Home » Archived » Buckminster » SettingIgnoring Worspace's API Base Line
|
Re: SettingIgnoring Worspace's API Base Line [message #664158 is a reply to message #664121] |
Thu, 07 April 2011 15:59 |
|
Hi Adolfo,
On 2011-04-07 16:33, Adolfo Sánchez-Barbudo Herrera wrote:
> Hi Bucky Team,
>
> Recently, I've tried to change my hudson job[1] to use buckminster 3.7. I got the following errors, which were warnings
> when using buckminster 3.6:
>
> Error: file /opt/users/hudsonbuild/workspace/buckminster-mdt-ocl-core-3.
> 1-nightly/buildroot/buckminster.workspace/plugins/org.eclips e.ocl: An API baseline has not been set for the current
> workspace.
>
> My questions are the following:
> a) Is there any way via buckminster commands to change the workspace settings so that not setting API Base line for it
> is a warning rather than an error. If there were someone, could I have some guidances to do that ?
>
If it is possible to control this via a project specific setting, then I would recommend that you do that and check it
in. If not, then what you need is a template workspace. There is large amount of settings that affect your builds in
various ways and the Eclipse IDE provides convenient preference pages for them. The mechanics are simple:
1. Start your IDE on a non-existent workspace.
2. Set all settings that you want in your build.
3. Exit your IDE.
4. Create a zip of the workspace. This is now the template.
5. Unzip the template to the location that you later refer to with -data as the first step in your headless build.
Buckminster will now build your workspace with the settings from #2.
> b) I have read some documentation concerning setting the API base line for buckminster's workspace [2]. I'm wondering if
> hudson server has any defined API base lines, so that I may simply add a "setpref" command to use said API base line.
> Otherwise, would I have to define a baseline ? May a p2 URL be used as the the API Base Line ?
>
A baseline has to be a target definition so that's something that you'll need to prepare in advance. IIRC, Johannes
Utzig made some experimental support for this in the Hudson Buckminster plug-in but I'm not sure what the status is on that.
- thomas
|
|
| |
Re: SettingIgnoring Worspace's API Base Line [message #664342 is a reply to message #664334] |
Fri, 08 April 2011 12:17 |
|
On 2011-04-08 14:11, Adolfo Sánchez-Barbudo Herrera wrote:
> Well, I could try this. However, I've noticed that EMF job is using Buckminster 3.7 and it doesn't seem the build needs
> this extra effort. Would you mind point Mickal to this thread to obtain his feedback ?
>
I can see two reasons for that. Either they use the project specific settings that I proposed, or they didn't have the
warnings at all when they used 3.6.
- thomas
|
|
|
Re: SettingIgnoring Worspace's API Base Line [message #664381 is a reply to message #664342] |
Fri, 08 April 2011 15:07 |
Michal Ruzicka Messages: 23 Registered: July 2009 |
Junior Member |
|
|
Hi All,
comments bellow.
Dne 8.4.2011 14:17, Thomas Hallgren napsal(a):
> On 2011-04-08 14:11, Adolfo Sánchez-Barbudo Herrera wrote:
>
>> Well, I could try this. However, I've noticed that EMF job is using
>> Buckminster 3.7 and it doesn't seem the build needs
>> this extra effort. Would you mind point Mickal to this thread to
>> obtain his feedback ?
>>
> I can see two reasons for that. Either they use the project specific
> settings that I proposed, or they didn't have the warnings at all when
> they used 3.6.
It is as Thomas have guessed: we didn't get the warnings with the 3.6
based buckminster so we don't have the problem with the 3.7 based one.
On the workspace level settings and workspace template front: I'm not a
huge fan of the workspace template idea, as it is not very flexible when
you are in a need to make a particular workspace level setting a
changeable parameter of a headless build. So the preferred solution IMO
would be to add the support for changing the setting to Buckminster.
This can prove to be very easy to do (as was the case of SSH preferences
- see [1] and [2]). So if you could find out how to change that setting
programmatically and provide a patch for that, I'd be happy to review
and apply it.
Cheers,
Michal
[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=310204
[2]
http://dev.eclipse.org/viewcvs/viewvc.cgi?view=revision& root=TOOLS_BUCKMINSTER&revision=11445
|
|
|
Re: SettingIgnoring Worspace's API Base Line [message #664391 is a reply to message #664381] |
Fri, 08 April 2011 15:50 |
Adolfo Sanchez-Barbudo Herrera Messages: 260 Registered: July 2009 |
Senior Member |
|
|
Mid-air collision.
