I am just curious as to how people are using SAMBA with QNX from the most basic features to the more advanced and creative solutions. I want get some ideas as to how I can incorporate SAMBA into our system. All I know about SAMBA is that it is used to share files and printers among different OS platform.
We have our entirely projects in severals QNX development machines. We’re used to run Eclipse self-hosted, and from there we compile (linking against our custom libraries and using lot of headers as you imagine). Now, as we know, the Neutrino-hosted version of the IDE is no longer supported in 6.4. So, we should considerate the idea of cross-development. But here’s the question: do I have to mount samba to share the folders in qnx where libs and headers resides?, is that the correct way?. I don’t think having all source files in a windows machine (in our case). Do somebody work in this way? (I’m talking about large distributed system, not embedded of course)
Other thing: we made, from inside the IDE, some references to other projects (actually they’re not projects, they’re the ‘include’ and the ‘lib’ directories of the whole project). So I can’t compile from the command line, because compiler don’t find headers or libs). How can I compile with the same project references as the IDE have inside from the command line?
Actually everything resides on the Windows machine. The headers, source files, libraries, etc. Nothing needs to reside on any QNX machine to edit and compile with the Windows version of Momentics IDE.
You edit and compile on Windows and then just transfer the binaries to a QNX machine to run/debug. I’ve only ever worked with having all the files on my local windows machine but you could definitely reference back to a common network server if you wanted to do so.
You do NOT need Samba for the transfer to QNX. The Momentics IDE under Windows has built in capabilites (via qconn) to transfer files, debug processes, run a pterm shell etc.
The only way I know to do this (compile in the IDE and command line) is to create your own Makefiles (we do this because we compile both ways) and setup the IDE to create a ‘standard Makefile’ project instead of an IDE project.
You probably need something like svn or CVS ( or any other source control system) to centralize the location of the projets.
Wait till you try the Windows version of the IDE is SO MUCH faster and far less buggy. I never though anyone would use the self-hosted version of Eclipse.
Thank you very much for feedback!. I would have to clarify some questions
Yes I know this, I used Eclipse from Windows many times (but just for debug, memory analysis, and some little compilations and tests), my question was oriented about if I want to see from my Windows machine (IDE) a QNX dir and keep compiling over that fs QNX tree where libs and headers resides.
Yes, we have our centralized repository, but we used to have one QNX machine (per project) synchronized with that unique CVS repository (now finally we start migrating to SVN)
Sincerely I developed a whole project using the self-hosted IDE in QNX 6.2.1 (first version) some years ago without majors difficulties and it was not so buggy, it seems that the IDE problems began to appear later. And since then we stopped using it.
Ok, I found other way to compile from the command line. It’s exporting some environment variables like this: