dev-qt/qt-creator-3.6.0 compiles fine but could not be installed: install -m 644 -p /var/tmp/portage/dev-qt/qt-creator-3.6.0/work/qt-creator-opensource-src-3.6.0/share/qtcreator/cplusplus/wrappedQtHeaders/QtCore/qobjectdefs.h /var/tmp/portage/dev-qt/qt-creator-3.6.0/image/usr/share/qtcreator/cplusplus/wrappedQtHeaders/QtCore/ make[2]: Leaving directory '/var/tmp/portage/dev-qt/qt-creator-3.6.0/work/qt-creator-opensource-src-3.6.0/share/qtcreator' make[1]: Leaving directory '/var/tmp/portage/dev-qt/qt-creator-3.6.0/work/qt-creator-opensource-src-3.6.0/share' /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment: line 839: estack_push: command not found /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment: line 828: estack_pop: command not found * ERROR: dev-qt/qt-creator-3.6.0::gentoo failed (install phase): * eshopts_pop: unbalanced push * * Call stack: * ebuild.sh, line 133: Called src_install * environment, line 2051: Called make_desktop_entry 'qtcreator' 'Qt Creator' 'QtProject-qtcreator' 'Qt;Development;IDE' * environment, line 1525: Called eshopts_pop * environment, line 828: Called die * The specific snippet of code: * estack_pop eshopts s || die "${FUNCNAME}: unbalanced push"; * * If you need support, post the output of `emerge --info '=dev-qt/qt-creator-3.6.0::gentoo'`, * the complete build log and the output of `emerge -pqv '=dev-qt/qt-creator-3.6.0::gentoo'`. * The complete build log is located at '/var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment'. * Working directory: '/var/tmp/portage/dev-qt/qt-creator-3.6.0/work/qt-creator-opensource-src-3.6.0' * S: '/var/tmp/portage/dev-qt/qt-creator-3.6.0/work/qt-creator-opensource-src-3.6.0' * QA Notice: command not found: * * /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment: line 839: estack_push: command not found * /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment: line 828: estack_pop: command not found Reproducible: Always
The real errors are /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment: line 839: estack_push: command not found /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment: line 828: estack_pop: command not found However this is very weird. make_desktop_entry() calls those functions, and they are defined in the same eclass... I'm confused.
Attach your environment file please.
Created attachment 420060 [details] qt-creator ebuild environment file
Created attachment 420110 [details] environment file This is my environment file from /var/tmp/portage/dev-qt/qt-creator-3.6.0/temp/environment
Strange, but when I change EAPI=6 to EAPI=5 in ebuild, it works.
Your env file contains the following function: pre_src_prepare () { if ! type epatch_user > /dev/null 2>&1; then local names="epatch_user epatch evar_push evar_push_set evar_pop estack_push estack_pop"; source <(awk "/^# @FUNCTION: / { p = 0 } /^# @FUNCTION: (${names// /|})\$/ { p = 1; } p { print }" /usr/portage/eclass/eutils.eclass); fi; epatch_user; for name in $names; do unset $name; done } Where does it come from?
(In reply to Davide Pesavento from comment #6) > Where does it come from? https://wiki.gentoo.org/wiki//etc/portage/patches
(In reply to Davide Pesavento from comment #6) > Your env file contains the following function: > > pre_src_prepare () > { > if ! type epatch_user > /dev/null 2>&1; then > local names="epatch_user epatch evar_push evar_push_set evar_pop > estack_push estack_pop"; > source <(awk "/^# @FUNCTION: / { p = 0 } /^# @FUNCTION: (${names// > /|})\$/ { p = 1; } p { print }" /usr/portage/eclass/eutils.eclass); > fi; > epatch_user; > for name in $names; > do > unset $name; > done > } > > Where does it come from? Actually, I have the same problem, and I also have the same code snippet in my /etc/portage/bashrc (also from https://wiki.gentoo.org/wiki//etc/portage/patches). If it is related to this issue (I can't check it right now), then the article should either be changed (so method works again) or marked as outdated.
That bashrc snippet is no longer needed (in fact it should be considered harmful) with EAPI=6 ebuilds. This bug is invalid. Cc'ing portage team as they might want to know about this.
*** Bug 571164 has been marked as a duplicate of this bug. ***
(In reply to Davide Pesavento from comment #9) > That bashrc snippet is no longer needed (in fact it should be considered > harmful) with EAPI=6 ebuilds. I will have to test as the scriptlet worked in the best phase for upstream patches. Could you please provide a link to docs for the new features so that one can see whether it is at least as useful as the scriptlet? > This bug is invalid. Non sequitur. You were talking about EAPI=6 but I seriously doubt all ebuilds are now using the new EAPI. I would like to fix and comment the scriptlet accordingly but I unfortunately don't have the time for testing right now.
I'm not getting the same errors. I'm using portage 2.2.26. Do I need to update anything else to see it?
(In reply to Pavel Šimerda (pavlix) from comment #11) > (In reply to Davide Pesavento from comment #9) > > That bashrc snippet is no longer needed (in fact it should be considered > > harmful) with EAPI=6 ebuilds. > > I will have to test as the scriptlet worked in the best phase for upstream > patches. Could you please provide a link to docs for the new features so > that one can see whether it is at least as useful as the scriptlet? > https://projects.gentoo.org/pms/6/pms.html#x1-13600011.3.3.7 In case people haven't noticed the problem yet, the scriptlet unsets random functions such as estack_push. > > This bug is invalid. > > Non sequitur. You were talking about EAPI=6 but I seriously doubt all > ebuilds are now using the new EAPI. > The bug happens only on EAPI=6 ebuilds. The number of ebuilds using EAPI 6 doesn't change the fact that it's invalid. The mere fact that the bug is triggered by a bashrc snippet makes it invalid.
*** Bug 571320 has been marked as a duplicate of this bug. ***
IIRC calling eapply_user is mandatory in EAPI 6, so the hack described in https://wiki.gentoo.org/wiki//etc/portage/patches is completely unnecessary for EAPI 6 ebuilds. Assuming the $EAPI variable is available when portage bashrc is sourced (I would assume it is), the snippet can be "fixed" by adding the following line at the *beginning* of pre_src_prepare(): [[ ${EAPI:-0} == [012345] ]] || return
(In reply to Davide Pesavento from comment #15) > > [[ ${EAPI:-0} == [012345] ]] || return Thanks, looks like this neutralizes the issue completely. Should this be present on the wiki page instead of a warning?
(In reply to selurvedu from comment #16) > (In reply to Davide Pesavento from comment #15) > > > > [[ ${EAPI:-0} == [012345] ]] || return > > Thanks, looks like this neutralizes the issue completely. Should this be > present on the wiki page instead of a warning? Sure. I've updated it for now, but please always feel free to edit the wiki.
bug #575074 also has same issue .. with .. the unset portion removed the script seems less toxic .. i.e. does not clobber existing eutil.eclass containing builds
