
16:19
<
lopex >
headius: how switch doesnt help ?
16:19
<
enebo >
we could have an eager main library entry point and a lazy one and you pick eager for AOT
16:20
<
headius >
lopex: I don't know what you mean by switch
16:20
<
headius >
switch/case?
16:20
<
enebo >
like in AOT version you call new JoniForceLoad()
16:20
<
lopex >
headius: case "foo"; return SomeEncoding.INSTANCE
16:20
<
headius >
ah, sure, it might work ok
16:21
<
headius >
or enebo's idea
16:21
<
lopex >
all the tables ?
16:21
<
lopex >
but that might be too much
16:21
<
headius >
so long as it doesn't cause all encoding to load at boot I don't care
16:21
tcrawley-away is now known as tcrawley
16:22
<
lopex >
and same problem with transcoders
16:22
<
lopex >
well, even bigger
16:23
<
lopex >
that's why Kevin alluded to lazy load the tables though
16:23
<
enebo >
lopex: you see the can joni parse faster issue
16:23
<
enebo >
lopex: Although he did not mean parse he meant execute the regexp
16:24
<
enebo >
lopex: yeah I think in last 12 hours
16:25
<
lopex >
yeah I see it
16:25
<
chrisseaton >
headius: how can there be too much data to eagerly load if you are doing AOT?
16:25
<
enebo >
chrisseaton: he means for us not doing AOT to share same code as you doing AOT
16:25
<
chrisseaton >
headius: I mean, what's the downside of having it in the AOT image?
16:25
<
chrisseaton >
Oh right, I thought you were trying it on Java 9 AOT
16:25
<
enebo >
at least that was my take
16:25
<
headius >
chrisseaton: see above, nirvdrum was asking about changes to make it AOT better in SVM
16:26
<
headius >
that's fine as long as it doesn't cause non-AOT case to eagerly load everything
16:26
<
enebo >
does Java 9 AOT allow classloading?
16:27
<
enebo >
or lazy but we instantly use them
16:27
<
enebo >
which is eager already
16:27
<
headius >
Java 9 AOT just boots your app immediately into C1-like optimizations
16:27
<
enebo >
so then we probably don't care past UTF-8 and ASCII-7BIT
16:27
<
lopex >
chrisseaton: newInstance is also a problem right ?
16:30
<
enebo >
chrisseaton: is any ruby manager releasing ruby+truffle builds yet?
16:30
<
chrisseaton >
lopex: are we talking SVM of J9 AOT? I know nothing about the latter
16:30
<
lopex >
chrisseaton: SVM
16:30
<
chrisseaton >
enebo: no, we're waiting on some internal discussions about how to release GraalVM without a clickthrough since that blocks it really
16:30
alex0ptr has joined #jruby
16:31
<
chrisseaton >
lopex: you can't newInstance in SVM
16:31
<
enebo >
chrisseaton: thanks
16:31
<
chrisseaton >
lopex: you will need our JVM bytecode interpreter for that, which is just a toy at the moment
16:31
akp has joined #jruby
16:31
<
lopex >
chrisseaton: so generated switch would suffice for now ?
16:31
<
chrisseaton >
lopex: yes
16:31
<
chrisseaton >
Or a lambda factory method kind of thing
16:34
marciol has joined #jruby
16:36
alex0ptr has quit [Read error: Connection reset by peer]
16:36
alex0ptr_ has joined #jruby
16:36
vtunka has quit [Quit: Leaving]
16:37
<
lopex >
enebo: but they seem to go through j.l.String, they could use Files.readAllBytes and scan for new lines at very least
16:38
prasun has quit [Ping timeout: 268 seconds]
16:40
<
enebo >
lopex: leave a comment please
16:40
<
lopex >
enebo: looking for a better idea now though
16:40
<
enebo >
lopex: I don't think that Java's regexp is good at tons of things so if they happen to be faster on some I hardly consider it a bug
16:41
<
enebo >
lopex: although perhaps it is
16:41
<
lopex >
enebo: they shouldnt
16:41
<
enebo >
lopex: I keep thinking of the Java devs vs PCRE devs when that engine came out
16:41
<
lopex >
enebo: i.e. j.u.Regexp
16:41
<
enebo >
lopex: shouldn't what? be faster at anything?
16:42
<
lopex >
enebo: we are likely slower at regexp construction though
16:42
<
enebo >
lopex: that was not an answer to my question was it?
