Datasets:

pheepa
/

jira-commentaries-mlm

comments stringlengths 0 36.6k
Birt Jetty and did not change. verified in JBDS Alpha
I agree option is the way to goI remembered after you brought it up in option above that currently the code is already duplicated in the tools jar. It is assumed that the network server will be running alongside the engine so classlister filters out the info classes fromThese classes are fairly stable. is the only one that has seen any change and that was to add a new identifier for JDK. That change was made in. The potential for problems with shadowing etc. are already there with the duplication in I suppose but I think the potential for problems is minor and the problem in a mixed environment would be fairly obvious to spot. I am coming into this issue late so might really be misunderstanding things but I dont like this idea of importing engine classes into the client no matter how stable we think they are. Doing this means that every line of code has to be evaluated for impactThis strategy was discussed on the list previously and the concensus I thought was that for client a copy would be put in the shared directory and at such time as we had a code sharing solution the files would be combined. If the files are really that stable it shouldnt be hard to resolve differencesAlso the description of this bug does not describe the circumstances under which a user might see this issue. I seem to recall that it imposed some sort of required classpath order. Is that correctIf so I think it is a regression and needs to be moved up to Blocker for. Hi Kathey I was wondering when you were going to join in on the funThe thing is that these classes have been imported into the client jar for as long as I have been on the project. So I cant imagine it can be a regression to put it back the way things wereIf you dont like the solution of reverting back to how things were can you suggest another oneThanks to answer your question yes my understanding is that the bug reveals itself when is in front of. Please feel free to upgrade to Blocker. I was working on implementing Solution as described above but Im on hold now until we can finish our discussion of potential issues around this solution and what the best alternative might be if any. Committed revision
PR applied. If youre able to build now please close this issue. reopened to fix Fix Version
. but I can do have such a thing without getting deployment exception nor any errorwarn in the logs. should have been. but I can do have such a bean with no Remove or Destroy without getting deployment exception nor any errorwarn in the logs. state correction . Yes absolutely
This is bad. Moving to so we at least give a to figure whats up here. Giving to Izaak to check tried to replicate the problem in trunk and couldnt. I attached the patch that shows what I did. I see that several things changed in trunk and one of them fixed this issue. Ill try to run my dev environment on trunk and see how it behaves. Feel free to apply this patch on trunk if you think it really adds robustness to the unit tests. Closing this issue as WONTFIX because there will be no more releases now that has been released.
I am closing this out it seems to be working nowregards you tested on weblogic?
Please review and we can discuss timing and impact. Actually thats hlink that has include attribute hcommandLinkhcommandButton dont have it. Nested fparam tags should be used.
Jasperthanks for the hint regarding the validator configuration DTDI just added it for Please check it outRegarding the improvement Id like to hear what other think about thischeers it would be nice but its hardly an issue for me. Still I agree it is something that should be added. thanks. This issue wil be solved with the suggested approach as well. Attached is a patch against trunk of XWIt should work out of the box against XW as wellPlease review and tell me what you thinkcheers in SVN as r patch looks good to me and I think this solution is the best of all worlds!
again sorry this is not a proper patch reference diff between patch and NET versions of with a minor change get now defers to get if report is null.
Rolling over to reduce workload for pre would be good to be able to have more than just three setups. My problem is that I am writing on Windows for Linux and I need to configure different paths for the different oss. Also I am using Google Maps which requires a new key for each machine. Somehow I would love a kind of inheritance model. Where the child would override the parent. Lets say my development machine is called duck and my production machine is called Server. If I was booting up grails on my development machine the bootstrap would first load it was on the production server then it would bootstrap was to boot up on any other environment where there was no configuration for that machine then it would just load. PS I dont like the idea of as a naming convention because we are swapping names around already from dev to development in was just looking for how to do this. I have different databases for dev test prod and the dataset I need to test some of the functions is rather large so I dont want to start over from scratch every time. Moving issues that arent going to make it into to votes to make a blocker
Good catch MarcoThis is a type that somehow escaped my automated renaming process from Java convention to CThis is now fixed in ver and. George.
