On 13/09/2009, Henrib <hbies...@gmail.com> wrote: > > Seems like integer (long versus int in test assertions) from what I could > dig. >
Yes, just found that too. This is a recent change in JEXL. The code is comparing file.length*2 with file.length - the former is being downcast to an Integer, which is wrong. > > sebb-2-2 wrote: > > > > On 13/09/2009, sebb <seb...@gmail.com> wrote: > >> On 12/09/2009, Rahul Akolkar <rahul.akol...@gmail.com> wrote: > >> > On Sat, Sep 12, 2009 at 9:49 AM, sebb <seb...@gmail.com> wrote: > >> > > On 12/09/2009, sebb <seb...@gmail.com> wrote: > >> > >> On 12/09/2009, Rahul Akolkar <rahul.akol...@gmail.com> wrote: > >> > >> > On Fri, Sep 11, 2009 at 10:45 AM, Paul Libbrecht > >> <p...@activemath.org> wrote: > >> > >> > > Hello Jexl developers, > >> > >> > > > >> > >> > > has ExpressionFactory disappeared? > >> > >> > > Did package-names maybe change? > >> > >> > > > >> > >> > > I guess I would appreciate to revert that bit, maybe by > >> explicitly depending > >> > >> > > on jexl 1? > >> > >> > > > >> > >> > > >> > >> > <snip/> > >> > >> > > >> > >> > You could depend on 1.1, yes. > >> > >> > > >> > >> > As gump is pointing out though, this is one breakage that'll > >> be in > >> > >> > parts seeing widespread use. On the JEXL side, thinking its > >> best to > >> > >> > add deprecated Expression/ScriptFactory versions that delegate > >> to the > >> > >> > JexlEngine API. > >> > >> > >> > >> > >> > >> +1 > >> > >> > >> > > > >> > > Just noticed that we both agreed to removing these classes - see > >> JEXL-72. > >> > > > >> > > It looks as though the classes were only deprecated in the 2.0 code > >> > > line, so I guess we should not be removing them yet. > >> > > >> > <snip/> > >> > > >> > Gump is deftly persuasive -- the cleaner, less redundant API will > >> have > >> > to wait. So yes, best to reopen JEXL-72. > >> > > >> > >> > >> Gump is now happy again. > > > > Wrote too soon! Although the compilation errors have gone, the test > > suite has some errors (e.g. variables not defined). > > > >> The deprecated methods now create a new JEXL engine each time, whereas > >> previously they shared a static instance of the engine. Is that > >> reasonable, or should the code try to be closer to the original and > >> use a shared instance? > > > > The test errors look as though the Jelly tests may be depending on a > > shared engine. I'll take a look at this shortly. > > > >> > >> > -Rahul > >> > > >> > > >> > --------------------------------------------------------------------- > >> > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > >> > For additional commands, e-mail: dev-h...@commons.apache.org > >> > > >> > > >> > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > > For additional commands, e-mail: dev-h...@commons.apache.org > > > > > > > > > -- > View this message in context: > http://www.nabble.com/-GUMP%40vmgump-%3A-Project-commons-jelly-%28in-module-commons-jelly%29-failed-tp25399459p25427936.html > Sent from the Commons - Dev mailing list archive at Nabble.com. > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org