Michal, Thomas thanks, I'll look into the better way to deal with this.
Best Regards,
Adolfo.
----------
>I can see two reasons for that. Either they use the project specific settings that I proposed, or they didn't have the
warnings at all when they used 3.6.
1. As said, there is not project specific settings for this. I mean, setting the API baseline for the workspace, and having a warning or error for not setting it up, is an exclusive workspace setting.
2. I also believe that the didn't have those warning when they used 3.6. Now, the question is why my build has those warnings (Bucky3.6) / errors (Bucky 3.7) and the EMF one doesn't have any, providing we are both using the same Hudson/Buckminster stuff
Perhaps, having any other API Analysys setting could cause this ?. For intance, I can see that the project org.eclipse.ocl has
- an org.eclipse.pde.api.tools.apiAnalysisNature nature.
- an org.eclipse.pde.api.tools.apiAnalysisBuilder builder.
In the case this was the cause, we can't obviously remove this, since this is a requirement for our projects to ensure that developers are doing well with APIs.
Could you please ask Michal if he has done anything setting concerning API Analysy Tooling .
Thanks in advance,
Adolfo.
|
|
|
Re: SettingIgnoring Worspace's API Base Line [message #664402 is a reply to message #664391] |
Fri, 08 April 2011 16:19 |
|
Adolfo, Michal, please add enhancement requests to our bugzilla for any additional setting that you think is
appropriate. The API support is relatively new and your experience in this area is important.
Thanks,
Thomas Hallgren
On 2011-04-08 17:50, Adolfo Sánchez-Barbudo Herrera wrote:
> Mid-air collision.
>
> Michal, Thomas thanks, I'll look into the better way to deal with this.
>
> Best Regards,
> Adolfo.
> ----------
>
>> I can see two reasons for that. Either they use the project specific settings that I proposed, or they didn't have the
> warnings at all when they used 3.6.
>
> 1. As said, there is not project specific settings for this. I mean, setting the API baseline for the workspace, and
> having a warning or error for not setting it up, is an exclusive workspace setting.
>
> 2. I also believe that the didn't have those warning when they used 3.6. Now, the question is why my build has those
> warnings (Bucky3.6) / errors (Bucky 3.7) and the EMF one doesn't have any, providing we are both using the same
> Hudson/Buckminster stuff
>
> Perhaps, having any other API Analysys setting could cause this ?. For intance, I can see that the project
> org.eclipse.ocl has
> - an org.eclipse.pde.api.tools.apiAnalysisNature nature.
> - an org.eclipse.pde.api.tools.apiAnalysisBuilder builder.
>
> In the case this was the cause, we can't obviously remove this, since this is a requirement for our projects to ensure
> that developers are doing well with APIs.
>
> Could you please ask Michal if he has done anything setting concerning API Analysy Tooling .
>
> Thanks in advance,
> Adolfo.
|
|
|
Re: SettingIgnoring Worspace's API Base Line [message #664988 is a reply to message #664402] |
Tue, 12 April 2011 17:26 |
Adolfo Sanchez-Barbudo Herrera Messages: 260 Registered: July 2009 |
Senior Member |
|
|
Thommas, Michal,
When reading some bugzillas[1] I've found that using the importprefs command could be an alternative way to tune up this setting (among others) when running my job. Michal, could you expand the information in said bugzilla in which you say that this mechanism is not flexible ?.
On the other hand, I've tried to fix the problem by the means of setting up the API Base Line. I've finally got it [2] using the following commands
# We import a target definition which will be our API Base Line
importtargetdefinition " ${checkout.location}/org.eclipse.mdt/org.eclipse.ocl/releng/ org.eclipse.ocl.releng.buckminster/APIBaseLine.target "
addbaseline -A "MDT/OCL API Base Line"
However, I have a couple of inconveniences/issues/questions:
1. There are a lot of [Fatal Error] :1:1: Premature end of file.
However, this doesn't prevent the build success. Besides, I read in a different forum's thread[3] that this would be fixed in the next milestone.