This patch should be applied at the bundleplugin directory not the root of the felix trunk. Fixed in r and deployed a new snapshot of the
URL must be valid URL if you do not know the URL do not provide one then a default one will be used. I think this can be rejected. Warren Gibson URL is a valid URL in ER and queries run fine. It started failing in ER. There is a being thrown but its getting masked by the NPE because were attempting to get the cause which must be null. So more than likely the underlying problem is a userpassword issue. We should still fix the problem with the exception though. updated the connectionimpl to not unwrap the source exceptions
Here is the WAR that contains the service endpoint with so you can recreate the issue on JBOSS GA. Please leave the scheduling and priority setup to the committers. Thanks. Can you please give me a time estimate of when this issue will be resolved so I can use this information to my PD management? Otherwise we cant advertise that our web services will be able to use. Native SOAP stack doesnt support JAXWS any more
Simone Daniele what do you think? Does it seem feasible. Simone Daniele do you want this done for? Can you help in that case. I am not sure well have enought time for this we are pretty much booked up until the end of the would probably leave it as unspecified. I saw Daniele was working a little on this topic. Shall I still keep it scheduled for move it to to? Open to suggestions. Changed to since ECW Plugin mechanism for GDAL isnt ready yet. Only support for windows based on plugin mechanism is actually support will be provided afterwards. Mass closing all resolved issues that have been closed for more than one month without further comments
We need to conform to with regards to what constitutes a Valid Email Address. Last time I looked Im pretty sure an email address of all spaces is considered valid. partially addressed with in a space check on username which should resolve the rest of this issue
Change made as requested. I have resolved this issue as the requested change has been made and the corrected version of the book is now publicly available. The change was made I guess we can close this now.
A check for the minimized window state is performed and no rendering occurs if the portlet is minimized. No other custom window state handling at this point. Backporting to branch
Sorry yes I missed that and I agree we do need it. From my observation the input vectors are immutable. They will be benefit from the sparse vector implementation. For centroids the vector works great. This adds back now as the parent of both sparse impls and extendingThere are certainly optimizations which can still be done and again no work has been done to dig through all the places where are created and choose whether it would be better to haveThis patch also has the fix for. bumpHey Grant have you tried out this version of the patch? Looks like it could resolve and all in one go if its looks good. Grant I can take over this one since its my changes. Ugh ok Ive revived this patch and it compiles again after all the math vector changes but its not passing all tests. Ill post it here as a progress indicator tomorrow anyways. Patch brought up to date with trunk again fairly invasive because it kills replaces with the two subclasses optimized for different use cases. In basically all cases throughout user code but this is probably not optimal. Cases where the vector being used is treated in a way it is more efficient to use but I did not hunt down all those situations. You want me to try this and commit it. It all passes for me. I had to delete another instance of in utils does that make sense? I admit I still had patch problems here but thought I manually fixed them allI can handle issues like copyright headers. bq. I had to delete another instance of in utils does that make sense? It does I moved it and deleted the old one but it must have snuck back into the patchIf youve got it cleaned up and ready to commit sure thanks!
Fixed in trunk svn revision could you please commit this fix also in the branch and close this issueIf you think this should not be included in please set the fixed version to and close this issue. Thanks. This is a weird glitch. I suppose the only reasonable explanation for this would be me accidentally pasting random garbage after opening the file for manual inspection after running the SLFJ conversion script. Sorry about thatMerged to in revision .
Just lazy. The problem is that the servlet mappings in are removed from the path before it is matched by the. You need to explicitly ask for the full path to be used if thats what you want. I can push the change to that branch if you want. Done. Not the most pretty use of MVC yet but it works and reads better I think. Closed with.
If forceCanonical is true subsequent accesses of vfssymlink resolve to othertmp which isnt found in either the permanent or real cacheIn what way are you accessing vfssymlinkAs I have again hard time reproducing thisWhat if both tmp and othertmp are added as permanent roots. Has the change for this been pulled in so that we can resolve this in the next release?
Simple factory that uses s feature to create a Socket with a connection timeout. Added in.
Issues for ant main are currently. Main build file checks for pixel bender. Easy fix. Unable to compile the odd pixel bender files. Can copy compile files from elsewhere to fix and make pixelbenderclean tasks do nothing. Spark AIR line renderMode fails to compile. Easy fix. fails to compile. Can be conditionally removed from. and fail to compile. Can be conditionally removed from. Javascript target fails missing code. Can remove javascript target from framework file but probably better way of fixing this. Why is code missingWith above modifications ant main and ant succeed although may be some modifications required. With above modifications all sub targets of ant binaryrelease except ant checkintests work. Gziped archive of all compiled pixel bender files. Fixed issued above so ant main and ant work. Fixed failure the test failure looks like a Flash Player issue. ignoring the Flash Player error and the issue with the sort tests targets all pass except asdocs same issue in generating docs for which doesnt exist in AIR. Can generate source and binary releases by running selected targets asdocs and checkin tests still fail.
