[JBoss JIRA] (FURNACE-120) Forge seems to be proxying java.nio.Path instances
by George Gastaldi (JIRA)
[ https://issues.jboss.org/browse/FURNACE-120?page=com.atlassian.jira.plugi... ]
George Gastaldi moved FORGE-2674 to FURNACE-120:
------------------------------------------------
Project: Forge: Furnace (was: Forge)
Key: FURNACE-120 (was: FORGE-2674)
> Forge seems to be proxying java.nio.Path instances
> --------------------------------------------------
>
> Key: FURNACE-120
> URL: https://issues.jboss.org/browse/FURNACE-120
> Project: Forge: Furnace
> Issue Type: Bug
> Reporter: Jess Sightler
>
> Class name: org.javassist.tmp.java.nio.file.Path_$$_javassist_e2aee44f-1c2f-4dea-9465-cd181bb52a83
> ClassLoader: ModuleClassLoader for Module "org.jboss.windup.exec.windup-exec:3.0.0-SNAPSHOT_45979342-4d41-4637-84bf-5f11f9de6cb8" from AddonModuleLoader
> Calling Files.isDirectory with this proxy object results in the following exception:
> Windup Execution failed due to: org.javassist.tmp.java.io.Closeable_$$_javassist_43fd2113-1fd1-4d0d-a9f6-7bf2667e61ab cannot be cast to java.nio.file.FileSystem
> java.lang.ClassCastException: org.javassist.tmp.java.io.Closeable_$$_javassist_43fd2113-1fd1-4d0d-a9f6-7bf2667e61ab cannot be cast to java.nio.file.FileSystem
> at org.javassist.tmp.java.nio.file.Path_$$_javassist_e2aee44f-1c2f-4dea-9465-cd181bb52a83.getFileSystem(Path_$$_javassist_e2aee44f-1c2f-4dea-9465-cd181bb52a83.java)
> at java.nio.file.Files.provider(Files.java:97)
> at java.nio.file.Files.readAttributes(Files.java:1737)
> at java.nio.file.Files.isDirectory(Files.java:2192)
> at org.jboss.windup.ext.groovy.GroovyWindupRuleProviderLoader.getScripts(GroovyWindupRuleProviderLoader.java:184)
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2674) Forge seems to be proxying java.nio.Path instances
by Jess Sightler (JIRA)
[ https://issues.jboss.org/browse/FORGE-2674?page=com.atlassian.jira.plugin... ]
Jess Sightler moved WINDUP-1131 to FORGE-2674:
----------------------------------------------
Project: Forge (was: WindUp)
Key: FORGE-2674 (was: WINDUP-1131)
> Forge seems to be proxying java.nio.Path instances
> --------------------------------------------------
>
> Key: FORGE-2674
> URL: https://issues.jboss.org/browse/FORGE-2674
> Project: Forge
> Issue Type: Bug
> Reporter: Jess Sightler
>
> Class name: org.javassist.tmp.java.nio.file.Path_$$_javassist_e2aee44f-1c2f-4dea-9465-cd181bb52a83
> ClassLoader: ModuleClassLoader for Module "org.jboss.windup.exec.windup-exec:3.0.0-SNAPSHOT_45979342-4d41-4637-84bf-5f11f9de6cb8" from AddonModuleLoader
> Calling Files.isDirectory with this proxy object results in the following exception:
> Windup Execution failed due to: org.javassist.tmp.java.io.Closeable_$$_javassist_43fd2113-1fd1-4d0d-a9f6-7bf2667e61ab cannot be cast to java.nio.file.FileSystem
> java.lang.ClassCastException: org.javassist.tmp.java.io.Closeable_$$_javassist_43fd2113-1fd1-4d0d-a9f6-7bf2667e61ab cannot be cast to java.nio.file.FileSystem
> at org.javassist.tmp.java.nio.file.Path_$$_javassist_e2aee44f-1c2f-4dea-9465-cd181bb52a83.getFileSystem(Path_$$_javassist_e2aee44f-1c2f-4dea-9465-cd181bb52a83.java)
> at java.nio.file.Files.provider(Files.java:97)
> at java.nio.file.Files.readAttributes(Files.java:1737)
> at java.nio.file.Files.isDirectory(Files.java:2192)
