Re: [JXPath] Whitespace in Expression/Property
Works!!! The problem was ME and my silly data structures :) Thanks HEAPS Matt. On Mon, Mar 23, 2009 at 10:55 PM, Andrew Hughes wrote: > Doesn't work, but I'll need to develop some testing code in order to > publically re-produce how this occurs. > p.s. both name and @name don't work. > > > p.p.s THANKS HEAPS MATT!!! > > > On Wed, Mar 18, 2009 at 2:42 AM, Matt Benson wrote: > >> >> Maybe /Earth/Australia/@name='New South Wales']/Sydney? :| >> >> -Matt >> >> --- On Mon, 3/16/09, Andrew Hughes wrote: >> >> > From: Andrew Hughes >> > Subject: [JXPath] Whitespace in Expression/Property >> > To: "Commons Users List" >> > Date: Monday, March 16, 2009, 11:42 PM >> > Hi Again, >> > How can I access a property named 'New South Wales' if it >> > has whitespace in >> > it? Example Expression: '/Earth/Australia/New South >> > Wales/Sydney' >> > >> > THANKS SO MUCH >> > >> >> >> >> >> - >> To unsubscribe, e-mail: user-unsubscr...@commons.apache.org >> For additional commands, e-mail: user-h...@commons.apache.org >> >> >
Re: [SCXML] Log output?
Please post in plain text or "internet style" as some email clients like to call it ... On Wed, Mar 25, 2009 at 9:39 AM, Anna Södling wrote: >> Hi, >> >> >> I want to use some logging in my scxml-file, for example upon entering >> and exiting a state. Unfortunately, I can't seem to get the hang of > how >> it works. I've got this line in my scxml-file: >> >> >> And the log output I get (in a console window in Eclipse) looks like >> this: >> 2009-mar-24 11:39:08 org.apache.comm >> ons.scxml.model.Log execute >> INFO: 'Renegade': Entering state: A >> 2009-mar-24 11:39:08 org.apache.commons.scxml.model.Log execute >> INFO: 'Renegade': Entering state: A >> > > > That seems OK to me (other than the fact it appeared twice, probably a > cut-n-paste error?). > > No, it was supposed to enter state A twice during the running of the > state machine...though I guess I didn't really have to paste both > outputs :) > However, I'm wondering if there is some way to "clean up" the output? If > possible, I would like to remove the "org.apache.commons.scxml.model.Log > execute" and put the date and info on the same row. > You'll have to look into a formatter / layout for the logging library you are using (or introduce one so you can do more than the commons-logging SimpleLog). >> Also, I'm doing some ordinary printed outputs (in Java) while I'm >> running my state machine, and the log info is not placed on the right >> place among these outputs either, but before all of them (so it > doesn't >> show up when I'm actually entering the state). >> And that's not really what I wanted. I would simply like to have an >> output that says "'Renegade': Entering state: A" (or something > similar) >> when I'm actually entering state A. Is this possible? Or have I >> missunderstood how the -tag works? >> > > > Please say more about the Java code which contains these print > statements. Are you using an AbstractStateMachine subclass, an > SCXMLListener or something else? ( content in the SCXML > document executes before code in both those specific cases). > > We're using an AbstractStateMachine...we haven't extended it (since this > didn't work for some reason), but rather copied it and added some parts. > But it's a kind of ASM very similar to the original. > We have a little problem with this line of code in the onentry-listener > though: > invoke(entered.getId()); > Without it, we only get the original output, but when we add it, we get > a whole lot of error messages. We're not sure what it is doing, or > invoking? > As the class Javadoc says, it calls the namesake method in the class for each state that has been entered. > > If you want more fine-grained control about the timing of your > procedural code within state changes (or sequencing even within a > particular or ), that kind of thing can be achieved > by using custom actions [1] to call your code from the appropriate > place(s) in SCXML executable content. > > > > We've tried with several log expressions, some in onentry-tags, and some > not. Even though they're not shown exactly where we want to, they seem > to be shown in the right order. But, is this maybe a coincidence? > I don't think so, but its hard to say anything without knowing the code. -Rahul > /Anna > > > > . > > - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
Re: [net] GSoC - Add SSH/SCP Support to Commons-Net