My mindI only test against null not against an exceptionSo let it return null. Bug has been marked as a duplicate of this bug. . Patch applied thanks IvanCan you please cross check and closeCheersMarcus
Created links for ajaxListanerloadBundle jsFunctionhtml componentControl. done
Remoting will need to get polished and documented for the next release. At the moment it is unlikely we can progress remoting before the release
This must be fixed in a CR. The ESB issue has been moved out of our release for the platform and back into the project release. We are currently targetting embedded console at ESB. Temporarily reopening to update release note info. Temporarily reopening to update release note info.
Attachment has been added with description The patch as has been added with description another logging applied. Im closing Simons issue here but Peters concerns can stillbe discussed on this bug page or on the list directly. transition resolvedfixed bugs to closedfixed
Patch to build prefsnio dll to. instead of jrebin. Thanks OliverPatch applied to makefiles in NIO and PREFS at repo revision rPlease check that the patch was applied as you expected. Looks good thanks Tim. Verified by Oliver.
Fixed by miniRTE which is displayed inline. Can we assign this issue to some new release? It was reopened so it was not handled within the release and probably will be in some future one. I just dont think we should have unresolved tasks in Changelog.
I think the problem is a is a stopword and strips stopwords? Try using instead. Yes the reason is the Analyzer. With there is no such problem.
patch that fixes the test. I dont think we need to change here. to commit. this is better i didnt want to really change just to fix the test. its a test bug pretty straight forward patch. I committed that to trunk. Thanks for reporting
and conf is also overwritten. went for a more simple solution called grails upgrade for nowwe will look at something more advanced once the plugin stuff is done
Made some simple changes to make it visually more clear that this is an overlay and give it a cleaner look. Theres also the option of hiding all the other gadgets and page content but I liked the idea of keeping it an overlay cluing in the user that youre still on the same page all has not been lost etc. etc. Definitely open to alternatives. Kristen Stan made a change to the css files and and are now removed. You will need to add the CSS changes to the less files. That was bad timing. Ill transfer my changes to use the less files. Reassigning to myself. Kristen has been down for over a week now. Ill review your patch and apply it. Created add function to handle full screen widget view instead of altering addOverlay. Patch applied thanks Kristen!
Cant we simply replace the current map for this new versionSome code will break but. Good idea it will reduce confusion in the long run and help embed the idiom of putting the function last so that it can be used with dodef.
Collision with. JDYa. i did not see that. Thanks for pointing it out. Same as
Pushing to to to to from the stat gathering implemented for the sysadmin set and set and exposed the nulldistinct counts on. see the reference and the developers guide for more on runtime updates of metadata.
fixed component and services config plugin too
This patch illustrates my idea. However it does not pass all the tests and an elegant fix will depend on. submitting patch. Moving the status back since the patch is not ready for commitYongqiang is working on this issue right now. I will reassign. then why not use ab for display and internal names etc in type defs and thrift. You can add displayName field to expr and use that field instead of column name while printing out the expressionIMO using translated column names instead of position names actually be more confusing if two expressions are similar. For this issue lets use col col firstLets fix the display name problem in. now works. The modifications made added an array of output column names to Operators like select groupby join etc removed plan regeneration code in instead of using integer as column internal names now use col col I tried to apply the patch ran into a lot of merge conflicts can you regenerate the for delay. is a new one against truck and intergrates with Zhengs suggestions. Namit thanks very much for the valueble feedbacks. Will work on them. Comment is a serious mistake. I usally did that for testting on eclipse and i forgot to restore it when i was trying to make the patchjoin and join are reporting error now will check and repair. Mostly looks good but I had some questions forOnly reduce is getting patched below the Select what if a groupBy is present doesnt the column names of group by also need to be patched or the assumption is that it is alwaysa subset of the select columns and therefore we dont need to do anything. the assumption is that it is always a subset of the select columns and therefore we dont need to do anythingyes. And a select operator is inserted before the groupby operator. When I ran the tests all tests failedAlso for I got errors in input join and which patch you applied. When I ran the tests all tests failedsorry my fault. does not include a overwrite test forAlso for I got errors in input join and join input is ok in my test. join is a newly added testcase and the previous patch does not include it inOne wired thing about join is that if i run the test with overwrite specified and every thing is ok. But if no overwrite specified an error is reported. So i will try figure it out and upload a new patch. looks good running tests right on it right now. please wait a minute. Added the new testcase but join does not get diff in my local. There is some behavior. The output of is different if run or if run as part of the which explains the diff. Namit Thanks for correction. Upload a new patch. I guess the diffs came from the order of file output operators column names. And the new patch sorted field schemas according to their names. Hope this can pass the tests. Again it passed in my local. running the. Thanks He
A simple refactormove. Tx for the review SidUpdated patch movingAlso needed to fix binyarnLets move the protocol names issue to another ticket. Trivial. bq. core tests. The patch failed these unit tests in bq. This is not caused by this ticket. Filed. Just committed this to trunk andThanks for the review Sid.
Punting issues from RC to final. Have not seen this failure in a long time.
this fixes the problem. please disregard previous attachment. it wasnt right. this one is correct. Applied Patchthanksdims
Fixed with other windows specific issues. Closing issue
Committed. Thanks Chris. Bulk close for release
Agreed although I would need a clearer definition of activeFully agreed on the other points. counts you me Andrew and Dirk at least in the top contributors. I guess that is sufficient. Then propose to close this one too in a few days. In addition weve been able to grow the amount of committers by two during incubation using Apache processes. Can we close this. for closing this one.
There is a bug in xmpbox that does not manage a unique rdfDescription containing both namespace fields and namespace definition. The last version of xmpbox fixes that problemThe attached document is now considered as valid
Scheduling to the next release for an for providing the test case. There are two problems Rewrite uses web filter for request processing and doesnt react correctly on filtering when it is invoked after forwards request repeating filter chain several times which again brokes which does activate manager and its scopes for investigating. Im looking forward to a solution.
Thank you for the comment Dan. I see your point regarding and hence leave it where it is. I will continue with this patch as described but give the community some days to respond to whether or not it can be included in this release. I am uncertain about this release note so it would be really great if someone reviewed itRelease note attemptSummary of ChangeMost authorization failures have new error codesSymptoms Seen by Applications Affected by ChangeIn the previous release authorization failures had error codes x and . In this release most of these errors have new error codes. The code changes are H I and J are now . are now is now B are now. Only the error codes have been changed error messages are not affectedIncompatibilities with Previous ReleaseApplications that rely on authorization error codes may failRationale for ChangeThe old error codes violated the SQL standard. The new error codes are correctApplication Changes RequiredApplications that rely on authorization error codes must be recoded to expect the new codesHi JrgenI think the Release Note looks good to me thxI glanced over the actual patch and have a comment thoughI see you have not modified any of the translated messagesIt looks like a tremendous amount of work but either the same thingas happens for English should happen for all other translatedmessages or the affected messages in the other languages should beremoved altogetherMyrna. Forgot to say modifying the translated message files can be done as a patch or subtask. Thank you for reviewing the patch Myrna. Ill make a subtask and start working on this straight away. Committed revision . KnutThanks for committing the patch. Closing issue as it has been tagged as fixed since May
Problem attaching the diff file. Post to instead. Fixed in nightly build.
Yongqiang can you post it to review board? Its much easier that way. . Thanks Yongqiang!
Reclassifying as a bug since this is actually blocking a desired use case for Pageflows. I think this can be closed as per Jacobs comment.
I face the same issue. This totally prevents site there a workaround. This problem starts as of version of also in Maven . A temporary workaround is to downgrade to. Is there a workaround for Maven ?
Removed the dead code. I also fixed some typos and java warnings. committed the patch. Thank you Brandon for the review.
This script is tested against DB V. Please help me review and apply it. Thanks. Moved to as there is no platform dependency on this work. The schema tool already covers this. I have followed up with jdc and discovered that this has only been partly addressed in the platform release. The schema tool has been updated but the platform does not include the additional createdrop back to the CP so that we can deal with this aspect. DB configuration added in revision .
Attached the patch file that changes all of these files plus added the new and files. There were several source files without any header in them and these were corrected. Committed the patch verifying that all unit tests pass. Reviewed all these issues and marking them as closed.
