When keepwork and keeptemp (and -stricter in my case) FEATURE flags are used, successive ebuild executions stack any reported QA warnings, due to the way warnings are grep-ed out from the log. The "problem" code is in this file: /usr/lib/portage/python2.7/install-qa-check.d/90gcc-warnings ... f=$(LC_CTYPE=C LC_COLLATE=C "${grep_cmd}" -E "${joined_msgs}" "${PORTAGE_LOG_FILE}") ... eqawarn "QA Notice: Package triggers severe warnings which indicate that it" eqawarn " may exhibit random runtime failures." eqawarn "${f}" ... It seems to me that the QA warnings become part of that log file (example log filename: /var/log/portage/ebuild.logs/build/sys-boot/grub-9999-r1:20150212-035520.log ) and because of the way they are reported they will match the grep again on a successive ebuild call so they will be reported again, effectively doubling them. The temp/logging/other file in which they are put(apparently) was about 3 megs at one point when I decided to investigate because it took too many seconds to display the QA warnings on console. Reproducible: Always Steps to Reproduce: 1. ebuild /usr/portage/sys-boot/grub/grub-9999-r1.ebuild merge 2. ebuild /usr/portage/sys-boot/grub/grub-9999-r1.ebuild merge 3. Expected Results: exclude already reported QA warnings from the existing .log file, so they don't keep accruing with each ebuild run.
looks like this: first time reported: * QA Notice: Package triggers severe warnings which indicate that it * may exhibit random runtime failures. * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implicit declaration of function ‘printf’ [-Wimplicit-function-declaration] second time reported: * QA Notice: Package triggers severe warnings which indicate that it * may exhibit random runtime failures. * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implicit declaration of function ‘printf’ [-Wimplicit-function-declaration] * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implic it declaration of function ‘printf’ [-Wimplicit-function-declaration] third time reported: * QA Notice: Package triggers severe warnings which indicate that it * may exhibit random runtime failures. * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implicit declaration of function ‘printf’ [-Wimplicit-function-declaration] * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implic it declaration of function ‘printf’ [-Wimplicit-function-declaration] * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implic it declaration of function ‘printf’ [-Wimplicit-function-declaration] * * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: imp licit declaration of function ‘printf’ [-Wimplicit-function-declaration] * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:532:5: warning: format ‘% u’ expects a matching ‘unsigned int’ argument [-Wformat=] * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:532:5: warning: format ‘% u’ expects a matching ‘unsigned int’ argument [-Wformat=]
failed the copy/pasting above, the lines got merged. I added newlines manually to the third, looks like this: * QA Notice: Package triggers severe warnings which indicate that it * may exhibit random runtime failures. * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implicit declaration of function ‘printf’ [-Wimplicit-function-declaration] * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implic it declaration of function ‘printf’ [-Wimplicit-function-declaration] * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implic it declaration of function ‘printf’ [-Wimplicit-function-declaration] * * * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: imp licit declaration of function ‘printf’ [-Wimplicit-function-declaration] * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:532:5: warning: format ‘% u’ expects a matching ‘unsigned int’ argument [-Wformat=] * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:532:5: warning: format ‘% u’ expects a matching ‘unsigned int’ argument [-Wformat=]
Created attachment 396232 [details] included whole 12meg log (as 260KiB bz2) to see exactly how it looks like when QA warnings get duplicated like that
While the QA warnings were being displayed, but because there were a lot of them and I had to interrupt with C-c, the next ebuild run shows this: # time ebuild /usr/portage/sys-boot/grub/grub-9999-r1.ebuild merge >>> Existing ${T}/environment for 'grub-9999-r1' will be sourced. Run >>> 'clean' to start with a fresh environment. >>> WORKDIR is up-to-date, keeping... * checking ebuild checksums ;-) ... [ ok ] * checking auxfile checksums ;-) ... [ ok ] * checking miscfile checksums ;-) ... [ ok ] >>> It appears that 'pretend' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.pretended' to force pretend. >>> It appears that 'setup' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.setuped' to force setup. >>> It appears that 'unpack' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.unpacked' to force unpack. >>> It appears that 'prepare' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.prepared' to force prepare. >>> It appears that 'configure' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.configured' to force configure. >>> It appears that 'compile' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.compiled' to force compile. * Skipping make test/check due to ebuild restriction. >>> Test phase [disabled because of RESTRICT=test]: sys-boot/grub-9999-r1 >>> It appears that 'install' has already executed for 'grub-9999-r1'; skipping. >>> Remove '/var/tmp/portage/sys-boot/grub-9999-r1/.installed' to force install. * Removing /lib/systemd * Removing /lib32/systemd * Removing /lib64/systemd * Removing /usr/lib/systemd * Removing /usr/lib32/systemd * Removing /usr/lib64/systemd * Removing /etc/systemd * checking 611 files for package collisions * This package will overwrite one or more files that may belong to other * packages (see list below). You can use a command such as `portageq * owners / <filename>` to identify the installed package that owns a * file. If portageq reports that only one package owns a file then do * NOT file a bug report. A bug report is only useful if it identifies at * least two or more packages that are known to install the same file(s). * If a collision occurs and you can not explain where the file came from * then you should simply ignore the collision since there is not enough * information to determine if a real problem exists. Please do NOT file * a bug report at http://bugs.gentoo.org unless you report exactly which * two packages install the same file(s). See * http://wiki.gentoo.org/wiki/Knowledge_Base:Blockers for tips on how to * solve the problem. And once again, please do NOT file a bug report * unless you have completely understood the above message. * * package sys-boot/grub-9999-r1 NOT merged * * Detected file collision(s): * * /usr/share/info/dir * * Searching all installed packages for file collisions... * * Press Ctrl-C to Stop * * None of the installed packages claim the file(s). * * Package 'sys-boot/grub-9999-r1' NOT merged due to file collisions. If * necessary, refer to your elog messages for the whole content of the * above message. real 0m4.424s user 0m3.520s sys 0m0.660s This goes away if I remove /var/tmp/portage/sys-boot/grub-9999-r1/.installed and rerun. But I'm unsure if the error message makes sense, maybe it's a new issue.