2. There a lot of ERROR: Error logged from API Tools Core:
java.io.IOException: No such file or directory
at java.io.UnixFileSystem.createFileExclusively(Native Method)
at java.io.File.createNewFile(File.java:883)
at org.eclipse.pde.api.tools.internal.ApiBaselineManager.persis tStateCache(ApiBaselineManager.java:339)
However, this doesn't prevent the build sucess. Any idea what about this issue ?.
[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=310204
[2] https://hudson.eclipse.org/hudson/job/buckminster-mdt-ocl-co re-3.1-nightly/425/
[3] http://www.eclipse.org/forums/index.php?t=msg&goto=66349 9
Cheers,
Adolfo.
|
|
| |
Re: SettingIgnoring Worspace's API Base Line [message #665000 is a reply to message #664992] |
Tue, 12 April 2011 18:21 |
|
The "Buckminster 3.7 Integration" is actually a copy of "Buckminster 3.6
Integration" right now, so please don't use that. Use "Buckminster 3.7".
It contains the latest bits.
- thomas
On 2011-04-12 19:44, Adolfo Sánchez-Barbudo Herrera wrote:
> BTW,
>
> I've run our job[1] in Eclipse Hudon using "Buckminster 3.7 Integration".
>
> 1. Indeed, the [Fatal Error] :1:1: Premature end of file. disappear.
>
> 2. The IOException concerning API tooling, still remains.
>
> 3. The worst is that the build fails with the following error:
>
> INFO: [start
> org.eclipse.ocl.doc:osgi.bundle$3.1.0.qualifier#buckminster. prebind]
> ERROR: org.eclipse.core.runtime.CoreException:
> /opt/users/hudsonbuild/workspace/buckminster-mdt-ocl-core-3.
> 1-nightly/buildroot/buckminster.workspace/plugins/org.eclips
> e.ocl.doc/build.xml:58: The following error occurred while executing
> this line:
> /opt/users/hudsonbuild/workspace/buckminster-mdt-ocl-core-3.
> 1-nightly/buildroot/buckminster.workspace/plugins/org.eclips
> e.ocl.doc/build.xml:77: The following error occurred while executing
> this line:
> /opt/users/hudsonbuild/workspace/buckminster-mdt-ocl-core-3.
> 1-nightly/buildroot/buckminster.workspace/plugins/org.eclips
> e.ocl.doc/build.xml:81: Problem: failed to create task or type
> help.buildHelpIndex
> Cause: The name is undefined.
> Action: Check the spelling.
> Action: Check that any custom tasks/types have been declared.
> Action: Check that any <presetdef>/<macrodef> declarations have taken
> place.
>
> This looks like a regression issue, doesn't it ?
>
> [1]
> https://hudson.eclipse.org/hudson/job/buckminster-mdt-ocl-co re-3.1-nightly/427/
>
>
> P.S: I'm assuming that "Buckminster 3.7 Integration" is newer than
> "Buckminster 3.7". Correct me if I'm wrong :p
>
> Cheers,
> Adolfo.
|
|
|
Re: SettingIgnoring Worspace's API Base Line [message #665010 is a reply to message #664988] |
Tue, 12 April 2011 18:29 |
|
On 2011-04-12 19:26, Adolfo Sánchez-Barbudo Herrera wrote:
> However, I have a couple of inconveniences/issues/questions:
> 1. There are a lot of [Fatal Error] :1:1: Premature end of file.
>
This is a known bug in p2 3.7M6. It causes no harm other than the ugly
printout and is already fixed in an integration build.
https://bugs.eclipse.org/bugs/show_bug.cgi?id=341508
> However, this doesn't prevent the build success. Besides, I read in a
> different forum's thread[3] that this would be fixed in the next milestone.
>
> 2. There a lot of ERROR: Error logged from API Tools Core:
> java.io.IOException: No such file or directory
> at java.io.UnixFileSystem.createFileExclusively(Native Method)
> at java.io.File.createNewFile(File.java:883)
> at org.eclipse.pde.api.tools.internal.ApiBaselineManager.persis
> tStateCache(ApiBaselineManager.java:339)
>
That might be worth reporting to the PDE team. I haven't seen that before.
- thomas
|
|
| |
Goto Forum:
Current Time: Wed Feb 05 10:42:55 GMT 2025
Powered by FUDForum. Page generated in 0.04757 seconds
|