Can you post a full stack trace. Changing to improvement I should be able to do it this weekend if not Ill bump to. Dave any news on this or should I bump it. added improvement to trunk
Now that work is underway with AS all previous community releases are. So all issues not directly assigned to an AS release are being closed. EAP is our supported enterprise version of AS and you can file your issues against any release during its very long support window via CSPRHNIf you believe your issue is still relevant to AS then please verify it and reopen. Due to feedback from the community I have split the AS and AS projects and reopened all unscheduled AS issues that are a year or less old. This will make it easier community members to find and work on themFuture releases beyond can be done provided a community member steps up to coordinate them.
This is the list of issues that we want resolved for posts it cvs branch.
Just checked with a debugger if the gtype of the geometry inside Oracle is a Polygon will be built. If the user gets a Multipolygon instead its because the database contains geometry of type theres no much we can do about this. Mass closing all issues that have been in resolved state for months or more without any feedback or update
Sorry for potentially not making any sense in the description. By main Javascript directory I mean the directory for the Javascript sources used by Jetty when invoked via mvn jettyrun. Id like to be able to make changes to those files in srcmainjs and have Jetty automatically pick them up so that I dont have to mvn jettyrun for each change I makeAs a side note itd be super if I could edit the description. The JS Import plugin is not designed to work the way that you want it to. JS Import will pick up from the targetclasses folderJetty is automatically configured to pick up from the targetclasses folder as changes are made. So I have to directly edit the deployed code in targetclasses even though the code by design is in srcmainjs? That makes no sense at all. Moreover this not the way other plugins of this type are designed. The GWT maven plugin for example allows you to make changes to your Java code in srcmainjava and automatically changes while Jetty is runningPlease tell me Ive misunderstood something and it isnt this bad. Hi TravisMy recommendation is to perform an mvn or mvn after your edit. This is certainly a manual step at the moment and were working on tooling to make things seamless in this regardOn processing sources your imports will be resolved and your code linted. The product of your code with imports resolved including html with the required script statements will be placed in the target folderhope that this helpsBTW this kind of question would be better asked on the forums rather than being raised as an issue. There may be more eyes on the forum and Im on vacation at the momentKind regards is all I need to do then at least I can rig up vim to do that for me on write. I think this is a valid improvement request which is why I filed it here. You might want to integrate with vim using mvn instead. Jetty is configured to present both targetclasses and target under the one context pathPerhaps the improvement request needs to be with respect to enhancing the documentation to show how this can all be used in the context of vim. I think that would be useful. Perhaps you can bed things down and then we can the request with that in mindI appreciate your feedback.
Yes definitely think this is wrong and inconsistent. Static from super should have lower priority. fixed but please comment if you think of any issues with provided perfect. Thanks.
This patch should fix the problem. Handler for case when the substring is longer than the specified string has been added. Thanks PaulexFixed in LUNI module at repo revision Please verify that this fully resolves your problem. the fix looks fine. Thank you Sergey and Tim. Verified by Paulex.
setting the fix to rc although i plan to start on it right away. theres no certainty that it will be finished by the end of the week so that is why I moved this issue. Whats the status of QA for ASWhich versions of it are going to be tested. postponed almost m currently checking out if i can use cargo. That should cover a lot of serversversions.
Some test cases and a minor fix to. A patch for the issue which also sets content type to textplain and character encoding to for propertiesI am not certain that it is correct to set all of these in the as the issue I observed was with the. So I would be grateful if you could carefully review the patch before applying. Irony has it that the encoding of the attached tests was incorrect. Trying again. Sounds to me like an easy fix we might want to include into the for checking and applying the path Carsten. Ive the changes you applied and confirm that my patch wasnt actually setting the content length yours does. Yes thanks Ive corrected the two tests.
can be removed now so maybe some of the issues mentioned here have already been resolved. The server module still has a dependency on the and there are still references to the. to so it does not look like anything has changed. I dont rember in what context I was seeing the problem but from a simple static dependency analysis depends on the is now distributed in serverlibA better decoupling of and is addressed here
Not sure what youre getting at having a username in Xircles JIRA and Confluence with a. is valid unless youve seen otherwiseThere are potentially some other bugs that exist we were getting killed with spam because JIRA doesnt really protect against sign up very well. Xircles does because its non standard. Luca Tagliani had issuesThe first account he created has a. in the username and he couldnt login to JIRAThe second account he created does not have a. in the username and he logged into JIRA just fine and dandyIf there are other root causes fair enough
Theres obviously a difference between Syck and in the way they handle the asterisk character. Until the matter is resolved you can work around this issue by enclosing the . rb in quotes or. Both YAML engines should handle that properly. Good idea. Yep that seems to work. Thanks. Fixed on trunk.
