We’re a company producing for 30++ years a LMS (Library Management System) written using any language you can think off (C, C++, Perl, SQL, Java, JS, ESQL/C, …). With ‘huge’ I mean 20++ SVN modules, some 60.000 source files. Our “CI” has two parts: one is a bunch of shell scripts using make (…) to produce at the end packages which must be deployed to a Linux host; and the 2nd part are Jenkins jobs producing EXEC files with Bitrock to be unpacked on Windows PC (the clients for the LMS, while the servers are running on Linux on top of PostgreSQL DBS). Parts of the CI scripts must be run as root. At the end, the result of the CI (we call ist nightly build) are around 20 packages for Linux, 8 EXEC for Windows which get automatically deployed to some Linux servers and PC for quality tests and later delivered to some 120 customers. The build always starts every night based on svn checkout to which developers are committing the changes. This as background.
We have reasons to think in GitLab (mostly for static security scans of the sources). I did a test “git svn clone…” which pulled out all branches with their history out of SVN and I will do a test bringing the modules as projects into GitLab. But my main concern is how to connect our nightly build processes to this afterwards, Rewriting all these script in a form of .gitlab-ci.yaml CI is for sure not an option.
Any ideas or comments about this? Or any similar C-projects?
Matthias