On 6/1/22 3:40 am, Paul Gilmartin wrote:
The SDSF development team have been using git for z/OS for the last 5 years or
so.
The source code is kept in zFS and our in-house build tool extracts to staging
PDSEs for the batch submission using jobgroups.
Why state to PDSE rather than build directly from zFS?
As Rob mentioned in a later post we have since written tooling to invoke
the PL/X and DTL compilers from zFS. Quite a few projects choose to stay
using PDS data sets and sync to the file system to use Git.
Some of us use SlickEdit rather than ISPF and we can instigate builds from
there using custom options on SlickEdit drop-down menu.
The builds use SSH from SlickEdit to run simple scripts on z/OS to construct
and submit JCL.
We use SDSF Rexx to capture build success and failure and can feedback that
info to the developer via email or straight into the SlickEdit console.
So my pessimism about a SlickEdit<->SDSF API interface was unwarranted. Does
the
captured information include job logs and SYSPRINTs?
I'm a big fan of the REXX SDSF API. Zowe provides a CLI that wraps the
z/OSMF REST API but it's not as rich as SDSF. It's simple to write REXX
shell scripts that create JSON which can be called from micro web
servers such as Python Flask to create REST APIs.
SlickEdit uses FTP as the transport to the zFS source code.
No NFS to share the source code?
I don't think we have NFS running. On Windows I use SSHFS
https://github.com/billziss-gh/sshfs-win. I spend most of my time in
WSL2 and use https://github.com/libfuse/sshfs but my IDE has file
syncing using FTP/SFTP so I don't really care about NFS.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN