<rdubya>
Does anybody know if this should be considered a valid UUID of some sort? "a0ee-bc99-9c0b-4ef8-bb6d-6bb9-bd38-0a11"
<rdubya>
there's an AR test using it to test a UUID column, but Java's UUID class doesn't like it and postgres bombs because it won't compare a string to a uuid column :(
<rdubya>
kares: enebo: is this a format we should try to detect or should I just exclude the test?
<rdubya>
apparently its considered a format that should be supported since there is a test explicitly testing that that format is supported, so I guess I'll see if I can do anything with it
lance|afk is now known as lanceball
<kares>
rdubya: that seems to be not valid - from first look (too many sections)
<kares>
but there are likely various uuid formats - so not sure what to do there
<kares>
am not against if you just exclude and open an issue for it
<kares>
we already might have issues with the extensive use of PG JDBC API
<kares>
e.g. I noticed I could not degrade to a previous version - since I wasn't able to compile
<kares>
not to be worried about right now but if it means more hacks around its API maybe delay ... for now
<rdubya>
it looks like it probably won't be too hard to support, just means more string processing for those :(
jrafanie has joined #jruby
jrafanie_ has joined #jruby
jrafanie has quit [Ping timeout: 260 seconds]
shellac has quit [Ping timeout: 240 seconds]
xardion has quit [Remote host closed the connection]
xardion has joined #jruby
elia has quit [Quit: Computer has gone to sleep.]
elia has joined #jruby
mistergibson has quit [Remote host closed the connection]
<Cu5tosLimen>
hi
<Cu5tosLimen>
is there channel for renjin?
claudiuinberlin has joined #jruby
NightMonkey has joined #jruby
<NightMonkey>
Howdy. Is there a doc that discusses (or blog post) about which Jruby releases are considered "stable" or are "supported"? Is 9.0.5.0 still considered "supported" by the jRuby community? :)
<NightMonkey>
Or, is only the 9.1.X.X tree that can be considered "supported" or "maintained"?
<lopex>
NightMonkey: all are considered stable
<lopex>
pick the latest one
nirvdrum has quit [Ping timeout: 240 seconds]
nirvdrum has joined #jruby
<NightMonkey>
Even 9.0.5.0, last touched in 1/2016? :)
<NightMonkey>
lopex: ^
<lopex>
NightMonkey: it's a release, why would it be touched since ?
<lopex>
NightMonkey: pick the latest one
<lopex>
9.1.16.0 as of now
<lopex>
enebo, headius: is there anything about versioning scheme for jruby anywhere on wiki ?
<enebo>
lopex: well we have only two versions listed on our website
<enebo>
lopex: but I don't think we have documented the versioning changing logic on the wiki at all
<lopex>
seems there's some confusion
<NightMonkey>
lopex: My concern is that any stability or security updates made to the MRI 2.2 series is also in jRuby 9.0.5.0, which is impossible if it hasn't been updated since 1/2016. :)
<enebo>
I guess we should be a bit more explicit about being end of life on some lines
<NightMonkey>
lopex: (and, of course MRI Ruby 2.2 is EOL'd soon.
<lopex>
NightMonkey: if you're not restricted by java compat then the simple answer is pick the latest one
<NightMonkey>
lopex: Ah! There's a JVM compatibility restriction?
<lopex>
as in any java project
<enebo>
The changes between 2.2 and 2.3 were so benign we said 9.0.x was dead and all should move to 9.1.x
<enebo>
lopex: NightMonkey: Java compat was Java 7 for both so that is not a change
<NightMonkey>
enebo: Ah, thank you! :D
<NightMonkey>
enebo: Might you point me to a public link where the team said that? :D Is it on a mailing list?
<enebo>
NightMonkey: our 2.5 support (skipping 2.4) will be out in a month or less but we will be supporting 9.1.x for some time yet
<lopex>
enebo: oh so I think I'll manage to land the case mapping
<NightMonkey>
enebo: Awesome.
<enebo>
NightMonkey: said that we had EOL'd 9.0.x? I think largely we have said it at talks but we should probably make a table on the wiki to help point people to these decisions. This is a failing on our part.
<enebo>
NightMonkey: but for your use I would not be using 9.0.x even if we were doing security patches since it has a number of serious memory issues.
<NightMonkey>
enebo: Thank you for sharing this.
<enebo>
NightMonkey: which may not be an issue...it just depends. We have also made 9.1.x a lot faster
<enebo>
9.0.x was part of a brand new architecture and we still had quite a few rough edges
<enebo>
lopex: cool
<lopex>
enebo: btw can you release jcodings ?
<enebo>
lopex: sure.
<lopex>
enebo: it has bumped to java to 7 and has that non reflective loading
<NightMonkey>
enebo: This information is super-helpful. I'm glad that jRuby is still active and moving forward.
<enebo>
lopex: you changes pom?
<enebo>
for the java change?
<lopex>
enebo: yeah source and target
<enebo>
NightMonkey: yeah. It will be really nice to have 2.5 out of the door soon.
<lopex>
enebo: you could have target lower than source in 5 and 6 days
<lopex>
enebo: I wonder if it;s javac or maven compiler plugin explicit limitation now
<lopex>
enebo: maybe they worry about the apis right ?
<enebo>
lopex: I don't think I personally care and the only consumers I think are nashorn (which is forked so it doesn't matter ... plus it is in something newer than 7 anyways) and jython
<enebo>
I don't know if jython cares or not
<enebo>
There might be other consumers
<lopex>
enebo: I looked for the consumers on mvnrepository
<enebo>
lopex: oh yeah?
<lopex>
none seem to be affected
<enebo>
lopex: ok well 7 is not such a big requirement any more and if we get big pushback we can go back I guess
<lopex>
enebo: I needed that just for strings in switch
<enebo>
lopex: I have lately been an advocate of more frequent point releases of our satellite projects so we are not depending on SNAPSHOT so much in core
<lopex>
so I assumed I could leave target at 1.5
<enebo>
SNAPSHOT makes bisecting horrible
<lopex>
but it didnt work
<enebo>
string switch is 7? I thought that was 8?
<lopex>
7
<enebo>
ah ok
<lopex>
enebo: yeah, I recall headius pointing the you care about theose being buildable
<enebo>
yeah so if someone complains we will just fix that switch and drop down
<lopex>
so snapshots are out
<enebo>
lopex: they are small have no manual QA and numbers are cheap
<enebo>
lopex: so let's rev more on joni/jcodings/jnr-*
<lopex>
enebo: but my concern was that strings in switch doesnt really require any new apis
<enebo>
lopex: it just requires the compiler
<lopex>
so there could be a possibility to leave the target at 1.5
<lopex>
yeah, but not at runtime which is the case
<enebo>
lopex: funny I commented on this a few weeks ago that Javac and java libraries should be decoupled
<lopex>
yea
<enebo>
lopex: because Java 10 type inference is nothing but compiler tricks
<lopex>
enebo: but it requires some metainfo
<enebo>
tricks is wrong word
<lopex>
sugar ?
<enebo>
lopex: I think they just formulate a synthetic type in worst cases
<enebo>
which maybe JVM needs to be able to sanely display at runtime so perhaps there is more than just compiler support
<lopex>
enebo: like for javac, "this construct uses 1.5 apis"
<lopex>
like strings in switch
<enebo>
yeah
<lopex>
since they use just hasCode for that
<lopex>
and equals
<enebo>
but generally it seems like some language-level features are independent of the APIs
<enebo>
I almost wish some of this stuff was done through compiler extensions
<lopex>
that the point
<enebo>
so you could potentially use new compiler feature in older JVM
<lopex>
enebo: same for possible macros, if the were there
<lopex>
which is the same thing at this point
<enebo>
yeah I sometimes wish we had hygenic macros in Java
<lopex>
oh well, the ast might change :P
<enebo>
we have a lot of repetitive code in places and things could look a lot nicer
<enebo>
but at least being hygenic it would be a proper AST on expansion
<enebo>
anyways I can dream :)
<lopex>
enebo: macros in scala are still experimental
<enebo>
Scala has it all!
<enebo>
I want less than all but it will be neat to see how well macros work for them
<lopex>
there's always two camps
<lopex>
macro camp and type system astronauts camp
jrafanie_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jrafanie has joined #jruby
<GitHub122>
[jruby] MSNexploder opened pull request #5124: :newline option to File.open implies text mode now. (Part of #5118) (master...newline) https://git.io/vxwbe