> at org.jboss.windup.ext.groovy.GroovyWindupRuleProviderLoader.getScripts(GroovyWindupRuleProviderLoader.java:184)
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2671) UIContext - Allow to open a file on a specific element / line number
by George Gastaldi (JIRA)
[ https://issues.jboss.org/browse/FORGE-2671?page=com.atlassian.jira.plugin... ]
George Gastaldi commented on FORGE-2671:
----------------------------------------
[~davsclaus],
MethodSource implements {{org.jboss.forge.roaster.model.source.LocationCapable}} that has the information needed. I think it would be something like (haven't fully tested):
{code:java}
MethodSource method = ...;
uiContext.setSelection(Selections.from(
(r) -> UIRegionBuilder.create(r)
.startPosition(method.getStartPosition())
.endPosition(method.getEndPosition())
.startLine(method.getLineNumber())
.endLine(method.getLineNumber()),
Arrays.asList(javaResource)));
{code}
> UIContext - Allow to open a file on a specific element / line number
> --------------------------------------------------------------------
>
> Key: FORGE-2671
> URL: https://issues.jboss.org/browse/FORGE-2671
> Project: Forge
> Issue Type: Feature Request
> Components: UI - API
> Affects Versions: 3.2.3.Final
> Reporter: Claus Ibsen
> Assignee: George Gastaldi
> Fix For: 3.3.0.Final
>
>
> [12:19:40] <davsclaus> gastaldi thanks for the open file - it works nicely
> [12:19:41] <davsclaus> https://github.com/fabric8io/fabric8-forge/blob/49efd47241fbe1805d5fc4218...
> [12:19:56] <davsclaus> one thing I would like to be able to do is to position the cursor on a specific element / line
> [12:20:02] <davsclaus> currently the cursor is open on line 1
> [12:20:12] <davsclaus> eg typically in a license header in the file etc
> [12:20:32] <@gastaldi> Ah
> [12:20:32] <davsclaus> i would like to set the cursor to a newly created class / method / field etc
> [12:20:45] <davsclaus> or in a .properties file / plain file - to say line 35 etc
> [12:21:00] <@gastaldi> Hm I think that's doable. Can you @issue it?
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2670) Command-completion is slow
by Daniel Platz (JIRA)
[ https://issues.jboss.org/browse/FORGE-2670?page=com.atlassian.jira.plugin... ]
Daniel Platz edited comment on FORGE-2670 at 8/14/16 9:02 AM:
--------------------------------------------------------------
Output of attaching a conditional breakpoint to DefaultFileOperations.fileExists for a single command-completion:
{code:java}
java.nio.file.Files.write(java.nio.file.Paths.get("C:/_dump/out.txt"), (f + "\n").getBytes(), java.nio.file.StandardOpenOption.APPEND);
return false;
{code}
was (Author: platzda):
Output of attaching a conditional breakpoint to DefaultFileOperations.fileExists for a single command-completion:
> Command-completion is slow
> --------------------------
>
> Key: FORGE-2670
> URL: https://issues.jboss.org/browse/FORGE-2670
> Project: Forge
> Issue Type: Bug
> Affects Versions: 3.2.3.Final
> Environment: Windows 7 64-Bit; JDK 8 64-Bit
> Intel Cor i7-4800MQ @ 2.7GHz
> 16 GB RAM
> SSD
> Reporter: Daniel Platz
> Assignee: George Gastaldi
> Attachments: jforge.png, out.txt, threaddumps.txt
>
>
> When typing e.g. "comm" and hitting Tab, it takes approximately 2 seconds for the command-completion to kick in and make suggestions/complete the command.