Hithx for providing this improvementThe interface changes that comes with will break old implementations. Each application that try to parse encrypted documents and using bouncy castle for his own software wouldnt work any moreSo I think this would be an improvement for the release of the pdfbox. Im not sure this will break applications using encrypted documents with. As I understand the code Bouncy Castle is a purely internal dependency no class from Bouncy Class leaks in the public API of. The user only interacts with standard from. So it should be safe to upgrade the dependency even for the line. For example If an project use the pdfbox and do some cryptographics and stay at bc with this acutal pdfbox no problem will occure. OK If the dependency would be updated to the code of the project will break without update to the new classes. ERROR If the project that using pdfbox try to force the bc version pdfbox will break while parsing encrypted documents trying to access not available objects. in rev the trunk is going so there should be no problem. Thank you ve updated the remaining dependencies in revision
Not as important as other things to get a valid therefore I move it to this is a nasty side effect. We have a real problem with uperlower casing in x. It should be fixed inI dont see any easy solution for step would be to return the attributes the user is waiting using the names he gave to the server in the attributes list to be returned. For instance if the user specify that he wants OBJECTCLASS then he adds OBJECTCLASS to the list of wanted attributes and we return all the objectClasses using OBJECTCLASS. Thanks for testing this out. Emmanuel was dead on with using the case of the attribute ID supplied in the search. Yes I agree that this is not too critical. It just annoys me personally. makes the server seem somewhat flaky if its not adhering to these unspoken expectations. Is it still an issue in. No it has been fixed in. But the problem is still present in enough for me Its one less bug for and one more for fixed it. Did I forgot to mark the issue resolved. This is great Emmanuel. Thanks! Especially because it will already work in the upcoming release. It was not a very important issue but I definitely did not like the big O in the search results. And yes currently the issue is marked as In Progress. Hmmmm. This issue stayed in the In Progess status too long. This has been fixed for some time and especially it works in the.
Verified. All test are passed successfully after the commit ecebdbcfdfeb was applied. Pushed to on Jenkins
David is this actually a regression issue against an earlier version or Castor. Castor or is this a result of you getting that far for the first time. The latter. Ive been stepping through the issues with the snapshot and this is the latest issue I had. I had to spend some time getting a standalone test case that demonstrates this. I havent tested this with earlier versions yet. However I know were not getting this issue with although I havent set up this specific test case to test that version. This schema is from existing code in our regular build which uses. Just realized that you are using the new automatic name conflict resolution mode. For the time being. until we have an educated idea as to why this is going wrong is there a way you could stop using this. Or would this involve writing biggish binding files. In my estimation this would require several hundred binding entries. Unless you somehow needed this to debug the problem I have no intention of going there. Okay thats what I thought. In this case please bear with us for some time being. It looks like you seem to have unearthed something very odd breaking the algorithm implemented in in a very odd way. i just attached a more restrictive patch that only changes the xpath for groups which are CHOICES. The patch also checks if the choice is a global element in this case no renaming should be performed.
Krasimir Would need some more details before we can help you include your whole filesystem then there will be time problems . it possible to see more detailed progressSomething like copying copying. This wont make it into this release. Rob. I just cant understand why one building war for application on Ganymede takes seconds and the same operation on Helios minutes. Unfortunatley nobody has had time to really into this. Delaying. not been able to reproduceif anyone can please attach war project archive setup. In last I cannot create WAR archive for project and I cannot figure it out why. Webtools integration seems to be installed as part of Tools but Project Archives view let me create only JAR archive. I just ran into this issue after an upgrade. Im using the archive builder from ant and can easily test different versions. I downloaded some older versions of the archive builder to determine when the problem has started to occur. I retrieved the following versions and built a fairly complex EAR on a core i seconds seconds minute and seconds minute and secondsSo the problem was introduced between betal and beta of somewhere between mid october and early december . I would be happy to test other buildsUnfortunately I am not able to share the project. still cant reproduce properly anyone that can attach a test case would be good. Its also worth noting that the bundled jars did not change at all during this period. potential new patch for this issue. All the unit tests still pass and speed seems to have improved tremendously on full builds. This should definitely fix the issue of external style users as well as UI users demanding a full build. This will also speed up the register event on a workspace restart which forces a full build to get the model up to synch. I tested a nightly build and can confirm the patch is working. Did not see an increase in speed though with respect to the old versions. Remmelt Krasimir or Rob please could you verify and close this issue? last I cannot create WAR. Project Archives view let me create only JAR archive. archive tools the JIRA was targeted to have been removed from jboss tools long time ago. JIRA is not relevant any more. Closing the JIRA.
