I’m trying to set up a CVS server on our mainframe to be used as the code repository for Eclipse (Running on Windows).
After mucking about with the inetd.conf, we got it going, but when I try to connect from the PC I get the following message:
cvs [login aborted]: dbg: res = -1, pccsid = 819, fccsid = 1047,
Any help would be appreciated.
One final question: Is anyone else using CVS to manage the code repository for eclipse developers, who are running on the PC (ie ASCII) platform.
We will look into this and get back to you within 24 hours.
We are able to reproduce the problem and are investigating what will be required to resolve. This may take some time. We will post again once this is resolved.
Just checking progress…
We have a kind of similar problem but reverse, we are using ssh/cvs to checkout files from a AIX-server to USS, but this gives garbled text on the screen. (probably a codepage problem too)
z/OS Version: 02.02
cvs client z/OS: 1.11.23
cvs server AIX: 1.11.20
AIX version: 188.8.131.52
Hi Hans, this is an inherent limitation related to codepage. Our CVS port does not do any automatic conversion between ASCII and EBCDIC. Our current focus for source code control system support is for git. A release is imminent for z/OS git with both ASCII and EBCDIC support with auto-conversion and auto-tagging. Although we intend to eventually enhance our CVS client port for code conversion, given our current focus on git and some other high profile languages and tools, it is unlikely we will get back to CVS any time soon unless a customer is using it and also paying for support.
Hi Peter, we played around with jgit, and that works ok, beside some issue’s which codepages, although it would be nice to have the git-hooks working, to bypass some of the issues.
Since z/Os 2.x supports unicode it would be very nice to have a native git version that supports unicode/ebcdic conversions and binary blobs over ssh or http connections.
So we certainly are interested in your upcoming git release.
In the mean time to use our CVS client port user has two options:
To understand how to use option 1 and how the upcoming git port will enable ASCII and EBCDIC projects, check out UNIX System Services Planning, Chapter 11 (Converting files between code pages) and in particular the section entitled Using Enhanced ASCII.
Sorry，I have not idea of it