#jython IRC Log

Index

IRC Log for 2015-04-29

Timestamps are in GMT/BST.

[0:04] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[0:41] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[0:45] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Client Quit)
[1:49] <jimbaker> gsnedders, agronholm - glad to hear this!
[1:51] <agronholm> oh ffs: https://github.com/NextThought/virtualenv/commit/b6919ae2fbbee5802fd4c8f4779697469f2715ec
[1:51] <agronholm> just when we got upstream fixed >_<
[1:54] <jimbaker> agronholm, well, at some point we will get it sorted out
[1:54] <jimbaker> :)
[1:54] <agronholm> the whole thing is blocked solely by this: https://github.com/pypa/virtualenv/pull/747
[2:00] <jimbaker> agronholm, let me discuss on #pypa-dev
[3:30] * gthank (~gthank@unaffiliated/gthank) Quit (Ping timeout: 248 seconds)
[3:31] * gthank (~gthank@thermi.strangled.net) has joined #jython
[3:31] * gthank (~gthank@thermi.strangled.net) Quit (Changing host)
[3:31] * gthank (~gthank@unaffiliated/gthank) has joined #jython
[3:48] * gthank (~gthank@unaffiliated/gthank) Quit (Ping timeout: 248 seconds)
[3:50] * gthank (~gthank@thermi.strangled.net) has joined #jython
[3:50] * gthank (~gthank@thermi.strangled.net) Quit (Changing host)
[3:50] * gthank (~gthank@unaffiliated/gthank) has joined #jython
[4:11] * trigoman (~trigoman@75-161-18-131.albq.qwest.net) has joined #jython
[4:38] * agronholm changes topic to 'Jython 2.7 final released (soft launch): http://tinyurl.com/mec6jpc | This channel is logged: http://jython.extreme.st/irclogs/ | Please update the wiki: http://wiki.python.org/jython | Jython Book: http://jythonbook.com | Podcast: http://jython.org/jythonpodcast/'
[4:51] <agronholm> something's wrong with that 2.7.0 jar
[4:53] <agronholm> oh hum I linked the wrong jar
[4:53] * agronholm changes topic to 'Jython 2.7 final released (soft launch): http://tinyurl.com/muqfuul | This channel is logged: http://jython.extreme.st/irclogs/ | Please update the wiki: http://wiki.python.org/jython | Jython Book: http://jythonbook.com | Podcast: http://jython.org/jythonpodcast/'
[4:57] <agronholm> jimbaker: should java -jar work with the standalone jar?
[4:58] <agronholm> or does that lead to the old problem of loading stdlib modules from the jar?
[6:08] * trigoman (~trigoman@75-161-18-131.albq.qwest.net) Quit (Read error: Connection reset by peer)
[6:08] * trigoman (~trigoman@75-161-18-131.albq.qwest.net) has joined #jython
[6:11] * ohumbel (5390f784@gateway/web/freenode/ip.83.144.247.132) Quit (Quit: Page closed)
[6:36] * koo6 (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 244 seconds)
[6:38] * trigoman (~trigoman@75-161-18-131.albq.qwest.net) Quit (Ping timeout: 255 seconds)
[9:33] <mbooth> Congratulations all on the tagging of 2.7.0 release :-)
[10:50] <grey_> We were evaluating Jython for use as an embedded interpreter. I threw away a bunch of stuff in a fairly crude manner, and got the jar shrinked down to 9.2 MB
[10:51] <grey_> 673 kb for Lib/site.py and it's dependencies
[10:52] <grey_> are there any thoughts or plans on a "core only" jar?
[10:58] <agronholm> grey_: the trouble is, different people need different things -- what would you consider to be "core" functionality?
[10:59] <agronholm> dropping the standard library altogether?
[11:01] <grey_> agronholm: I know it's a difficult question. From the angle of _my_ use case, it would be safe to drop everything related to network and file system access
[11:01] <grey_> since those would have to be sandboxed anyway.
[11:03] <grey_> I let site.py in just to test, but I would imagine none of it would be required for scripts that calculate and return numeric and boolean values (as Java objects)
[11:04] <agronholm> grey_: can you sandbox java's networking and file stuff then?
[11:04] <agronholm> there's nothing stopping the interpreter from using those
[11:19] <grey_> I understand that it should be possible with a SecurityManager
[11:19] <grey_> Haven't tried it yet, though
[14:07] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[14:12] * xemdetia (xemdetia@nat/ibm/x-hcqcbdlpewbnozhx) has joined #jython
[14:52] <jimbaker> agronholm, grey_ - i did some work on ensuring jython works with a SecurityManager properly, especially around file ops
[14:52] <jimbaker> re shrinking the jar - that sounds great. it would be interesting to come up with a set of profiles for jython
[14:52] <jimbaker> grey_, did you also try removing *.py files in favor of just $py.class?
[14:53] <jimbaker> on windows, we did get this bug report: http://bugs.jython.org/issue2346
[14:53] <jimbaker> i don't like how this leaking out jython.exe was built using my z: drive, that mapped my home directory on OSX
[15:01] <jimbaker> agronholm, i just noticed that noncommitters can set the milestone in bugs.jython.org, which sort of defeats the purpose as i understand it
[15:02] <jimbaker> re issue 2346 - great bug number! - it looks like release notes to me