I am currently using this workaround: f=$(LC_CTYPE=C LC_COLLATE=C "${grep_cmd}" --invert-match -E $'^( \x1b\[33;01m\*\x1b\[0m |\x1b\[31;01m \* \x1b\[39;49;00m)' "${PORTAGE_LOG_FILE}" | "${grep_cmd}" -E "${joined_msgs}") original line was: f=$(LC_CTYPE=C LC_COLLATE=C "${grep_cmd}" -E "${joined_msgs}" "${PORTAGE_LOG_FILE}") file: /usr/lib/portage/python2.7/install-qa-check.d/90gcc-warnings line:85 or so And it looks like this(with the 12meg log file in place): * QA Notice: Package triggers severe warnings which indicate that it * may exhibit random runtime failures. * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:505:4: warning: implicit declaration of function ‘printf’ [-Wimplicit-function-declaration] * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:532:5: warning: format ‘%u’ expects a matching ‘unsigned int’ argument [-Wformat=] * /var/tmp/portage/sys-boot/grub-9999-r1/work/grub-9999/grub-core/disk/diskfilter.c:532:5: warning: format ‘%u’ expects a matching ‘unsigned int’ argument [-Wformat=] haven't tested it when the '*' are red(the 31 above), if they are, ever.
The issue is a known result of the design. I doubt it's worth fixing. Keeptemp/keepwork are not normal use conditions and running src_install() multiple times is not supported.
The only reason I use keepwork and keeptemp is because, at this time, I don't know how else to keep the sourcecode in place while tinkering with the code(grub in this case) in the hopes of fixing some (grub)bug. I should probably read on how to use ebuilds. Thanks.
We could make it remove ${T}/logging/${EBUILD_PHASE} before it executes a particular phase. However, given the low-level nature of the ebuild(1) command, I think the existing behavior is fine. If you run the same phase multiple times without cleaning, then I think it's logical all of the elog messages to accumulate.
Just to be clear, the problem is not that the C reported warnings accrue(as I understand is meant in Comment #8 ), but that those warnings reported in the QA Notice block (also)accrue. Since the .log file is kept, it makes sense that those C warnings accrue, but it doesn't make sense to re-report in the next QA block those warnings from all the previous QA blocks. (and they also look nested) My workaround in Comment #5 completely disregards the warnings reported in all the previous QA Notice blocks from the .log, so the QA Notice block contains only those reported by C (or, to be accurate, any warnings that are not within the previous QA blocks from the log). This means the C-reported warnings still accrue, but this is way better than having those in QA block also accrue - in this latter case you would have to spend a lot of time waiting for them to scroll up the screen when they are displayed. (or just remove the .log file after a few compilations I guess) So, if I get 2 warnings in the QA Notice block, after 1st time compilation, and I don't clear the log for the next 13 compilations, then on the 14th compilation the QA Notice block will show me about 2^14 or 16384 warnings on screen, which will take a while to display. With the workaround, supposedly I'll only see 2*14 or 28. Sorry for the noise, I just wanted to make sure I was somewhat clearer, because it seemed to me this can be misinterpreted in the 2 ways I've just attempted to describe in the first paragraph. That's all from me:)
have to update it to make sure those env vars are passed to the second invocation of grep too f=$(LC_CTYPE=C LC_COLLATE=C "${grep_cmd}" --invert-match -E $'^( \x1b\[33;01m\*\x1b\[0m |\x1b\[31;01m \* \x1b\[39;49;00m)' "${PORTAGE_LOG_FILE}" | LC_CTYPE=C LC_COLLATE=C "${grep_cmd}" -E "${joined_msgs}" because this says they previously didn't: $ A=1 bash -c 'echo $A'|bash -c 'cat;echo $A.' 1 . so this works: $ A=1 bash -c 'echo $A'|A=2 bash -c 'cat;echo $A.' 1 2.
(In reply to Michał Górny from comment #6) err, running `ebuild ... install` multiple times is def supported. this shows up unrelated to keepwork/noclean.
