[jdk8u-dev] RFR: 8374899: [8u] Fully handle clang as the toolchain in flags.m4 [v4]

[Andrew John Hughes]

On Tue, 27 Jan 2026 16:15:55 GMT, Severin Gehwolf <sgehwolf at openjdk.org> wrote:

> > > I'm not super-keen on the debug logs in GHAs. It makes the real build failure harder to spot (in the general case). YMMV.
> > 
> > 
> > Can you elaborate? In the past, I've found it impossible to find issues without being able to see what compiler flags are being passed. The reason I merged the 8u482 update into here was to get more verbose output for the Windows issue.
> 
> To me it's a two-step process. First figure out where the compilation error is (which seems good with the default log level), then dive into it by increasing the log level. When it comes to warning analysis and actual errors the default log level is nicer. The plus is that we don't hog disk space as well.

Ok, I can see your point on the size issue. I don't know how much GitHub archives these things. While I've been building this way for the best part of twenty years of OpenJDK development locally, I don't tend to keep around logs of anything but the most recent build of a particular JDK & configuration. I am aware that they can be pretty big.

I'll revert that commit when I merge in the pending other changes (the Windows fix, the JNF workaround). However, I can see me needing to keep this patch around for debugging. It would be impossible to find e.g. https://github.com/openjdk/jdk8u-dev/pull/742 without it.  Do you want me to also revert the other cases that were added prior to this change as well? They have been there without any issue to my knowledge since August 2024.

-------------

PR Comment: https://git.openjdk.org/jdk8u-dev/pull/740#issuecomment-3807126298