Hi all, Regarding including SSH/SCP in commons-net. Some of projects that i'm involved in, due to non-technical reasons prohibit direct usage of any cryptogrphic libraries (binary-dependency). Including, for example, any SSH/SCP libraries (j2ssh, jsch, etc.). Thus, i would appreciate adding SCH/SSH functionality as a new library for commons, not as an addition to commons-net. One more thing. Friend of mine from MIPT laboratory created small commons-ssh library ( http://commons-ssh.sourceforge.net/ ) that behaves like an thin adapter to known SSH implementation. He is okay to move it under apache commons project and extends to support SCP. -- Sergey On Thu, Mar 26, 2009 at 3:45 PM, Torsten Curdt wrote: > Hey Vlad, > > Welcome! A few suggestions... > > o engage with the community beforehand (you did that - good!) > o some projects even require a first patch (to judge if the student is > up to the work) > o communicate in an open manner - communicate often (given the past > experience I would fail students that cannot manage to communicate at > least once a week) > o while your application is expected to give a high level outline of > project plan, also provide details that show that you know what you > are talking about > > Also feel free to contact me also off the list if you need further help. > > cheers > -- > Torsten > > - > To unsubscribe, e-mail: user-unsubscr...@commons.apache.org > For additional commands, e-mail: user-h...@commons.apache.org > > -- Sergey Vladimirov - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
Re: [ANN] commons-exec 1.0 released - Part 2
Hi Gary, Source and binary distributions can be downloaded from the corresponding directories in an Apache Software Foundation mirror at: http://www.apache.org/dyn/closer.cgi/commons/exec See also the download page at: http://commons.apache.org/exec Cheers, Siegfried Goeschl Gary Gregory wrote: > You should include links to the site and download area... > > Gary > > >> -Original Message- >> From: Siegfried Goeschl [mailto:siegfried.goes...@it20one.at] >> Sent: Thursday, March 26, 2009 11:35 AM >> To: Jakarta Commons Developers List; Jakarta Commons Users List >> Subject: [ANN] commons-exec 1.0 released >> >> The commons-exec-team is pleased to announce the commons-exec-1.0 release! >> >> A library to reliably execute external processes from within the JVM >> >> Changes in this version include: >> >> New features: >> -- >> --- >> >> o Added one additional test : DefaultExecutorTest.testExecuteWithFancyArg >> Issue: EXEC-26. Thanks to Benjamin Bentmann. >> o Added convinience method to add two parameters to the CommandLine >> using one method invocation. >> o Added test scripts for OpenVMS - he seems to be the last human having >> access to an OpenVMS box ... :-) >> o With the help of the Apache Commons community I added the first results >> of cross-OS testing. Thanks to Simone Gianni, Bindul Bhowmik, >> Niall Pemberton, Sebastian Bazley. >> o The regression tests now also works on Windows - so it should >> work now on Linux, Windows and Mac OS X >> o Added DebugUtils to improve cross-platform testing. >> o Made DefaultExecutor.launch() protected to enable mocking. >> Issue: SANDBOX-62. Thanks to Jeremy Lacoste. >> o Made ProcessDestroyer optional and pluggable when using Executor. >> Issue: SANDBOX-107. Thanks to Niklas Gustavsson. >> o CommandLine can now expand the given command line by a user-suppied >> map. This allows to execute something like "${JAVA_HOME}/bin/java -jar >> ${myapp}" >> o Added methods to provide pre-quoted arguments. >> Issue: SANDBOX-192. Thanks to Reinhold Fuereder. >> o Exposing a ExecuteWatchdog.destroy() to kill an asynchrounous process >> manually. This formalizes a workaround described in the JIRA >> Issue: SANDBOX-193. Thanks to Reinhold Fuereder. >> o Extending exit value handling to support applications returning an error >> code. Issue: SANDBOX-203. >> o Added a few regression tests for the watchdog since they were missing. >> Issue: SANDBOX-204. >> >> Fixed Bugs: >> -- >> --- >> >> o Removed useless synchronized statement in >> OpenVmsProcessingEnvironment.createProcEnvironment >> Issue: EXEC-37. Thanks to Sebastien Bazley. >> o Using System.in for child process will actually hang your application - >> see JIRA for more details. Since there is no easy fix an >> IllegalRuntimeException is thrown when System.in is passed. >> Issue: EXEC-33. >> o Fixing a few findbugs issues. Issue: EXEC-35. Thanks to Luc Maisonobe. >> o Handle null streams consistently. Issue: EXEC-32. Thanks to Marco >> Ferrante. >> o After a long discussion we decided to stick to following groupId >> "org.apache.commons" instead of "commons-exec". >> o The Ant build now works even when junit is not on the classpath >> Thanks to Kevin Jackson. >> o Fixed broken "groupId" from "org.apache.commons" to "commons-exec" >> o Renamed EnvironmentUtil to EnvironmentUtils to align with other classes >> in this project and commons in general. Please note that this change >> could break existing clients (but would be rather unlikely). >> Issue: EXEC-27. Thanks to Benjamin Bentmann. >> o Make environment variables respect casing rules of platforms. Under >> Windows >> "PATH", "Path" and "path" would access the same environment variable >> whereas >> the real name is "Path". Issue: EXEC-30. Thanks to Benjamin Bentmann. >> o Invoking DefaultExecutor.execute(CommandLine command, Map environment) >> using >> a 'null' Map results in inheriting all environment variables of the >> current >> process while passing an empty map implies starting the new process >> with no >> environment variables. In short 'null' is not the same as an empty map. >> Issue: EXEC-31. Thanks to Benjamin Bentmann. >> o Using variable substitution within CommandLine broke the regression >> tests >> under Windows. Found also another bug when calling >> CommandLine.getExecutable() >> the result was not substituted at all. As a general rule we do variable >> substitution and file separator fixing on the command line executable >> and >> variable substitution but NO file separator fixing for the command line >> arguments. Issue: EXEC-25. >> o Implemented better regression test for OpenVMS affecting also >> the Executor and CommandLauncher interface. >> o Cleaned up the source code to get rid of javadoc errors and >> unuse
Re: [ANN] commons-exec 1.0 released
Great job seeing this through, Siegfried! -Matt --- On Thu, 3/26/09, Siegfried Goeschl wrote: > From: Siegfried Goeschl > Subject: [ANN] commons-exec 1.0 released > To: "Jakarta Commons Developers List" , "Jakarta > Commons Users List" > Date: Thursday, March 26, 2009, 1:34 PM > The commons-exec-team is pleased to > announce the commons-exec-1.0 release! > > A library to reliably execute external processes from > within the JVM > > Changes in this version include: > > New features: > - > > o Added one additional test : > DefaultExecutorTest.testExecuteWithFancyArg > Issue: EXEC-26. Thanks to Benjamin Bentmann. > o Added convinience method to add two parameters to the > CommandLine > using one method invocation. > o Added test scripts for OpenVMS - he seems to be the last > human having > access to an OpenVMS box ... :-) > o With the help of the Apache Commons community I added the > first results > of cross-OS testing. Thanks to Simone Gianni, Bindul > Bhowmik, > Niall Pemberton, Sebastian Bazley. > o The regression tests now also works on Windows - so it > should > work now on Linux, Windows and Mac OS X > o Added DebugUtils to improve cross-platform testing. > o Made DefaultExecutor.launch() protected to enable > mocking. > Issue: SANDBOX-62. Thanks to Jeremy Lacoste. > o Made ProcessDestroyer optional and pluggable when using > Executor. > Issue: SANDBOX-107. Thanks to Niklas Gustavsson. > o CommandLine can now expand the given command line by a > user-suppied > map. This allows to execute something like > "${JAVA_HOME}/bin/java -jar > ${myapp}" > o Added methods to provide pre-quoted arguments. > Issue: SANDBOX-192. Thanks to Reinhold Fuereder. > o Exposing a ExecuteWatchdog.destroy() to kill an > asynchrounous process > manually. This formalizes a workaround described in > the JIRA > Issue: SANDBOX-193. Thanks to Reinhold Fuereder. > o Extending exit value handling to support applications > returning an error > code. Issue: SANDBOX-203. > o Added a few regression tests for the watchdog since they > were missing. > Issue: SANDBOX-204. > > Fixed Bugs: > - > > o Removed useless synchronized statement in > > OpenVmsProcessingEnvironment.createProcEnvironment > Issue: EXEC-37. Thanks to Sebastien Bazley. > o Using System.in for child process will actually hang your > application - > see JIRA for more details. Since there is no easy > fix an > IllegalRuntimeException is thrown when System.in is > passed. > Issue: EXEC-33. > o Fixing a few findbugs issues. Issue: EXEC-35. > Thanks to Luc Maisonobe. > o Handle null streams consistently. Issue: EXEC-32. > Thanks to Marco > Ferrante. > o After a long discussion we decided to stick to following > groupId > "org.apache.commons" instead of "commons-exec". > o The Ant build now works even when junit is not on the > classpath > Thanks to Kevin Jackson. > o Fixed broken "groupId" from "org.apache.commons" to > "commons-exec" > o Renamed EnvironmentUtil to EnvironmentUtils to align with > other classes > in this project and commons in general. Please note > that this change > could break existing clients (but would be rather > unlikely). > Issue: EXEC-27. Thanks to Benjamin Bentmann. > o Make environment variables respect casing rules of > platforms. Under > Windows > "PATH", "Path" and "path" would access the same > environment variable > whereas > the real name is "Path". Issue: EXEC-30. > Thanks to Benjamin Bentmann. > o Invoking DefaultExecutor.execute(CommandLine command, Map > environment) > using > a 'null' Map results in inheriting all environment > variables of the current > process while passing an empty map implies starting > the new process > with no > environment variables. In short 'null' is not the > same as an empty map. > Issue: EXEC-31. Thanks to Benjamin Bentmann. > o Using variable substitution within CommandLine broke the > regression tests > under Windows. Found also another bug when calling > CommandLine.getExecutable() > the result was not substituted at all. As a general > rule we do variable > substitution and file separator fixing on the > command line executable and > variable substitution but NO file separator fixing > for the command line > arguments. Issue: EXEC-25. > o Implemented better regression test for OpenVMS affecting > also > the Executor and CommandLauncher interface. > o Cleaned up the source code to get rid of javadoc errors > and > unused imports. Issue: SANDBOX-204. > > > Removed: > o Removed commons-logging integration > > Have fun! > > -commons-exec-team > > > - > To unsubscribe, e-mail: user-unsubscr...@commons.apache.org > For additional commands, e-mail: user-h...@commons.apache.org > >
[ANN] commons-exec 1.0 released
The commons-exec-team is pleased to announce the commons-exec-1.0 release! A library to reliably execute external processes from within the JVM Changes in this version include: New features: - o Added one additional test : DefaultExecutorTest.testExecuteWithFancyArg Issue: EXEC-26. Thanks to Benjamin Bentmann. o Added convinience method to add two parameters to the CommandLine using one method invocation. o Added test scripts for OpenVMS - he seems to be the last human having access to an OpenVMS box ... :-) o With the help of the Apache Commons community I added the first results of cross-OS testing. Thanks to Simone Gianni, Bindul Bhowmik, Niall Pemberton, Sebastian Bazley. o The regression tests now also works on Windows - so it should work now on Linux, Windows and Mac OS X o Added DebugUtils to improve cross-platform testing. o Made DefaultExecutor.launch() protected to enable mocking. Issue: SANDBOX-62. Thanks to Jeremy Lacoste. o Made ProcessDestroyer optional and pluggable when using Executor. Issue: SANDBOX-107. Thanks to Niklas Gustavsson. o CommandLine can now expand the given command line by a user-suppied map. This allows to execute something like "${JAVA_HOME}/bin/java -jar ${myapp}" o Added methods to provide pre-quoted arguments. Issue: SANDBOX-192. Thanks to Reinhold Fuereder. o Exposing a ExecuteWatchdog.destroy() to kill an asynchrounous process manually. This formalizes a workaround described in the JIRA Issue: SANDBOX-193. Thanks to Reinhold Fuereder. o Extending exit value handling to support applications returning an error code. Issue: SANDBOX-203. o Added a few regression tests for the watchdog since they were missing. Issue: SANDBOX-204. Fixed Bugs: - o Removed useless synchronized statement in OpenVmsProcessingEnvironment.createProcEnvironment Issue: EXEC-37. Thanks to Sebastien Bazley. o Using System.in for child process will actually hang your application - see JIRA for more details. Since there is no easy fix an IllegalRuntimeException is thrown when System.in is passed. Issue: EXEC-33. o Fixing a few findbugs issues. Issue: EXEC-35. Thanks to Luc Maisonobe. o Handle null streams consistently. Issue: EXEC-32. Thanks to Marco Ferrante. o After a long discussion we decided to stick to following groupId "org.apache.commons" instead of "commons-exec". o The Ant build now works even when junit is not on the classpath Thanks to Kevin Jackson. o Fixed broken "groupId" from "org.apache.commons" to "commons-exec" o Renamed EnvironmentUtil to EnvironmentUtils to align with other classes in this project and commons in general. Please note that this change could break existing clients (but would be rather unlikely). Issue: EXEC-27. Thanks to Benjamin Bentmann. o Make environment variables respect casing rules of platforms. Under Windows "PATH", "Path" and "path" would access the same environment variable whereas the real name is "Path". Issue: EXEC-30. Thanks to Benjamin Bentmann. o Invoking DefaultExecutor.execute(CommandLine command, Map environment) using a 'null' Map results in inheriting all environment variables of the current process while passing an empty map implies starting the new process with no environment variables. In short 'null' is not the same as an empty map. Issue: EXEC-31. Thanks to Benjamin Bentmann. o Using variable substitution within CommandLine broke the regression tests under Windows. Found also another bug when calling CommandLine.getExecutable() the result was not substituted at all. As a general rule we do variable substitution and file separator fixing on the command line executable and variable substitution but NO file separator fixing for the command line arguments. Issue: EXEC-25. o Implemented better regression test for OpenVMS affecting also the Executor and CommandLauncher interface. o Cleaned up the source code to get rid of javadoc errors and unused imports. Issue: SANDBOX-204. Removed: o Removed commons-logging integration Have fun! -commons-exec-team - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
