<GitHub46>
[jruby] thedarkone commented on commit 7d82684: @chrisseaton I think this will lead to trouble: yes, GC will collect `LeapRope` string objs from the weak reference, but the key-value mapping (and its associated `Map.Entry`) will stick around, leaking memory.... https://git.io/vVHEa
<GitHub197>
[jruby] chrisseaton commented on commit 7d82684: Yes I suppose it's quite likely someone could run in a loop formatting random integers for a long time.... https://git.io/vVHuZ
AckZ has joined #jruby
<GitHub76>
[jruby] thedarkone commented on commit 7d82684: > Yes I suppose it's quite likely someone could run in a loop formatting random integers for a long time.... https://git.io/vVHaX
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 276 seconds]
<GitHub145>
[jruby] thedarkone commented on commit 7d82684: After some thought, consider e.g. Rails's ORM: there will be a ton of `Product.find(params[:id])` calls, where `params[:id]` is HTTP provided string, that will be parsed into a `Fixnum` (for security purposes) and then subsequently internally used to generate a SQL query: `"SELECT * FROM products WHERE id = #{id}"` ...... https://git.io/vVHV9
<GitHub96>
[jruby] chrisseaton commented on commit 7d82684: Yes I think I'll try the lazy integer node method. I won't leave this cache in place as in as it is not limited by anything. https://git.io/vVHwe
yfeldblum has joined #jruby
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 260 seconds]
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 268 seconds]
adgtl has joined #jruby
adgtl has joined #jruby
adgtl has quit [Changing host]
knu has quit [Read error: Connection reset by peer]
knu has joined #jruby
johnsonch_afk is now known as johnsonch
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 244 seconds]
knu has quit [Read error: Connection reset by peer]
<GitHub61>
jruby/master fbcc60d kares: can safe more allocations on type populate by delaying methods map allocations...
<GitHub61>
jruby/master 4d82ce2 kares: finalize defineAnnotatedConstant on RubyModule as well + avoid an array allocation
<GitHub61>
jruby/master 96a38e8 kares: move the label instances where they are used - only inside default: switch block
Aethenelle has joined #jruby
pawnbox has joined #jruby
<GitHub90>
[jruby] chrisseaton commented on commit 7d4d142: I think this might need to be `$JRUBY_BIN/jruby $JRUBY_BIN/gem` otherwise the `gem` command shell script will try to use a system JRuby. That could be anything on some of our CI systems. https://git.io/vVQJg
raeoks has quit [Ping timeout: 264 seconds]
raeoks has joined #jruby
pawnbox has quit [Remote host closed the connection]
drbobbeaty has quit [Read error: Connection reset by peer]
drbobbeaty has joined #jruby
ebarrett has quit [Quit: WeeChat 1.2]
bbrowning has joined #jruby
tcrawley-away is now known as tcrawley
lance|afk is now known as lanceball
pawnbox has quit [Remote host closed the connection]
vtunka has joined #jruby
Aethenelle has joined #jruby
<GitHub75>
[jruby] eregon commented on commit 7d4d142: Or alternatively `$JRUBY_BIN/jruby -S gem install ...` https://git.io/vV7uT
enebo has joined #jruby
tcrawley is now known as tcrawley-away
pawnbox has joined #jruby
tcrawley-away is now known as tcrawley
<GitHub61>
[jruby] eregon commented on commit 7d82684: One could also locally cache with `@Cached` in Fixnum#to_s. It would probably be interesting to rewrite Integer.toString ourselves to be more PE-friendly. If it turns out multiple integers output a String of the same length, it might allow interesting optimizations further. https://git.io/vV72F
pawnbox has quit [Ping timeout: 276 seconds]
kares has quit [Quit: ...]
