|Re: [cdt-dev] missing org.eclipse.cdt.debug.mi.core in Neon?|
If Liviu can quickly fix the issues with the missing dependencies we can then look at the remaining API issues. From what I understand from Jonah's investigation there are probably only two of those; and Jonah was able to fix one of them for TCF, so maybe the same could be done here.
That would leave one issue which probably cannot be fixed without changing CDT. However, assuming we can find a decent solution involving both CDT and the ARM plugins, I'm willing to look into reintroducing the removed API. This would need approval as it is very late, so I can't make promises, but we can try.
We have until Monday to make this happen.
The most serious of the cases involves multi-launches and race conditions between them. Have you ever had the GNU ARM plug-ins behave unexpectedly in multi-core cases?
Kichwa Coders Ltd.
On 24 May 2016 at 23:29, Liviu Ionescu <ilg@xxxxxxxxxx> wrote:
> On 25 May 2016, at 01:23, Jonah Graham <jonah@xxxxxxxxxxxxxxxx> wrote:
> The point is rather moot because the problem with the ARM plugins is not the removal of org.eclipse.cdt.debug.mi.core. That removal only led to your first visible error. Had CDT left org.eclipse.cdt.debug.mi.core in CDT 9.0 and users installed the ARM plug-ins the problems the users would have faced would be weird class errors instead and would not have triggered until that specific code was run.
I see. in this case things are really bad. but if the 8.x classes are that broken, I wonder how the GNU ARM plug-ins worked till now...
Back to the top