16:42
<
lopex >
enebo: yeah, we should be faster at matching in general
16:42
<
enebo >
lopex: ah ok so another comment would be to precompile the regexp and remove it from the loop
16:43
<
enebo >
oh they did
16:43
<
lopex >
not a problem
16:43
<
enebo >
lopex: yeah so they get bytes in the loop and transcode them
16:43
<
lopex >
problem is getBytes as ususal
16:44
thedarkone2 has joined #jruby
16:44
<
lopex >
enebo: so we get extra bonus by not transocing
16:44
<
lopex >
*transcoding
16:45
<
enebo >
lopex: yeah so their LineDecoder could work with bytes directly
16:46
<
lopex >
yeah, but then new lines might become problematic
16:46
<
enebo >
lopex: which would not only kill our UTF16-LE -> UTF-8 to make bytes but also kill the original text -> UTF16-LE
16:46
<
lopex >
you'd have to skip char heads at very least
16:46
<
lopex >
well, you can use jcodings for that
16:47
<
lopex >
enebo: or just swotch regexp to multiline ?
16:47
<
enebo >
lopex: but you mean you would have to mbclen over and over (or whatever we call it in jcodings) to walk to the newline
16:47
<
enebo >
lopex: multiline could work depending on the regexp
16:50
<
lopex >
enebo: and then joni wouldnt copy anything (even StringNode buffers are shared - unless not escaped)
16:50
<
lopex >
in the regexp itself
16:50
<
lopex >
er, unless escaped
16:50
<
enebo >
lopex: yeah so only cost is walking the bytes in Java
16:50
<
lopex >
enebo: help ruby, since lots of regexps are fed from arbitrary strings
16:50
<
enebo >
but that C++ is also making it into UTF16-LE
16:51
<
enebo >
so we may even come out ahead depending on how well the newline walking code is optimized
16:51
<
enebo >
lopex: oh yeah once Java makes the String the \n finding bit is in Java too
16:52
<
enebo >
lopex: but walking the mbc and transcoding to UTF16LE is in C++ in Java right?
16:52
<
lopex >
enebo: you mean native ?
16:52
<
enebo >
lopex: are you hung up that it might be in C?
16:53
<
lopex >
enebo: looking
16:53
<
enebo >
lopex: haha
16:53
<
enebo >
lopex: it hardly matters...I should have just said native as I just mean some compiler chomped into what it is
16:53
<
lopex >
enebo: I'm more concerned with laziness though
16:53
<
enebo >
So one native loop walking bytes to make characters but then a java loop looking for \n character
16:54
<
enebo >
It has to do that much work to pass it to Java.lang.Regexp for matching
16:54
ankitr has quit [Ping timeout: 240 seconds]
16:54
<
enebo >
The reporter can make something which reads bytes and walks once (in Java) and return byte[] for a line
16:55
<
enebo >
which one is faster? I have no idea but it seems possible we might win then
16:55
<
enebo >
in any case it is much less work than what the reporter is doing now
16:58
<
lopex >
that's the idea
17:00
<
enebo >
lopex: I can comment to this on the issue if you want
17:00
<
lopex >
enebo: readLine() uses int read() anyways
17:01
<
enebo >
lopex: but from a stream of a reader?
17:01
<
lopex >
enebo: looking for any alternative
17:02
<
enebo >
lopex I think read(BUFSIZE) and maintaining start + length is what I would do
17:03
<
enebo >
but hey working with byte[] is more complicated than String
17:03
<
enebo >
which may involve some amount of System.arraycopy to move end bytes onto front when buffer is exhausted
17:03
cremes has quit [Quit: cremes]
17:03
<
lopex >
enebo: oh, wait BufferedReader reads directly from it's buffer
17:09
<
GitHub138 >
jruby/master a2cdff8 Charles Oliver Nutter: Do not precompile target script with BEGIN or END. For #713.