enebo has quit [Quit: enebo]
<GitHub115>
[jruby] eregon commented on commit d166ef9: In the worst case, it looks like the reference could be collected between these 2 `get()`. https://git.io/vV7Vx
enebo has joined #jruby
johnsonch_afk is now known as johnsonch
vtunka has quit [Quit: Leaving]
pawnbox has joined #jruby
enebo has quit [Quit: enebo]
kares has joined #jruby
<GitHub186>
[jruby] chrisseaton commented on commit 7d82684: We're flying blind here - I'll produce some benchmarks to show what I'm trying to do. https://git.io/vV76u
<GitHub29>
[jruby] chrisseaton commented on commit d166ef9: Ah yes of course - thanks. https://git.io/vV76o
vtunka has joined #jruby
<GitHub11>
[jruby] eregon commented on commit 8ae215d: String concat without boundary here. https://git.io/vV7MB
<GitHub102>
[jruby] chrisseaton commented on commit 8ae215d: Yes and looks like there are some others - thanks I'll review. https://git.io/vV7MV
bjfish2 has joined #jruby
cprice has joined #jruby
donV has quit [Ping timeout: 244 seconds]
<GitHub99>
[jruby] eregon commented on commit 9fcd6ad: Well done :+1: https://git.io/vV7Se
enebo has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #jruby
skade has quit [Client Quit]
<GitHub96>
[jruby] nirvdrum commented on commit 9fcd6ad: Having comments is great. I'd probably indicate what "size of method" means. Is it node count? Is it byte size? Just so those not well-versed in Truffle can follow better. https://git.io/vV7pg
CaptainHeavy has quit [Remote host closed the connection]
skade has joined #jruby
<GitHub34>
[jruby] chrisseaton created truffle-cexts (+5 new commits): https://git.io/vV5UJ
<GitHub34>
jruby/truffle-cexts ba4a5e7 Chris Seaton: [Truffle] Create jruby-cext-c utility for compiling C extensions using Sulong.
<GitHub34>
jruby/truffle-cexts 5c96568 Chris Seaton: Ignore Sulong libraries (.su)
<GitHub34>
jruby/truffle-cexts d228b4a Chris Seaton: [Truffle] Minimal C extension for testing.
vtunka has quit [Quit: Leaving]
<GitHub144>
[jruby] nirvdrum commented on commit 7e7f844: That's fine. I was a bit lazy originally and didn't enforce only LeafRopes being in the table. This change really shouldn't affect that -- I'd expect the conversion to produce a LeafRope. But if we're going to expand usage we'll want to make sure that we're not caching massive ropes I think. https://git.io/vV5Ip
thedarkone2 has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #jruby
skade has quit [Client Quit]
skade has joined #jruby
<GitHub25>
[jruby] mkristian pushed 1 new commit to jruby-1_7: https://git.io/vV5OI
<GitHub25>
jruby/jruby-1_7 e16e536 Christian Meier: exclude some test on main-extended
lanceball is now known as lance|afk
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #jruby
skade has quit [Client Quit]
brauliobo has quit [Ping timeout: 244 seconds]
shellac has quit [Quit: Ex-Chat]
nirvdrum has quit [Ping timeout: 276 seconds]
nirvdrum has joined #jruby
<GitHub49>
[jruby] mkristian force-pushed jruby-1_7 from e16e536 to aac2d70: https://git.io/Y42-2g
<GitHub49>
jruby/jruby-1_7 aac2d70 Christian Meier: exclude some test on main-extended
<GitHub135>
[jruby] mkristian force-pushed jruby-1_7 from aac2d70 to b2cbfa1: https://git.io/Y42-2g
<GitHub135>
jruby/jruby-1_7 b2cbfa1 Christian Meier: exclude some test on main-extended
<enebo>
mkristian: #3145
<enebo>
mkristian: It is another ubuntu libcrypt problem and I was wondering if you had any time to look at this
<enebo>
mkristian: I do not strictly think we need to fix this for 1.7.25 but if you perceive the fix as simple/non-risky then I am willing to rerelease the jnr planet
<mkristian>
enebo, let me look
<mkristian>
enebo, so the main implementation collects ld paths from /etc/ld.so.conf and /etc/ld.so.conf.d/* and there are no hardcoded paths anymore. it sounds redhat is doing it differently than ubuntu/debian/mint
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
bbrowning is now known as bbrowning_away
<GitHub179>
[jruby] mkristian force-pushed jruby-1_7 from b2cbfa1 to 5693183: https://git.io/Y42-2g
<GitHub179>
jruby/jruby-1_7 5693183 Christian Meier: do not run broken mri tests for embedded jruby...
<enebo>
mkristian: saw your reply here and to atamb0
<enebo>
mkristian: I think we punt this to next point reelase once we get more info on how RHLinux handles this
blandflakes_ has joined #jruby
blandflakes has quit [Disconnected by services]
blandflakes_ is now known as blandflakes
bbrowning_away is now known as bbrowning
shellac has quit [Quit: Computer has gone to sleep.]
<mkristian>
enebo, I will get a cents image and poke around so we have something for 1.7,27 or so
subbu is now known as subbu|away
thedarkone2 has quit [Read error: Connection reset by peer]
skade has joined #jruby
<enebo>
1.7.26
<enebo>
:)
<enebo>
mkristian: thanks this env stuff is almost never ending
<enebo>
mkristian: btw you saw atamb0’s response? Dopes jffi follow includes + globbing?
