I build all packages in my system with USEFLAGS=debug -ggdb so that when my ~amd64 systems breaks, I can give *complete* backtraces to pretty much any package. I was led to this bug by bug 292237 - compile of chromium was just hanging in the middle. I investigated, turning off debug opts, turning down CFLAGs, doing everything, same result, across multiple versions. Then I did a pstree and found that emerge was ultimately calling python, which was stuck doing an svn command. when I did svn, i got a ton of debug spam on the console itself. tried to look for svn config in /etc to see if I could turn debug output on console off, no such luck. kream@satori ~ $ svn --version 2> svndebugspam.txt svn, version 1.6.6 (r40053) compiled Nov 2 2009, 02:26:21 Copyright (C) 2000-2009 CollabNet. Subversion is open source software, see http://subversion.tigris.org/ This product includes software developed by CollabNet (http://www.Collab.Net/). The following repository access (RA) modules are available: * ra_neon : Module for accessing a repository via WebDAV protocol using Neon. - handles 'http' scheme - handles 'https' scheme * ra_svn : Module for accessing a repository using the svn network protocol. - with Cyrus SASL authentication - handles 'svn' scheme * ra_local : Module for accessing a repository on local disk. - handles 'file' scheme kream@satori ~ $ svndebugspam.txt is appended. similar output is seen on any svn command. Reproducible: Always
Created attachment 210799 [details] 2.5mb debug output on "svn --version"
this is actually a bug in dev-libs/apr - when dev-libs/apr is emerged with debug then svn spams info.
You enable debug info, you get debug info. Works as advertised.
debug is enabled globally across the system. 99.9% of console packages do not spew (toolchain-breaking) debug information to the prompt when run after being compiled with +debug in USE 0.1% of console packages spew (toolchain-breaking) debug information to the prompt when run after being compiled with +debug in USE. Is the fault with the first group or the second? I don't enable debug in USE for fun, I take the performance hit so that in my ~amd64 system, when things go wrong, I can be a responsible free software citizen and help the maximum extent, as I have been doing since bug #226 . It should not be "works as advertised" if enabling debug == broken compiles two packages away down the line. Now, this might be a bug in apr or in subversion, but this is a bug or at the very least an enhancement request in one of those packages. Kick it upstream if you have to (and i'll help with that as well) but please do treat it as a bug/feature request. --Gentoo user and contributor since 1.0_rc3-r12
According to the useflag description, debug messages are the main purpose of the debug use flag: debug - Enable extra debug codepaths, like asserts and extra output. If you want to get meaningful backtraces see http://www.gentoo.org/proj/en/qa/backtraces.xml Since your goal is to be always ready to provide meaningful bug reports (yay for dedicated users), it should be ok to turn off USE=debug for apr and just keep -ggdb and FEATURES=nostrip (or splitdebug). Would that be a suitable solution, or do you still think a feature request is needed?
