Dear Jason, apologies for the late response, somehow your message got carried away in a flood of mail over the past few weeks. You could checkout your code first using a native version control software (subversion, git, or whatever you use), and then import it in TRex. You could also split it in smaller chunks and import these individually (although that may be non-trivial or even impossible). I do agree that 1.5GB is just way too large, and some restructuring is probably long overdue so that you do not have to manage all of that all at the same time. In order to use more than 2GB of memory, you would have to use a 64-bit version of Java. I can imagine it being slow and unstable, given the size of the codebase, I’m afraid it couldn’t be any different at this point, with other tools probably being pushed to their limits as well. We are working on some backend improvements which should reduce the resource usage and improve stability and speed, but even with them, the code base is just too large. I’d highly recommend considering some restructuring so it can be easily managed in smaller chunks. I don’t think anyone can quite grasp what is going on across 25 000 files. Ultimately it comes down to what you want to do with TRex. If you just want the editor and code completion, splitting the code in smaller chunks (sub-projects or simply importing individual directories) may go a long way. If you do want to do something more advanced, you may want to look at some of the commercial offerings, although I’m not sure how well they can manage such large code bases either. Best regards, Philip On 02 Oct 2013, at 14:12, Avramidis, Jason (NSN - GR/Athens) <jason.avramidis@nsn.com> wrote:
Thank you for your prompt response Philip
I didn’t have any luck with the plug-in version of TRex. Already played with the –Xmx variable, but Eclipse gets unbearably slow during project buildup.
I tried the standalone installer for TRex and it seems to be working somewhat better. Initially it would crash whilst creating (checkout) the Project (which is too large ~ 1.5GB and 25K files in total)! So, I edited –Xmx and set it to 1500. That allowed me to Checkout the code. Looking at the Task Manager, trex.exe tops at 1.620.000K– which means it’s already utilized the 1500 I allocated to it. I tried setting the –Xmx to a high value e.g. 2000 or more, but Java would through up an error and TRex wouldn’t start. My system has 4GB of RAM, so I should have been allowed to use more e.g. 2000, but the JVM won’t have it L Any workaround to this?
Even though TRex is actually running and sort of it does what I want – it’s very slow; it often crashes after some light-to-moderate usage. Are there any options from within the TRex advanced config that I can safely disable – in order to reduce memory usage?
Regards Jason
From: Trex-user [mailto:trex-user-bounces@informatik.uni-goettingen.de] On Behalf Of ext Philip Makedonski Sent: Tuesday, October 01, 2013 2:51 PM To: TRex Users Mailing-List Subject: Re: [Trex-user] FW: T-Rex (newbie) query!
Dear Jason,
from the looks of it, it seems like a memory issue (GC overhead limit exceeded). You mentioned working with admittedly rather large TTCN-3 code base which would also suggest potential memory issues. The general remedy is trying to increase the memory settings even further (-Xmx1024m or more). Your intuition is otherwise correct, this is indeed one of the intended functionalities of TRex. But before you perform any of those actions, you need to make sure that initial processing is finished (depending on the codebase size it may take a few minutes to parse and index all resources).
It could also be a bug of course, and I will have a look at this as far as time permits, but generally memory consumption with large test suites is a known limitation of the current implementation and careful memory selection mostly helps (~2 GB for 200 000+ LOC TTCN-3 code, for example).
Best regards,
Philip
On Oct 1, 2013, at 12:42 PM, "Avramidis, Jason (NSN - GR/Athens)" <jason.avramidis@nsn.com> wrote:
OK – seems that attachments are not permissible. The error was:
=== An internal error occurred during: searching for markers Decoratio calculation GC overhead limit exceeded ===
Eclipse does not return any results nor following through the code as I would expect it to do so…
Jason
_____________________________________________ From: Avramidis, Jason (NSN - GR/Athens) Sent: Tuesday, October 01, 2013 1:36 PM To: 'trex-user@informatik.uni-goettingen.de' Subject: T-Rex (newbie) query!
Dear all
I am trying to use the T-Rex build-in functionality “open declaration” and “find references”. Nothing seems to happen and after a while (hanging Eclipse) I get the following errors:
<Picture (Device Independent Bitmap) 1.jpg>
I am using the plugin of T-Rex on Eclipse 3.7.2 . As recommended, I have increased the mem usage to 512m, but still seems to have no effect. I am using Subclipse for my TTCN script repo (which is relative large) and the default Java perspective (if that has anything to do with it).
What I want to do here is to enable Eclipse to ‘resolve’ functions, declarations etc so that I can navigate through the code quicker – without having to manually look where a particular function is stored. I presume T-Rex should allow me to do this, right?
Any thoughts that might help me figure what’s wrong here?
Cheers Jason
_______________________________________________ Trex-user mailing list Trex-user@informatik.uni-goettingen.de https://user.informatik.uni-goettingen.de/mailman/listinfo/trex-user
_______________________________________________ Trex-user mailing list Trex-user@informatik.uni-goettingen.de https://user.informatik.uni-goettingen.de/mailman/listinfo/trex-user