17:20
AckZ has joined #jruby
17:23
cremes has joined #jruby
17:26
bbrowning_away is now known as bbrowning
17:29
hone has joined #jruby
17:30
cremes has quit [Quit: cremes]
17:32
shellac has quit [Ping timeout: 240 seconds]
17:32
yosafbridge has quit [Quit: Leaving]
17:32
cremes has joined #jruby
17:36
yosafbridge has joined #jruby
17:42
alex0ptr_ has quit [Remote host closed the connection]
17:53
Guest24425 has quit [Remote host closed the connection]
18:00
camlow325 has quit [Quit: WeeChat 1.5]
18:20
thedarkone2 has quit [Quit: thedarkone2]
18:27
Specialist_ has quit [Ping timeout: 255 seconds]
18:27
Specialist has joined #jruby
18:34
camlow325 has joined #jruby
18:45
enebo_ has joined #jruby
18:48
Specialist has quit [Remote host closed the connection]
18:48
msg has joined #jruby
18:48
msg is now known as Guest5443
18:53
tcrawley is now known as tcrawley-away
18:55
tcrawley-away is now known as tcrawley
19:36
thedarkone2 has joined #jruby
19:36
<
GitHub120 >
[jruby] nerdrew opened issue #4500: IO#close_write seems to close the read side too for unix sockets
https://git.io/vyvol
19:42
Guest5443 has quit [Remote host closed the connection]
20:22
<
GitHub134 >
jruby/master 0424413 Thomas E. Enebo: Fixes ruby/spec failure for File#join where we would process windows delimited path on non-windows platforms
20:22
<
GitHub134 >
jruby/master e2d8cdf Thomas E. Enebo: Fixes several ruby/spec failures involving File#join
20:30
msg has joined #jruby
20:31
msg is now known as Guest14186
20:46
bbrowning is now known as bbrowning_away
21:09
_whitelogger_ has joined #jruby
21:10
_whitelogger has quit [Remote host closed the connection]
21:25
lampad has joined #jruby
21:27
<
lampad >
does anyone have a good recommendation for how to debug the java side of things in an IDE? The wiki and stuff I found online seem to focus on the ruby level
21:28
<
lampad >
Another thing that might complicate it is that I'm working in jruby-openssl and would like to debug something in that jar (in my IDE), but i'm executing a ruby test using a jruby provided by rvm after compiling and packaging jruby-openssl
21:30
<
lampad >
if any of that makes sense...
21:30
marciol has quit [Remote host closed the connection]
21:32
bbrowning_away is now known as bbrowning
21:34
marciol has joined #jruby
21:39
<
GitHub114 >
[jruby] whwilder opened pull request #4502: Implemented optional buffer argument for Array#pack (ruby-2.4...ruby-2.4-array-pack)
https://git.io/vyvN9
21:52
alex0ptr has joined #jruby
21:57
alex0ptr has quit [Ping timeout: 260 seconds]
22:01
bbrowning is now known as bbrowning_away
22:03
akp_ has joined #jruby
22:04
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
22:07
akp_ has quit [Ping timeout: 240 seconds]
22:07
akp has quit [Ping timeout: 268 seconds]
22:10
akp has joined #jruby
22:14
cremes has quit [Quit: cremes]
22:14
akp has quit [Ping timeout: 260 seconds]
22:29
tcrawley is now known as tcrawley-away
22:31
marciol has quit [Remote host closed the connection]
22:31
<
chrisseaton >
lampad: are you familiar with conventional JVM remote debugging?
22:32
marciol has joined #jruby
22:43
<
lampad >
chrisseaton: yeah, just do a normal java invocation passing the jruby jar, debugging opts and additional ruby-level args that way?
22:45
<
GitHub91 >
jruby/master eab2c92 Thomas E. Enebo: Fixes #4501. JRuby native no longer loads on windows. Well it does but we had no gettimeofday impl on windows. Working around it for now. Ultimate fix is in jnr-posix
22:54
hone has quit [Ping timeout: 260 seconds]
22:56
cremes has joined #jruby
23:04
<
chrisseaton >
(replied in DM)
23:04
<
GitHub165 >
jruby/master c1b5f15 Thomas E. Enebo: Fixes failures from ruby/spec for File#join where leading delimiters using alt delim (\) will now be seen
23:05
alex0ptr has joined #jruby
23:09
marciol has quit [Remote host closed the connection]
23:18
enebo_ has quit [Quit: enebo_]
23:18
enebo has quit [Quit: Leaving.]
23:25
cremes has quit [Quit: cremes]
23:26
cremes has joined #jruby
23:30
Guest14186 has quit [Remote host closed the connection]
23:40
msg has joined #jruby
23:41
msg is now known as Guest41479
23:45
Guest41479 has quit [Ping timeout: 255 seconds]