Prasun has joined #jruby
shellac has joined #jruby
thedarkone2 has joined #jruby
<nirvdrum>
enebo: Any ideas why this would throw an OOM? bin/jruby -J-Xmx16g -w -e 'x = "f" * (2**31 - 1); nil; p x.size'
<enebo>
nirvdrum: not off hand… we are constainted I thought to size of long but otherwise I would not expect that string to be an issue
<nirvdrum>
I think you're constrained to the size of an int.
<nirvdrum>
Bah. It looks like a platform-specific limit. Changing to (2**31 - 3) worked.
<enebo>
nirvdrum: oh for a byte array
<enebo>
nirvdrum: ah yeah true
<enebo>
nirvdrum: nonetheless this should fit
<nirvdrum>
Our ropes are currently limited to int size. I did some work to make them long, but it's a bit ugly.
<enebo>
nirvdrum: but you still need to chain primitives limited at int?
<nirvdrum>
That's part of why it was ugly. I just inserted bounds checks in a bunch of places and raised an exception.
<nirvdrum>
jcodings is pretty tied to 32-bit values, too.
<nirvdrum>
I'd really have to rework things to traverse the entire rope structure and aggregate the results of operations at leaves, which would be limited to 32 bits each.
<enebo>
nirvdrum: does truffle convert regexps via inspect to always be US-ASCII?
<nirvdrum>
Our regexps are an utter mess. We mostly just use your runtime objects still.
<enebo>
nirvdrum: so I just punted an issue another point release but so far it appears inspect will make the inspect value US-aSCII
<enebo>
to be fair not many things probably depend on actual encoding of an inspect’d regexp string but who knows
<mkristian>
but the error is :LoadError: cannot infer basepath
<mkristian>
from file:/Users/cmeier/projects/active/jruby/lib/jruby.jar!/jruby/kernel19/kernel.rb:9:in `require_relative'
<mkristian>
which we had before and was a tricky one
Prasun has quit [Ping timeout: 250 seconds]
<enebo>
mkristian: then do you think we need to fix this for 1.7.25?
<mkristian>
enebo, not for 1.7.25 but we should fix it
<mkristian>
at least on master IMO
Prasun has joined #jruby
<enebo>
mkristian: ok but I am hop[ing this week for 9.1 also
<enebo>
mkristian: but 1.7.25 will be first so the clock is running :)
<mkristian>
I read about your plan somewhere today :)
<enebo>
well I keep saying this week or next week for last three weeks
<enebo>
so I think you read my plan once a week
<enebo>
but 9.1 is pretty sorely needed too since we fixed from AOT bugs and we had some prety wacky perf regressions in 9.0.5
<enebo>
the 2.3 features are nice but those two other things are driving us forward
<enebo>
Oh and also the regexp thread safety explosion
<GitHub44>
[jruby] pitr-ch commented on commit 7d4d142: Thanks, updating. I am not sure about `-S`, it uses `PATH` to look up the script, doesn't it? https://git.io/vV5xu
skade has quit [Quit: Computer has gone to sleep.]
pawnbox has quit [Remote host closed the connection]
shellac has quit [Quit: Computer has gone to sleep.]
<electrical>
Ohw? more towards jruby it self? saw you active in there as well
<mkristian>
actually the current project is moving away from jruby - slow startup, big library, osgi problems
<mkristian>
but on the other hand learning more about jruby with this ;)
<electrical>
hehe okay :-) nice.
<electrical>
I've been messing around with java lately.
<electrical>
learning bits and pieces of it
<electrical>
Now looking for a treetop replacement for java :-)
<mkristian>
I am wrong person to ask here ;)
<GitHub180>
[jruby] nirvdrum pushed 2 new commits to master: https://git.io/vVd38
<GitHub180>
jruby/master 56de066 Kevin Menard: [Truffle] Moved RepeatingRope construction out to nodes....
<GitHub180>
jruby/master 579d882 Kevin Menard: [Truffle] Added maximum length check to rope concatenation.
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
<electrical>
mkristian: hahaha ;-)
<electrical>
fair enough
jeremyevans has quit [Ping timeout: 248 seconds]
yfeldblum has joined #jruby
<GitHub128>
[jruby] eregon commented on commit 7d4d142: That's what MRI's `-h` says but I do not believe it. It can be used to locate gem binaries for instance.... https://git.io/vVdCt
shellac has quit [Quit: Computer has gone to sleep.]
shellac has joined #jruby
shellac has quit [Client Quit]
jeremyevans has joined #jruby
bjfish2 has quit [Quit: bjfish2]
lan5432 has joined #jruby
Balzrael has quit [Remote host closed the connection]
enebo has quit [Quit: enebo]
<lan5432>
kares: didn't you say there was an package refactoring you were gonna cc me in? I'm pretty busy lately with exams, but I may have time to work on some things