This just makes things confusing - why do the RPM and DEB versions (both x86 and ARM) and Windows of v9.3.3 have build hash `75595d8f83ef`, but when you look at the solaris UFs, the build hash is `740e48416363` ?! What gives? This just makes our lives more difficult when trying to organize large-scale downloads for users in a heterogeneous environment...
Currently, Linux x64/arm, Windows_x64, MacOS enterprise and UF packages are built from the same repo and have the same commit hashes. Linux s390x/ppcle, Windows 32-bit, FreeBSD, Solaris and AIX UFs, are built from a different repo. I understand the pain though, we didn't do the recent split lightly and we'll re-evaluate it at some point.
Solaris and other universal forwarder platforms are built from a separate repo. Historically, those legacy platforms have caused delays in upgrading the main platforms to the latest toolchains (compilers, libraries, etc). Hence, the different commit hash.
Sure but when I have end users depending on download links, that makes it confusing and difficult for them. Doesn't seem to be that it is an unreasonable ask to use the same commit hash for each build once it's versioned that way - in other words, you don't tag your repo with multiple of the same version numbers, you just tag it once and then compile the installer from that version. No reason to use different commits, as that only introduces additional variance in the way the installer behave, unless these are all built from separate repos, in which case I would expect the windows and linux and MacOS packages to all have different UF hashes, but they don't, so solaris shouldn't either in my estimation.