Attached patch for reviewNote that the corresponding docbook updates are outstanding. I will attach a patch for these separately. attached docbook for the review comments Keith. Here is a further patch that addresses your feedback. Patch applied. Reassigning back to Phil for the completion of the docs. Patch applied. Reassigning back to Phil for the completion of the docs. attached docbook docbook applied. No comments.
Isnt it related to a previous issue you had opened. In a sense yes but this patch has a much smaller scope the only thing it does is to allow compiler clients to optionally provide a custom class loader for loading AST transformations just as they have always been able to provide a custom class loader for loading compile dependencies. As such it is a very natural extension of groovyc that both fixes immediate problems and simplifies future efforts such as all without changing the class loading behavior of existing compiler clients. there are some small adjustments to the patch I would like to see. may have a different then the or. So I would like to see the transform loader in the same class. Also I would probably name it to compiler or better only extensionLoader. Also per default the laoder should be the same as what get produces. these xnullyz checks I dont like very of course should not get the loader in the constructor but request it from source as it does with the normal classloader. Ill let you apply this patch and amend it as appropriate. I now applied the patch but with a few changes. Firstly I removed the reflective code you added. For a transform to work properly you need access to the compiler classes so there is no point in excluding the annotation classes for the transform in th the compiler. Also I changed transform loader so it has to be a. The loader used for transforms might be a but this loader has the ability compile scripts into classes on its own. But since we are currently in a compilation this must not happen. The result would be a deadlock or a cyclic compilation. that means a file is to be compiled which needs another file which then is compiled but depends on the first file. So if the compiler spawns a new compilation process outside the current one it would have to spawn new ones again and again. So a special loadClass method is used to avoid this cycle. The compiler then identifies classes to compile itself instead of depending on for course that means the gmaven patch has to be changed to use instead of but that shouldnt be a problem. I close this issue as fixed now. should we keep open. This patch makes local transforms work in the presence of a custom class loader. It also includes a thorough unit test. patch applied
I added two methods which allow the programmatic initialization of componentscheck the code for the fisheye component for a working usecaseThe trick is to call the render method and for of all components you should be done with that one. The component works now with the programmatically approach.
Im thinking were going to want the option to enabledisable the save event on every project. We may even want to consider creating a custom project type for Sling projects. We will also definitely need the ability to define paths to ignore. For example you certainly wouldnt want your code to be pushed to Sling every time your target was updated. dan I do agree on both the suggestionsJust bear in mind though that at the moment a file in order to be also saved to the repository must contains jcrroot in the path. robert I agree this could be a problem. I could not reproduce the issue though up to nowAny hint. thanks for your patch Robert. It would be nice to handle folder as well thoughIt does handle them quite well I think. I know there is an issue with empty file and folder thoughAre you sure that the issue above is not isolated to jcrroot folder though? if yes this could be handled for your patch Robert. I have applied it in revision .
Trivial patch to fix issue. Oh pretty good. just committed this. Thanks David
Well the only way to make it work is to make synchronized I if you look in you will see that the implementation created already contains a synchronized block so it is already in the current release. Closing issues.
testcase verifies that the primary group of the user submitting the job is used as pool name. This is done in on of the testcasesNote that the will not be avail to tasks as the JT does not the to works against needs to work against the idea of primary group is good. We should make a new JIRA for it. Committed to and trunk. Thanks Alejandro!