> I have written about it already here: https://developer.jboss.org/message/960679#960679
> Not sure what best to provide to make analysis easier. I am also Ok with debugging it myself if there maybe is a pointer in the right direction where to look into.
> Regarding the questions in the forum post:
> * I did not run any previous commands; it was this way from the beginning. I also deleted the ~/.forge folder to start afresh. Same issue
> * I used the forge console from within eclipse and from the cmd. Both the same.
> * I am using the 3.2.3 Final version
> I have attached a screenshot of the jvisualvm sampler. Steps i did:
> 1. Type "com"
> 2. Start sampler
> 3. Type "TAB"
> 4. Wait for command-completion
> 5. Stop sampler
> It took approximately ~2 seconds for the command-compeltion to respond. it seems the extensive checks like DefaultFileOperations.fileExists are a big part of it as it accumlates to ~1 seconds which sound much.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2670) Command-completion is slow
by Daniel Platz (JIRA)
[ https://issues.jboss.org/browse/FORGE-2670?page=com.atlassian.jira.plugin... ]
Daniel Platz updated FORGE-2670:
--------------------------------
Attachment: out.txt
Output of attaching a conditional breakpoint to DefaultFileOperations.fileExists for a single command-completion:
> Command-completion is slow
> --------------------------
>
> Key: FORGE-2670
> URL: https://issues.jboss.org/browse/FORGE-2670
> Project: Forge
> Issue Type: Bug
> Affects Versions: 3.2.3.Final
> Environment: Windows 7 64-Bit; JDK 8 64-Bit
> Intel Cor i7-4800MQ @ 2.7GHz
> 16 GB RAM
> SSD
> Reporter: Daniel Platz
> Assignee: George Gastaldi
> Attachments: jforge.png, out.txt, threaddumps.txt
>
>
> When typing e.g. "comm" and hitting Tab, it takes approximately 2 seconds for the command-completion to kick in and make suggestions/complete the command.
> I have written about it already here: https://developer.jboss.org/message/960679#960679
> Not sure what best to provide to make analysis easier. I am also Ok with debugging it myself if there maybe is a pointer in the right direction where to look into.
> Regarding the questions in the forum post:
> * I did not run any previous commands; it was this way from the beginning. I also deleted the ~/.forge folder to start afresh. Same issue
> * I used the forge console from within eclipse and from the cmd. Both the same.
> * I am using the 3.2.3 Final version
> I have attached a screenshot of the jvisualvm sampler. Steps i did:
> 1. Type "com"
> 2. Start sampler
> 3. Type "TAB"
> 4. Wait for command-completion
> 5. Stop sampler
> It took approximately ~2 seconds for the command-compeltion to respond. it seems the extensive checks like DefaultFileOperations.fileExists are a big part of it as it accumlates to ~1 seconds which sound much.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2670) Command-completion is slow
by Daniel Platz (JIRA)
[ https://issues.jboss.org/browse/FORGE-2670?page=com.atlassian.jira.plugin... ]
Daniel Platz reopened FORGE-2670:
---------------------------------
Assignee: George Gastaldi
Hi George,
reopening so i can attach another file. Maybe you have some insight into this based on the new information provided.
Cheers,
Daniel
> Command-completion is slow
> --------------------------
>
> Key: FORGE-2670
> URL: https://issues.jboss.org/browse/FORGE-2670
> Project: Forge
> Issue Type: Bug
> Affects Versions: 3.2.3.Final
> Environment: Windows 7 64-Bit; JDK 8 64-Bit
> Intel Cor i7-4800MQ @ 2.7GHz
> 16 GB RAM
> SSD
> Reporter: Daniel Platz
> Assignee: George Gastaldi
> Attachments: jforge.png, threaddumps.txt
>
>
> When typing e.g. "comm" and hitting Tab, it takes approximately 2 seconds for the command-completion to kick in and make suggestions/complete the command.