[15:04] * koo6 (~koo5@236.152.broadband3.iol.cz) has joined #jython
[16:30] <gsnedders> I need to stop finding weird obscure bugs when it comes to lone surrogates in Jython. :)
[16:38] * dyce (~quassel@ns3290920.ip-5-135-184.eu) has joined #jython
[16:45] <jimbaker> gsnedders, no worries, file away!
[16:45] <gsnedders> At some point I'll probably try and fix them.
[17:32] * xemdetia_ (xemdetia@nat/ibm/x-rygddinxhaxlkmdr) has joined #jython
[17:35] * xemdetia (xemdetia@nat/ibm/x-hcqcbdlpewbnozhx) Quit (Ping timeout: 250 seconds)
[17:45] * koo6 (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 264 seconds)
[19:05] * trigoman (~trigoman@75-161-67-24.albq.qwest.net) has joined #jython
[19:24] <agronholm> jimbaker: I have no idea if that can be fixed easily
[19:57] * xemdetia_ is now known as xemdetia
[20:26] * koo6 (~koo5@236.152.broadband3.iol.cz) has joined #jython
[21:24] <jimbaker> agronholm, no worries - just want to minimize triage here
[21:24] <jimbaker> since that's the purpose of that milestone
[22:23] <pjenvey> jimbaker - now jython has the same problem pypy does, with a jython executable being written in python itself: you get extra traceback lines of the executable's code (some of them having that Z: drive) =P
[22:24] <pjenvey> does it really fork a subprocess everytime?
[22:31] <jimbaker> pjenvey, yes, it does
[22:31] <jimbaker> it does the absolutely simplest braindead thing possible
[22:31] <jimbaker> but only on windows
[22:31] * gsnedders (~gsnedders@5.2.16.23) Quit (Ping timeout: 244 seconds)
[22:32] <jimbaker> or if needs to capture (example: profiling). otherwise it uses os.execvp
[22:33] <agronholm> could we not implement the launcher as a powershell script?
[22:34] <jimbaker> agronholm, sure
[22:36] <agronholm> what's the minimum windows version we want to support?
[22:37] <jimbaker> agronholm, only ones that support java 7
[22:37] <agronholm> I don't know what those are
[22:37] <jimbaker> i assume this excludes windows xp, etc
[22:37] <jimbaker> yeah, i don't either
[22:37] <agronholm> powershell 2 is integrated into win7 and is available for xp sp3
[22:37] <jimbaker> the whole point of writing jython.py was to minimize such knowledge :)
[22:39] <pjenvey> fyi i made pypy3's runner hide its own stack frames, pypy2 still shows them but it could maybe be fixed too with some more effort, https://bitbucket.org/pypy/pypy/commits/894b0fa
[22:39] <pjenvey> that was kind of a fun hack
[22:40] <jimbaker> pjenvey, i'm sure we could do something similar
[22:40] <pjenvey> yea
[22:40] <pjenvey> i don't think jython's pyframe has a hidden_applevel like attribute, you'd have to add one, not too much more work. pypy luckily already had it
[22:41] <jimbaker> pjenvey, we need to rework stack trackbacks anyway, to support better java info when that's the causing exception
[22:41] <jimbaker> agronholm, it's a reasonable list for windows - http://java.com/en/download/help/sysreq.xml
[22:42] <jimbaker> so we could do this with powershell
[22:42] <jimbaker> again, it does need to be an exe
[22:43] <pjenvey> er it's on the code object
[22:46] * xemdetia (xemdetia@nat/ibm/x-rygddinxhaxlkmdr) Quit (Ping timeout: 276 seconds)
[23:00] * koo6 (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 276 seconds)
[23:01] <jimbaker> however, there are certain advantages of having the launcher written in python. for instance, i'm pretty sure everyone here knows how to write python ;) more importantly, we have a single source solution across platforms
[23:03] <jimbaker> i also looked at jython.py with pyqver (https://github.com/ghewgill/pyqver), and it should be straightforward to backport to 2.4, which should nearly every cpython out there that we would care about, such as RHEL5
[23:03] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[23:05] <jimbaker> 2.4 brought in subprocess support, and it's easy enough to replace subprocess.check_output, etc
[23:09] * gsnedders (~gsnedders@5.2.16.23) has joined #jython
[23:09] <gsnedders> https://pypi.python.org/pypi/html5lib/0.9999 ??? weee!
[23:10] <agronholm> 0.9999? seriously?
[23:11] <agronholm> still, glad to hear it's been released
[23:36] <gsnedders> agronholm: there's a list of API breakages planned for 1.0 as well as better docs for 1.0
[23:37] <gsnedders> agronholm: till then we're releasing when people want us to, asymptotically approaching 1.0 :)
[23:44] <jimbaker> gsnedders, :)
[23:45] <gsnedders> (Honestly, we should probably have called a release 1.0 a year or two ago and just shipped without decent docs or those API breakages and just push them to 2.0)
[23:46] <gsnedders> (But it now seems /so close/ that it seems silly.)
[23:51] <gsnedders> But yeah, in general we should always be in a place where we can ship master, so if there's anything on master you guys want in a release just prod. I'm almost always on freenode, even not in here.

Index

These logs were automatically created by JythonLogBot on irc.freenode.net using a slightly modified version of the Java IRC LogBot.