Alex has committed a fix for this in Groovy but the problem remains that we are getting proxies for collection elements. This breaks instanceof and that cannot be worked around in Groovy. So my question is why are collection elements being proxied? If I read the Hibernate docs correctly this is only supposed to happen with fetching. I believe that this issue is not isolated to collections. We are seeing the same thing with associations and abstract an eager fetch resolves this by avoiding the proxy but its a bit of a hack. We are getting this Exception also in javassist
Ouch. Thats a nasty issue. Thanks for your time in investigating it. Well have a fix for the next milestone. Fixed on master. closing with the release of
updated forrest doc. an image showing the generated html. Looks failure is from Hudson attempting to apply the. gif instead of the patchI just committed this. Thanks NicholasNote that many changes to docs are from which didnt run forrest before committing its changes.
With this in hand and after reviewing the code more deeply Ive figured out the proper configuration to get this pretty close with the existing code. It turns out that the existing dependencies of the features project are parsed and included by reference if they exist but the transitive dependency graph is not pruned in the manner I described here. Im going to code up this pruning to see what I can come up with. Deep in the bowels of the beast were more clues. Pruning is accomplished by setting a dependency to provided scope. I guess this is all is really necessaryIf theres some desire for documentation please create an issue and assign it to me. please grab this one somehow Im not able to assign this to you
Fixed Updated in stratos controller dashboard module and service stub Attached patch file with commit edfaecdacefaacadfeecea
Is backwards compatible or we need to make code changes. Everything seems to be working fine with. Im going to commit this small change tonight if there is no objection. looks fine. Committed to trunk. Thanks Tom for checking.
This is only for the branch. . like this was committed to the branch. Can it be closed?
One of the failing test test pass atm
this is something we will look into after and Eclipse is both too rigid in the current form to be done easily. It is resolved as an duplicate bug there are no reason to have two issues open for the same thing. Not sure how appfuse can help the tooling in being more flexible about project archetypes are worseequal to what we have today since it is all code generation instead of adjusting existing projects artifact style random maven or any other codegen will not result for a guaranteed working does not mean maven archetypes is bad just mean that as a foundation for a wizard doing a specific thing I do not belive it is a good enough foundation. I think this bug is more general than related and should be open. Now this looks like we dosent want to fix this. Many people are using maven for many years and now are excited about seam but using this technologies together is real nightmare. There is no help or very small help from for people than are using maven. Spring was very nice for this intergration with maven was easy and smooth. I understand than jboss people loves ant but many developers think maven integration is big problem for seam adoption. All developers of seammaveneclipse are required to build all configurationsenvironments from zero. mariusz seam uses maven im not sure why you are discussing maven support for seam herethe tools work with any project structure now what is it you are asking for. Did you go and enable seam support for the project?
see destination refs are resolved to a in the stage and then handled like.
Per phone triage with Burr and Max on. in progress core done still work left.
the source code for better mistakesorry!
Attaching patch has been applied thanks Richard one more patch point
Unlikely to be leaking of classes more likely to be order of deploymentPlease attach a log file showing the complete sequence. attached it shows successful start of repo during server startup and then unsuccessful. Our version of lucene only seems to exist to support a JCR message store so changing may be straight forward. cant find any source reference to lucene so we may even be able to remove it. contains Lucene for SOA. If you want to use hsearch from seam it wont work with lucene . The seam version of comes in via hibernate to remove the esb version as part of the SOA will create a new JIRA to handle Lucene Seam revision in ER
I am not all that satisfied with how closures are serialized since that usually requires serialization of the class the closure is defined in as well. But besides that nothing really speaks against that idea I you dont serialize the Closures then what is it you want to serialize? You would be fine with an empty Expando after deserialization?
Done. Updated Branch to and that allowed the of the solution as mentioned aboveThe fix should also indicate a potential release but this has be removed from JIRA.
Per Burr Sutter on this is a blocker for to Beta since that is where these should be verifiedresolved. we got all the basic tooling parts in concerns are forge plugins readyness . reverse engineering and identifying our projects is not yet verifedtestedpossible.
Heres a patch with the proposed refactor. Good call simplify the signaturesUpdating the JIRA description now patch to follow. Patch with changed signature. thanks.
This is the right fix. Its not something primarily related to but rather to. Committed revision Thank you Doug for reviewing.
In the example given above you dont really need to specify config and schema parameters. By default Solr looks in instanceDirconf for them. However if you name your xml files different from the defaults then this bug will cause you problems. Fixed as part of close resolved issues