> I have written about it already here: https://developer.jboss.org/message/960679#960679
> Not sure what best to provide to make analysis easier. I am also Ok with debugging it myself if there maybe is a pointer in the right direction where to look into.
> Regarding the questions in the forum post:
> * I did not run any previous commands; it was this way from the beginning. I also deleted the ~/.forge folder to start afresh. Same issue
> * I used the forge console from within eclipse and from the cmd. Both the same.
> * I am using the 3.2.3 Final version
> I have attached a screenshot of the jvisualvm sampler. Steps i did:
> 1. Type "com"
> 2. Start sampler
> 3. Type "TAB"
> 4. Wait for command-completion
> 5. Stop sampler
> It took approximately ~2 seconds for the command-compeltion to respond. it seems the extensive checks like DefaultFileOperations.fileExists are a big part of it as it accumlates to ~1 seconds which sound much.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2670) Command-completion is slow
by Daniel Platz (JIRA)
[ https://issues.jboss.org/browse/FORGE-2670?page=com.atlassian.jira.plugin... ]
Daniel Platz commented on FORGE-2670:
-------------------------------------
Hi George,
I also would like to understand, so i am not giving up yet :-)
Maybe it is a combination of my system beeing slow due to some service but also a lot of file-io done by forge.
Can you maybe on how much file-access you would expect during such a command-completion event.
I am especially looking into DefaultFileOperations.fileExists as i was seen this a huge part of the execution-time in jvisualmvm.
I have set an Exception-breakpoint to count the number of executions and for what files like this:
{code:java}
java.nio.file.Files.write(java.nio.file.Paths.get("C:/_dump/out.txt"), (f + "\n").getBytes(), java.nio.file.StandardOpenOption.APPEND);
return false;
{code}
What i get is attached file with a huge amount calls (~2500 file-system calls via DefaultFileOperations.fileExists) during the command-completion. Is this expected/the same on your system?
In case you do not oberserve this huge amounts of file-access, let me know if i can provide you anything that might help further analysis. I really want to get the forge console running! :-)
> Command-completion is slow
> --------------------------
>
> Key: FORGE-2670
> URL: https://issues.jboss.org/browse/FORGE-2670
> Project: Forge
> Issue Type: Bug
> Affects Versions: 3.2.3.Final
> Environment: Windows 7 64-Bit; JDK 8 64-Bit
> Intel Cor i7-4800MQ @ 2.7GHz
> 16 GB RAM
> SSD
> Reporter: Daniel Platz
> Attachments: jforge.png, threaddumps.txt
>
>
> When typing e.g. "comm" and hitting Tab, it takes approximately 2 seconds for the command-completion to kick in and make suggestions/complete the command.
> I have written about it already here: https://developer.jboss.org/message/960679#960679
> Not sure what best to provide to make analysis easier. I am also Ok with debugging it myself if there maybe is a pointer in the right direction where to look into.
> Regarding the questions in the forum post:
> * I did not run any previous commands; it was this way from the beginning. I also deleted the ~/.forge folder to start afresh. Same issue
> * I used the forge console from within eclipse and from the cmd. Both the same.
> * I am using the 3.2.3 Final version
> I have attached a screenshot of the jvisualvm sampler. Steps i did:
> 1. Type "com"
> 2. Start sampler
> 3. Type "TAB"
> 4. Wait for command-completion
> 5. Stop sampler
> It took approximately ~2 seconds for the command-compeltion to respond. it seems the extensive checks like DefaultFileOperations.fileExists are a big part of it as it accumlates to ~1 seconds which sound much.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2670) Command-completion is slow
by George Gastaldi (JIRA)
[ https://issues.jboss.org/browse/FORGE-2670?page=com.atlassian.jira.plugin... ]
George Gastaldi closed FORGE-2670.
