Builds are too verbose. I build with "make -s --no-print-directory" and yet the name of every file that is compiled still gets printed like this:
This is due to REPORT_BUILD, the commit log for which was this:
3.181 <firstname.lastname@example.org> 2000-02-24 19:55
Adding commands to print out the file getting built. This increases the usefulness of make -s
That's right, it took non-trivial effort to produce that spew. I guess ol' Jim Nance wanted something between "silence" and "great big compiler invocations". Sorry, Jim; silence is golden.
Patch coming in a bit. If this breaks something about the build I'm gonna cry.
Created attachment 575354 [details] [diff] [review]
This patch reduces the number of lines of output during a clobber debug build on my Linux64 box from 12687 to 7114.
Comment on attachment 575354 [details] [diff] [review]
Review of attachment 575354 [details] [diff] [review]:
Fine with me, I never use "make -s", and if you are using it it seems like you'd want less output.
Is there a way to get the names of compiled files without all the trash output that 'make' otherwise generates? I really like(d) the output from make -s, since I could tell at a glance where in the build process a given directory was. I often have several builds going on at once, or a build and something else, and this clean output makes triage easier and more efficient than if there is no output at all.
If we were using PyMake as the official build tool, we could hook into its API and build rich reporting not otherwise possible with GNU make. I have some work around this in bug 683438. It just needs to be hooked up to the build system.
I agree with bhackett, and am bummed about this. The spew wasn't overwhelming (the way normal spew is), and without it there's no way to tell how far along a build is. It's also nice to know what had to be rebuilt when changing headers and IDL and the like.
Ted, how would you feel about putting this back in and making the 'echo' configurable?
(In reply to Bobby Holley (:bholley) from comment #7)
See bug 705332 in this bug's dependencies.