|RES: [dsdp-mtj-dev] Importing MTJ-projects|
Actually the information that is presented on .mtj and device.xml files are different. The information on device.xml is a workspace information that describe all devices that are imported on mtj. On the other hand, the .mtj file keep symbol set data that is associated to the different runtime configurations that a mtj project can have. The symbol sets are used on the java code preprocessing feature.
The user is able to edit / add / delete new fields when he adds a new configuration to the project. That why just having a reference to the device / group on .mtj file would not be enough. We usually refer to the data on the .mtj file as project level symbolset.
We presented a talk during last eclipsecon that describe the muti-configuration support on mtj. All of those concepts were described on the talk. Slides are available here. If you have some questions about it, feel free to send to the list. Suggestions about how to improve the feature or make them more clear are also welcome (specially with a patch that implements it J).
De: dsdp-mtj-dev-bounces@xxxxxxxxxxx [mailto:dsdp-mtj-dev-bounces@xxxxxxxxxxx] Em nome de Olsén, Ingemar
Q: Why is so much device information saved in the .mtj file? Isn't most of that information workspace specific?
A: The information you see on the .mtj file is the runtime configuration information required for each device configuration you have o your project. These information is project specific since the project configuration must kept across workspaces.
Yes, but the workspace have to have the same device that is connected to the project in its devices.xml file that is located in “.metadata/.plugins/org.eclipse.mtj.core”. All information regarding the device is already in that file. To me it seems like redundant information. Since you have to import that device to any workspace where you want to import your mtj-project anyway…
Wouldn’t just device name and group be enough?
I have a question regarding the import mechanism of MTJ projects. When working in a multiuser environment people may want to import a colleagues project into their workspace. This is troublesome if the device that is associated with the project is not located in the exact same path as the one that created the project, since the .mtj file is containing device specific information. Has this problem been discussed?
I tried to import an mtj-project that had been created on another computer. I got some errors so I change the device in the Java ME properties to a device that is available on my computer. Some of the errors disappeared it still produces an error saying "No device definition is associated to this project". I cannot find a way around this, apart from creating a new project instead.
Can you please
open a bug on MTJ (see link bellow) and describe a step by step process of
reproducing the issue so we can fix it.
Why is so much device information saved in the .mtj file? Isn't most of that information workspace specific?
The information you see on the .mtj file is the runtime configuration information required for each device configuration you have o your project. These information is project specific since the project configuration must kept across workspaces.
A nice way to handle it would be to have the importer ask the user which device he wants to associate with the project when he imports it.
Eclipse team, SonyEricsson
dsdp-mtj-dev mailing list