----------------------------------
Resolution: Cannot Reproduce Bug
Closing as cannot reproduce, but I am curious to know what is causing that. I appreciate if you add a comment here when you find out the cause. Thank you!
> Command-completion is slow
> --------------------------
>
> Key: FORGE-2670
> URL: https://issues.jboss.org/browse/FORGE-2670
> Project: Forge
> Issue Type: Bug
> Affects Versions: 3.2.3.Final
> Environment: Windows 7 64-Bit; JDK 8 64-Bit
> Intel Cor i7-4800MQ @ 2.7GHz
> 16 GB RAM
> SSD
> Reporter: Daniel Platz
> Attachments: jforge.png, threaddumps.txt
>
>
> When typing e.g. "comm" and hitting Tab, it takes approximately 2 seconds for the command-completion to kick in and make suggestions/complete the command.
> I have written about it already here: https://developer.jboss.org/message/960679#960679
> Not sure what best to provide to make analysis easier. I am also Ok with debugging it myself if there maybe is a pointer in the right direction where to look into.
> Regarding the questions in the forum post:
> * I did not run any previous commands; it was this way from the beginning. I also deleted the ~/.forge folder to start afresh. Same issue
> * I used the forge console from within eclipse and from the cmd. Both the same.
> * I am using the 3.2.3 Final version
> I have attached a screenshot of the jvisualvm sampler. Steps i did:
> 1. Type "com"
> 2. Start sampler
> 3. Type "TAB"
> 4. Wait for command-completion
> 5. Stop sampler
> It took approximately ~2 seconds for the command-compeltion to respond. it seems the extensive checks like DefaultFileOperations.fileExists are a big part of it as it accumlates to ~1 seconds which sound much.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months
[JBoss JIRA] (FORGE-2670) Command-completion is slow
by Daniel Platz (JIRA)
[ https://issues.jboss.org/browse/FORGE-2670?page=com.atlassian.jira.plugin... ]
Daniel Platz commented on FORGE-2670:
-------------------------------------
Hi George,
I disabled the virus scanner (it is TrendMirco OfficeScan) and it is still slow.
Then i tried windows safe mode: Good speed now. So, there must be something on my system that is makingit slow.
I think this is not an issue with Forge, so if you agree, i will close the issue.
Thanks,
Daniel
> Command-completion is slow
> --------------------------
>
> Key: FORGE-2670
> URL: https://issues.jboss.org/browse/FORGE-2670
> Project: Forge
> Issue Type: Bug
> Affects Versions: 3.2.3.Final
> Environment: Windows 7 64-Bit; JDK 8 64-Bit
> Intel Cor i7-4800MQ @ 2.7GHz
> 16 GB RAM
> SSD
> Reporter: Daniel Platz
> Attachments: jforge.png, threaddumps.txt
>
>
> When typing e.g. "comm" and hitting Tab, it takes approximately 2 seconds for the command-completion to kick in and make suggestions/complete the command.
> I have written about it already here: https://developer.jboss.org/message/960679#960679
> Not sure what best to provide to make analysis easier. I am also Ok with debugging it myself if there maybe is a pointer in the right direction where to look into.
> Regarding the questions in the forum post:
> * I did not run any previous commands; it was this way from the beginning. I also deleted the ~/.forge folder to start afresh. Same issue
> * I used the forge console from within eclipse and from the cmd. Both the same.
> * I am using the 3.2.3 Final version
> I have attached a screenshot of the jvisualvm sampler. Steps i did:
> 1. Type "com"
> 2. Start sampler
> 3. Type "TAB"
> 4. Wait for command-completion
> 5. Stop sampler
> It took approximately ~2 seconds for the command-compeltion to respond. it seems the extensive checks like DefaultFileOperations.fileExists are a big part of it as it accumlates to ~1 seconds which sound much.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
7 years, 9 months