Re: Commons XMLConfiguration Issue --XPath query with /root
Thanks Oliver for clarifying this Oliver Heger-3 wrote: > > bnreddy schrieb: >> Hi christian, >> >> From my understanding (correct me if I am wrong)... "The root element >> is >> ignored when constructing keys". but this is when XMLConfiguration is >> using >> DefaultExpressionEngine >> >>Is root element ignored even when the XPathExpressionEngine is being >> used >> by XMLConfig... >> >> The problem is ...there are preexisting XPaths and we need to fit this >> in >> that... >> >> as per shouldn't the XMLConfiguration be able to support all legal >> XPath's >> instead of trimming root elements... > > XMLConfiguration used to ignore the root element from the very > beginning. Expression engines were introduces later, but when the > refactoring was done this behavior was not changed to remain backwards > compatible. > > When XPathExpressionEngine was implemented it seems only natural to > mimic the behavior of DefaultExpressionEngine to be consistent. > > Oliver > >> >> >> >> this is noted in the user guide: XPathExpressionEngine expressions >> ignore the root element - so just leave it away. What problems could >> arise from this? >> >> christian! >> > > > - > To unsubscribe, e-mail: user-unsubscr...@commons.apache.org > For additional commands, e-mail: user-h...@commons.apache.org > > > -- View this message in context: http://www.nabble.com/Commons-XMLConfiguration-Issue---XPath-query-with--root-tp22704510p22723428.html Sent from the Commons - User mailing list archive at Nabble.com. - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
[BETWIXT] Bitwixt XMLOutput
> Hi, > > I have a question here. > I am using bitwixt to write objects as xml in a file. > But there are line feeds added after each line in the output. > E.g. > > > John Smith > 21 > > > I want to get rid of these line feeds and display the whole xml in a > single line. > Like this:John > Smith21 > > > Any ideas ? > > Thanks, > Thanks, > Sujit Srivastava > > - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
Re: [net] GSoC - Add SSH/SCP Support to Commons-Net
Hey Vlad, Welcome! A few suggestions... o engage with the community beforehand (you did that - good!) o some projects even require a first patch (to judge if the student is up to the work) o communicate in an open manner - communicate often (given the past experience I would fail students that cannot manage to communicate at least once a week) o while your application is expected to give a high level outline of project plan, also provide details that show that you know what you are talking about Also feel free to contact me also off the list if you need further help. cheers -- Torsten - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
Re: [VFS] problem on runtime connection to SFTp server
It seems that the problem comes from release function, when I call It after copying files, the second SFTP connection fail. lmk wrote: > > I have > commons-net-2.0 > commons-vfs-1.0 > jsch-0.1.40 > > > the first call works..but the second one does not. > > > > > Jörg Schaible-2 wrote: >> >> lmk wrote at Dienstag, 24. März 2009 14:58: >> >>> >>> Hi Sergey, >>> >>> here the exception trace >>> >>> org.apache.commons.vfs.FileSystemException: Unknown scheme "sftp" in URI >>> "sftp://login:passw...@host/path/";. >> >> This is an error message, but not a stack trace. However, it seems that >> you >> do not have commons-net and jsch in your deps ... >> >> - Jörg >> >> >> - >> To unsubscribe, e-mail: user-unsubscr...@commons.apache.org >> For additional commands, e-mail: user-h...@commons.apache.org >> >> >> > > -- View this message in context: http://www.nabble.com/-VFS--problem-on-runtime-connection-to-SFTp-server-tp22664327p22718714.html Sent from the Commons - User mailing list archive at Nabble.com. - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org
Re: why is my withArgName argument being overwritten
sagun shakya wrote at Donnerstag, 26. März 2009 02:38: > Hi, > > I have an Option created as: > Option fooOption = OptionBuilder.withDescription("add foo") >.withLongOpt("add_foo") >.hasArgs(2) >.withArgName("foo") >.withArgName("bar") > .create("baz"); Because the OptionBuilder builds always only one option at a time. http://commons.apache.org/cli/usage.html - Jörg - To unsubscribe, e-mail: user-unsubscr...@commons.apache.org For additional commands, e-mail: user-h...@commons.apache.org