I have been using using my Windows 10 laptop as a client to connect to an xpra server running on a virtual desktop instance (VDI) of RHEL6 at work. I don't have root access on my VDI (company security policy) in order to install the xpra packages, but I was able to work around this by extracting ALL of the required rpms manually and updating my ENV accordingly to point to my "installation".
The company is very slow to upgrade to new versions of RHEL/CentOS because of EDA tool compatibility issues and a conservative "if it isn't broke, don't fix it" philosophy. Therefore, I anticipate being stuck on xpra version 1.x for a while. In the meantime, I imagine that there have been performance improvements and bug fixes in 2.x and beyond that are not being back-ported to 1.x. It is possible that we MIGHT move to CentOS 7.3 relatively soon, but even that has been YEARS in the making, and the latest version of xpra isn't supported on CentOS 7.3 anyway. As another workaround, I've been playing around with the idea of compiling xpra from the source. I do have a recent version of python installed (3.7.1), and have been able to install many of the dependencies by creating a virtual environment and installing packages with pip. However, the fact that RHEL3 does not support gtk3 is kind of a show stopper. I haven't yet attempted to build gtk3 from source. This has got me thinking about the fact that xpra is monolithic package with both client and server together, and I suspect that the gtk3 dependencies relate mostly (or maybe only) to the client. In my use case however, I'll NEVER be running the client directly on the VDI, but only from my remote Windows 10 laptop, on which I can install the latest xpra client version. Is there some way to build xpra with ONLY server support? I'm guessing not, but its something to think about. I'm sure there are other use cases out there like mine. _______________________________________________ shifter-users mailing list [email protected] https://lists.devloop.org.uk/mailman/listinfo/shifter-users
