You can subscribe to this list here.
2008 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(1) |
Oct
(4) |
Nov
|
Dec
(2) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2009 |
Jan
(5) |
Feb
(10) |
Mar
(5) |
Apr
(18) |
May
(31) |
Jun
(3) |
Jul
(3) |
Aug
(4) |
Sep
(11) |
Oct
(6) |
Nov
(18) |
Dec
|
2010 |
Jan
|
Feb
|
Mar
(14) |
Apr
|
May
(2) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Jonas K. <jon...@us...> - 2010-05-14 16:58:01
|
Hi Nico let's continue this on the user list. 1. I will investigate this behaviour and come back to you. Problem is, that my laptop is currently broken, so it might take some time ;-) 2. Ajax synchronization should not be done by just waiting for some time, this is not very reliable for test execution. Use one of the waitForXXX methods: http://www.molyb.org/confluence/display/molyb/Molybdenum+Command+Reference e.g. for validating form data: waitForTextPresent "please provider your first name" If you don't care about any page interactions, just wait for the Ajax event to return successfully via: http://www.molyb.org/confluence/display/molyb/waitForAjax-action-moly Regards, Jonas Zitat von Nico de Haen <ma...@nd...>: > Hi list, > > I have 2 questions: > > 1. Is it possible to have loops in schedules? > > I have a test running 10 times in a loop. If I let it run in a testsuite > it runs through all loops, but if I include the test suite in a > schedule, the variable ${molybdenum_loop} is not substituted and so I > cannot load the loop specific data. > > 2. Is it possible, to define as a default for schedules and testsuites > to run in "walk" mode? > > Since I have a lot of Ajax stuff and most of the commands need a little > bit more time for execution as in the default "run" mode it would be > helpful to define the timegap between command execution > > Thanks, > Nico > |
From: Nico de H. <ma...@nd...> - 2010-05-09 08:00:56
|
Hi list, I have 2 questions: 1. Is it possible to have loops in schedules? I have a test running 10 times in a loop. If I let it run in a testsuite it runs through all loops, but if I include the test suite in a schedule, the variable ${molybdenum_loop} is not substituted and so I cannot load the loop specific data. 2. Is it possible, to define as a default for schedules and testsuites to run in "walk" mode? Since I have a lot of Ajax stuff and most of the commands need a little bit more time for execution as in the default "run" mode it would be helpful to define the timegap between command execution Thanks, Nico |
From: Claas T. <ct...@ct...> - 2010-03-15 20:55:31
|
There is no builtin functionality for creating JIRA issues. There is only the linking feature in the description of tests and commands: http://www.molyb.org/confluence/display/molyb/Features +0.7.3#Features0.7.3-desc We discussed a solution, which will work with other issue tracking systems than JIRA as well, but its not implemented. And it should be implemented in Molybdenum itself, not in the ANT tasks so it can be utilised without the build automation as well. Perhaps this were too much constraints to make it real... > Hi Claas, > > is this still up-to-date? > http://www.molyb.org/confluence/display/molyb/FAQ#FAQ-IsitpossibletocreateJIRAissuesfrombrokentestcases%3F > > You have built something superior, haven't you? > > Regards, > Jonas > > ------------------------------------------------------------------------------ > Download Intel® Parallel Studio Eval > Try the new software tools for yourself. Speed compiling, find bugs > proactively, and fine-tune applications for parallel performance. > See why Intel Parallel Studio got high marks during beta. > http://p.sf.net/sfu/intel-sw-dev > _______________________________________________ > Molybdenum-developer mailing list > Mol...@li... > https://lists.sourceforge.net/lists/listinfo/molybdenum-developer |
From: Claas T. <ct...@ct...> - 2010-03-15 20:20:25
|
let us discuss the design a little bit more in detail. Using XUL windows as a container could perhaps a good idea. Our module concept could stay as well. If we are embedding html in XUL, which is simply possible, we have to deal with namespaces and some things are quite different because the XUL document and window has not the same properties as the HTML pendants. For instance, there is no XMLHttpRequest object in XUL. Such things will break frameworks like jQuery. We could put all UI in a XUL browser element. Can we call all the XPCom stuff from there? We have to find out. Will we use jQueryUI or any other framework? Sometimes XUL is adding a big value, for instance for preferences dialogs. How do we mix up things here by holding the style consistent at the same time? > my +1 on both HTML and JQuery over XUL, preferably with a soft migration > path, e.g. keeping the dialogs in XUL and starting with migrating the > main window. Definitively need to get a closer look on Jetpack. > > Jonas > > ------------------------------------------------------------------------------ > Download Intel® Parallel Studio Eval > Try the new software tools for yourself. Speed compiling, find bugs > proactively, and fine-tune applications for parallel performance. > See why Intel Parallel Studio got high marks during beta. > http://p.sf.net/sfu/intel-sw-dev > _______________________________________________ > Molybdenum-developer mailing list > Mol...@li... > https://lists.sourceforge.net/lists/listinfo/molybdenum-developer |
From: Claas T. <ct...@ct...> - 2010-03-15 20:08:03
|
I am a fan of stupid upcounting versions and not sure if removing a feature legitimates a bigger version number jump ;-) Lets see, what features we are collecting for the next version... ...and release managers are having more weight voting for version number changes ;-) I created a ticket for FF2.0 removal ( MOLYB-330). We should collect code cleanup tasks as child tickets. > my +1 and I think we should move forward to 0.8 with that change and - > while we on it - then target full german language support for version 1.0 > Would you like to create a ticket for german language support describing in detail what you want to see (ui, command interface, report, help)? Thanks Claas > Jonas > > ------------------------------------------------------------------------------ > Download Intel® Parallel Studio Eval > Try the new software tools for yourself. Speed compiling, find bugs > proactively, and fine-tune applications for parallel performance. > See why Intel Parallel Studio got high marks during beta. > http://p.sf.net/sfu/intel-sw-dev > _______________________________________________ > Molybdenum-developer mailing list > Mol...@li... > https://lists.sourceforge.net/lists/listinfo/molybdenum-developer |
From: Claas T. <ct...@ct...> - 2010-03-15 19:49:43
|
Our current implementation is checking the type of a property (is it a function?) and its name. The type check is fine but checking the name doesn't really make sense because the very same property can be replaced with a different function object at runtime. Other than in Java, the 'interfaces' are not fixed - they can change dynamically. This is a feature often used in javascript and is - in the right hands - a big advantage. I want to use this feature - and at the same time - I want to check the arguments the function was called with. At the moment, I think, this can be done only inside the function itself without adding any new restriction. The reason is: There is no fixed execution context for a function in javascript. Where to put the contract other than inside the function itself? Furthermore, currently we are checking the name of the function only. We don't check any argument signature. But this is much more important than checking the function name. So for me, it sounds reasonable to add caller contract checks inside functions. It seems to be the most efficient way to have a contract applied in javascript. For instance, if we have the observer pattern and we like to add an observer to object OBSERVER. It will work like this: OBSERVER gets an object to be added as a SUBSCRIBER. OBSERVER checks the type of the object. If it is a function, it will be added to the list of subscribers. The time OBSERVER is calling the SUBSCRIBER, this is done with the arguments. No check of a signature is done in any way. The call is surrounded with a try catch block to catch unintended behavior of the SUBSCRIBER. The logging is important here, because this is the hint for a contract violation. The SUBSCRIBER is checking the passed arguments against the contract. If the contract is violated, an exception is thrown or the error is ignored silently and logged only, likewise. Of course, technically pseudo interfaces and the convention described above can coexist in our code. but is there a usecase where pseudo interfaces are bringing a higher benefit than the 'inside function contracts'? > Hi Claas, > > thanks for the input. I generally agree though I don't see that we're > too far away from those concepts or even going too far by having too > many lines of code. The object fulfilling the contract calls the > "fulfills" function itself to check if its signature matches the > predefined one (which could even be done in the constructor), i.e. > there's no external module checking that "interface" objects and > "implementation" objects have the same signature. > > I propose to call this "pseudo interface" from now on, just to formulate > our intention clearly. > > So it really boils down on how to check whether a consumer matches a > (potentially evolving) provider contract. If it's convenient to do this > by iterating over the object's functions and comparing them with those > of another object: fine. If this is not sufficient or even misleading: > fine as well. Let's just do what works best. > > If we take the dictionary for example, the "pseudo interface" makes no > sense as not every dictionary has to contain every command, we want them > to evolve. On the other hand, there a few entries every dictionary has > to contains like, > > - mS > - mW > - unknown > > etc. It certainly makes sense to check their availability on a central > place manually, without falling back to pseudo interfaces. > > Side note: > I know what you mean by mentioning SOA and REST and you probably know > this paper, emphasizing your point: > http://martinfowler.com/articles/consumerDrivenContracts.html > However, strictly speaking, I'm not sure that REST itself comes with > less formal interfaces by definition, or does it? > > Cheers > Jonas > > > > On 03/14/2010 12:04 PM, Claas Thiele wrote: > > sorry for bringing up all these things on one weekend. I am feeling, Molybdenum needs some refreshment the very next > > time. It's already starting with all the ideas around: cloud integration fluent interface and ajax spidering supporting > > the old idea of web ui coverage. > > > > I thought indeep about the idea of bringing interfaces to javascript after using javascript libraries like jquery more > > intensivly. > > And seriously, more and more I am going to dislike interfaces in javascript. The provider-consumer contract is built a > > different way in most js frameworks and this is making development and evolvement of frameworks and code based on > > frameworks much more easier from my point of view. > > > > Just an example: > > > > In jquery, a plugin needs an element as parameter to be constructed. This parameter can be a DOM element, a jquery > > object or a selector string the element is derived from. > > > > Following the interface idea strictly (interfaces defining the contract of usage), we have to provide 3 functions or we > > have to find a way to form an abstract element object with concret implementations of dom element, jquery object and > > selector to be passed to one constructor function. Thats all adding much lines of code and less value in my opinion. > > Evolving the plugin furthermore, all this lines of code have to be adapted. > > > > Instead of this, I like the idea to put some checking code into the constructor, which has to find out if the passed > > parameter fulfills the requirements and if not, try to adapt the passed one or to throw a sensible message to the user > > (developer). > > > > Looking around, how the usage of languages with strict interfaces like Java is evolving, we can see that several > > techniques are sourrounding the language feature effectivly. There is the IAdaptable in the Eclipse framework, there is > > all this bytecode manipulation to adapt things to each other not fitting the strict way. Latter one will effectivly > > making some strict checks done during development and compile time a useless thing, because lateron they are violated. > > > > Same evolvement I see in the world of services. All this strict stuff like WSDL SOAP and so is used with a bad mood of > > being inflexible. Instead of this REST APIs are coming up more and more. > > > > WDYT? How should coding style of Molybdenum evolve? > > > > > > Claas > > > > ------------------------------------------------------------------------------ > Download Intel® Parallel Studio Eval > Try the new software tools for yourself. Speed compiling, find bugs > proactively, and fine-tune applications for parallel performance. > See why Intel Parallel Studio got high marks during beta. > http://p.sf.net/sfu/intel-sw-dev > _______________________________________________ > Molybdenum-developer mailing list > Mol...@li... > https://lists.sourceforge.net/lists/listinfo/molybdenum-developer |
From: <je...@ma...> - 2010-03-15 12:58:50
|
+1 regards, jens Quoting Claas Thiele <ct...@ct...>: > Firfox 3.6 is out a while and adopted well by the users. > Firefox 2.* is end of life for a long time now. > > So it's time to stop supporting Firefox 2.* in Molybdenum I think. > Doing so, we can cleanup our code and it would be > easier to use new 3.* features of Firefox without the need to > integrate code for fallback scenarios. > > What about removing 2.* compatibility with the next release (0.7.4)? > > my +1 > > > Claas > > ------------------------------------------------------------------------------ > Download Intel® Parallel Studio Eval > Try the new software tools for yourself. Speed compiling, find bugs > proactively, and fine-tune applications for parallel performance. > See why Intel Parallel Studio got high marks during beta. > http://p.sf.net/sfu/intel-sw-dev > _______________________________________________ > Molybdenum-developer mailing list > Mol...@li... > https://lists.sourceforge.net/lists/listinfo/molybdenum-developer > |
From: Claas T. <ct...@ct...> - 2010-03-14 11:04:22
|
sorry for bringing up all these things on one weekend. I am feeling, Molybdenum needs some refreshment the very next time. It's already starting with all the ideas around: cloud integration fluent interface and ajax spidering supporting the old idea of web ui coverage. I thought indeep about the idea of bringing interfaces to javascript after using javascript libraries like jquery more intensivly. And seriously, more and more I am going to dislike interfaces in javascript. The provider-consumer contract is built a different way in most js frameworks and this is making development and evolvement of frameworks and code based on frameworks much more easier from my point of view. Just an example: In jquery, a plugin needs an element as parameter to be constructed. This parameter can be a DOM element, a jquery object or a selector string the element is derived from. Following the interface idea strictly (interfaces defining the contract of usage), we have to provide 3 functions or we have to find a way to form an abstract element object with concret implementations of dom element, jquery object and selector to be passed to one constructor function. Thats all adding much lines of code and less value in my opinion. Evolving the plugin furthermore, all this lines of code have to be adapted. Instead of this, I like the idea to put some checking code into the constructor, which has to find out if the passed parameter fulfills the requirements and if not, try to adapt the passed one or to throw a sensible message to the user (developer). Looking around, how the usage of languages with strict interfaces like Java is evolving, we can see that several techniques are sourrounding the language feature effectivly. There is the IAdaptable in the Eclipse framework, there is all this bytecode manipulation to adapt things to each other not fitting the strict way. Latter one will effectivly making some strict checks done during development and compile time a useless thing, because lateron they are violated. Same evolvement I see in the world of services. All this strict stuff like WSDL SOAP and so is used with a bad mood of being inflexible. Instead of this REST APIs are coming up more and more. WDYT? How should coding style of Molybdenum evolve? Claas |
From: Claas T. <ct...@ct...> - 2010-03-14 10:29:28
|
Firfox 3.6 is out a while and adopted well by the users. Firefox 2.* is end of life for a long time now. So it's time to stop supporting Firefox 2.* in Molybdenum I think. Doing so, we can cleanup our code and it would be easier to use new 3.* features of Firefox without the need to integrate code for fallback scenarios. What about removing 2.* compatibility with the next release (0.7.4)? my +1 Claas |
From: Claas T. <ct...@ct...> - 2010-03-14 10:28:51
|
currently the ui of Molybdenum is based on XUL. Looking around, I see some movement away from XUL for addons. There is Firebug, where the ui is build with HTML nearly entirly and there is the upcoming JetPack API which is enforcing developers to use HTML for developing addons. And seriously, it is a hard job to develop with XUL - no test tools, no framworks, no broad documentation and experiences in the community. Thinking about new planned features like the fluent language interface with sophisticated autocompletion, I would like to implement such things using common frameworks like jquery or others. Lets start a discussion from here about if it is a good idea to use html for the user interface of Molybdenum and if so, what frameworks want we use and how it is simplifying the testing of Molybdenum ui. What other implication will popping up eventually (build system, ...) Thanks Claas |
From: Jonas K. <jon...@us...> - 2010-03-08 20:27:04
|
Hi guys, I've finished a draft of the core part (ExpressionBuilder.mmod). A testcase demonstrating the behaviour by converting our internal molybdenum test report back and forth is called 'fluentReportTest.js'. See its output here: https://porter.flavia-it.de/hudson/job/Trunk%20XPI/55/console e.g. Click the element with id or name 'oneAlert' As you can see, the commands not present in the dictionary (english.mmod) are written like this: Execute command '$C'[ with target 'T'[ and value '$V')]] e.g. Execute command 'verifyAlert' with target 'Store Below 429 degrees F!' Feel free to play around with it, for example by adding more dictionary entries to english.mmod. Next steps: - adding a significant amount of dict entries, start with the most common commands - introduce a prefs dialog for report generation: switch between fluent and classical view, chosing language - generate report html using fluent english language - come up with a dialog for adding commands in a fluent manner - add german language support Cheers, Jonas |
From: <ct...@ct...> - 2009-11-10 08:06:49
|
> Had a look at the report on Hudson: > https://porter.flavia-it.de/hudson/job/Trunk%20XPI/39/artifact/molybdenum-xpi/target/site/molybdenum-ff36/report.html > > Seems to work great from a functional point of view. I like the fading > in of additional content (comments) when clicking the small info button. I'd made it appearing faster. Jens,you wasn't happy with this anim before, how is it with the anim going faster? > How about extending that concept further: > I would not hide any infos the user might want to sort for like overall > test duration, because we might want to add that feature in near future. > But the re-run button, the anchor link and the tags would be good > candidates IMHO. > Depends, how the report is used. If the daily usage is catching deep links and putting them in a bug tracking system, it would be annoying to have an additional click to be done. We should collect all the feature requests and we should collect users feedback about report usage. > Also removing the common parts out of the test suite location > (file:///data/hudson/jobs/Trunk...) would remove visual noise Definitly, it is the prerequisite for AUTO-16 (http://molyb.org/jira/browse/AUTO-16) as well. > and what > about even coloring the toggle button to red or green depending on the > test result in order to get rid of the additional red cross resp. green > checkmark on the level of test suites, tests and bricks and only keep > them on command level? Something like an attached icon as you have in Netbeans and Eclipse (composed Icons)... > > When we've reduced the content to the bare minium, we could then think > about adding additional style if needed. > > Besides, how do we make the re-run button work on hudson? Moly would > need to be given an http url during maven build, so that this one will > appear in the report instead of the links based on file:// protocol. > Ideally we would take the svn url of the associated schedule files, > being checked out by hudson to the local working copy. What about > extending that idea to a CI mode of the molybdenum maven plugin > accessing the SCM section of the POM? It's in already in a way (http://www.molyb.org/confluence/display/molyb/Maven) : relocateto. It has to be configured only. But the testserver has to be there as well and unittests cannot run from a remote server. Using the scm stuff is a nice idea. Since this is not working in all cases in Maven, i.e. tagged versions, I would prefer the manual configuration. > > Regards, > Jonas > |
From: Jonas K. <jon...@us...> - 2009-11-09 22:59:25
|
Had a look at the report on Hudson: https://porter.flavia-it.de/hudson/job/Trunk%20XPI/39/artifact/molybdenum-xpi/target/site/molybdenum-ff36/report.html Seems to work great from a functional point of view. I like the fading in of additional content (comments) when clicking the small info button. How about extending that concept further: I would not hide any infos the user might want to sort for like overall test duration, because we might want to add that feature in near future. But the re-run button, the anchor link and the tags would be good candidates IMHO. Also removing the common parts out of the test suite location (file:///data/hudson/jobs/Trunk...) would remove visual noise and what about even coloring the toggle button to red or green depending on the test result in order to get rid of the additional red cross resp. green checkmark on the level of test suites, tests and bricks and only keep them on command level? When we've reduced the content to the bare minium, we could then think about adding additional style if needed. Besides, how do we make the re-run button work on hudson? Moly would need to be given an http url during maven build, so that this one will appear in the report instead of the links based on file:// protocol. Ideally we would take the svn url of the associated schedule files, being checked out by hudson to the local working copy. What about extending that idea to a CI mode of the molybdenum maven plugin accessing the SCM section of the POM? Regards, Jonas |
From: Jonas K. <jon...@us...> - 2009-11-09 21:02:19
|
>> I would propose to mark Molybdenum 0.7.3 as 3.6 compliant on AMO. >> >> my +1 >> >> Please vote. +1 Jonas |
From: Jonas K. <jon...@us...> - 2009-11-08 19:41:25
|
Claas Thiele wrote: > Thanks for testing Jonas - and for being strongheaded. > Please, try again the trunk version. Still does not work for me, same with FF 3.5.5 When manually generating the report, log says: Writing report to file 'C:\TEMP\newreport2\report.html.xml' Component returned failure code: 0x80520001 (NS_ERROR_FILE_UNRECOGNIZED_PATH) [nsILocalFile.initWithPath] Regards, Jonas |
From: <je...@ma...> - 2009-11-07 23:29:36
|
i've tried it last week, looked good for me, so here is my +1 regards, jens Zitat von Claas Thiele <ct...@ct...>: > I've added Firefox 3.6b1 to the testsuite, Hudosn environment is updated > Molybdenum is out of the box compliant with Firefox 3.6. > As 3.6 has readyState support included, I switched off homemade > readyState support for Firefox versions greater and equals than 3.6a1. > > I would propose to mark Molybdenum 0.7.3 as 3.6 compliant on AMO. > > my +1 > > Please vote. > > > > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Molybdenum-developer mailing list > Mol...@li... > https://lists.sourceforge.net/lists/listinfo/molybdenum-developer > |
From: Claas T. <ct...@ct...> - 2009-11-07 21:47:56
|
I've added Firefox 3.6b1 to the testsuite, Hudosn environment is updated Molybdenum is out of the box compliant with Firefox 3.6. As 3.6 has readyState support included, I switched off homemade readyState support for Firefox versions greater and equals than 3.6a1. I would propose to mark Molybdenum 0.7.3 as 3.6 compliant on AMO. my +1 Please vote. |
From: Claas T. <ct...@ct...> - 2009-11-07 08:57:44
|
... and found the problem with chunk writing. getElementsByName() returns a NodeList. This NodeList is *not* immutable, it might change, if the document it comes from is changed... Thanks for testing Jonas - and for being strongheaded. Please, try again the trunk version. > ... and works fine for me with FF 3.5.5 under Win XP. > > > > > > > > Hm, still not working for me using FF 3.1 beta 2 under windows. > > > > Would it be possible to use the released version, 3.5.4 or later? > > > > I'll give the 3.5 Firefox a chance today. > > > > > Directory "report.html_files" is present, but no metafiles inside > > > besides the screenshots and the source code. So it's only the > static > > > data (images, css and javascript) which is missing. > > > > > > One more thing: I tried to create the report manually via the > menu, no > > > maven, ant or commandline involved. > > > > Exactly what I'm doing... > > > > > > > > Will perform more tests to give you feedback, right now the only > error > > > message I see says something like "Component returned failure > code ... > > > initWithPath" without any further hints. > > > > > > Regards > > > Jonas > > > > > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ Molybdenum-developer mailing list Mol...@li... https://lists.sourceforge.net/lists/listinfo/molybdenum-developer |
From: <ct...@ct...> - 2009-11-06 12:27:18
|
collapse and expand is much, much slower than before ;-( |
From: <je...@ma...> - 2009-11-06 09:09:16
|
latest svn works fine with xp and ff3.5.x for me, too. kind regards, jens Quoting "ct...@ct..." <ct...@ct...>: > > ... and works fine for me with FF 3.5.5 under Win XP. > >> >> >> > Hm, still not working for me using FF 3.1 beta 2 under windows. >> >> Would it be possible to use the released version, 3.5.4 or later? >> >> I'll give the 3.5 Firefox a chance today. >> >> > Directory "report.html_files" is present, but no metafiles inside >> > besides the screenshots and the source code. So it's only the static >> > data (images, css and javascript) which is missing. >> > >> > One more thing: I tried to create the report manually via the menu, no >> > maven, ant or commandline involved. >> >> Exactly what I'm doing... >> >> > >> > Will perform more tests to give you feedback, right now the only error >> > message I see says something like "Component returned failure code ... >> > initWithPath" without any further hints. >> > >> > Regards >> > Jonas >> > >> > |
From: <ct...@ct...> - 2009-11-06 07:29:32
|
... and works fine for me with FF 3.5.5 under Win XP. > > > > Hm, still not working for me using FF 3.1 beta 2 under windows. > > Would it be possible to use the released version, 3.5.4 or later? > > I'll give the 3.5 Firefox a chance today. > > > Directory "report.html_files" is present, but no metafiles inside > > besides the screenshots and the source code. So it's only the static > > data (images, css and javascript) which is missing. > > > > One more thing: I tried to create the report manually via the menu, no > > maven, ant or commandline involved. > > Exactly what I'm doing... > > > > > Will perform more tests to give you feedback, right now the only error > > message I see says something like "Component returned failure code ... > > initWithPath" without any further hints. > > > > Regards > > Jonas > > > |
From: <ct...@ct...> - 2009-11-06 06:15:33
|
> Hm, still not working for me using FF 3.1 beta 2 under windows. Would it be possible to use the released version, 3.5.4 or later? I'll give the 3.5 Firefox a chance today. > Directory "report.html_files" is present, but no metafiles inside > besides the screenshots and the source code. So it's only the static > data (images, css and javascript) which is missing. > > One more thing: I tried to create the report manually via the menu, no > maven, ant or commandline involved. Exactly what I'm doing... > > Will perform more tests to give you feedback, right now the only error > message I see says something like "Component returned failure code ... > initWithPath" without any further hints. > > Regards > Jonas > |
From: <ct...@ct...> - 2009-11-06 06:12:17
|
http://www.molyb.org/jira/browse/INTE-20 > [ERROR] BUILD FAILURE > [INFO] > ------------------------------------------------------------------------ > [INFO] Compilation failure > c:\Projects\molybdenum-trunk\interpreter\src\main\java\org\molyb\interpreter\CommandlineParser.java:[91,58] > cannot find symbol > symbol : method copyOfRange(java.lang.String[],int,int) > location: class java.util.Arrays > > Arrays#copyOfRange is Java 6 only. Is this intended? If no, I'll lower > the target version of the compiler plugin and change the code above. If > yes, I'll include the maven-enforcer-plugin to enforce the minimum > required java version. > > Regards, > Jonas > |
From: Jonas K. <jon...@us...> - 2009-11-05 22:27:51
|
[ERROR] BUILD FAILURE [INFO] ------------------------------------------------------------------------ [INFO] Compilation failure c:\Projects\molybdenum-trunk\interpreter\src\main\java\org\molyb\interpreter\CommandlineParser.java:[91,58] cannot find symbol symbol : method copyOfRange(java.lang.String[],int,int) location: class java.util.Arrays Arrays#copyOfRange is Java 6 only. Is this intended? If no, I'll lower the target version of the compiler plugin and change the code above. If yes, I'll include the maven-enforcer-plugin to enforce the minimum required java version. Regards, Jonas |
From: Jonas K. <jon...@us...> - 2009-11-05 22:22:23
|
Hm, still not working for me using FF 3.1 beta 2 under windows. Directory "report.html_files" is present, but no metafiles inside besides the screenshots and the source code. So it's only the static data (images, css and javascript) which is missing. One more thing: I tried to create the report manually via the menu, no maven, ant or commandline involved. Will perform more tests to give you feedback, right now the only error message I see says something like "Component returned failure code ... initWithPath" without any further hints. Regards Jonas Claas Thiele wrote: > Not sure about the effect you described. > > But I'd booted my Win now and found another issue, fixed it (r1316) and > it is working now under Win using Firefox 3.0.x (on my machine) > >> Very curious to try it out, but still not working under windows. The >> chunks are written now, the base html file as well, but all accompanying >> files like css, javascript, images etc. are missing, albeit the folder >> named "xyz_files" is present. |