summaryrefslogtreecommitdiffstats
path: root/bin/isolated-functions.sh
Commit message (Collapse)AuthorAgeFilesLines
* Move filtering of misc bash variables from save_ebuild_env() toZac Medico2009-10-181-4/+1
| | | | | | | filter_readonly_variables(). Add additional variables found in the output of `env -i bash -c 'declare -p'`. svn path=/main/trunk/; revision=14676
* Remove reference to non-existent _source_ebuild function.Zac Medico2009-10-181-1/+0
| | | | svn path=/main/trunk/; revision=14663
* Preserve variables which have been set in global scope using 'declare'.Arfrever Frehtes Taifersar Arahesis2009-10-181-2/+2
| | | | svn path=/main/trunk/; revision=14662
* Simplify the conditionals from bug #273648. Thanks to Jonathan CallenZac Medico2009-10-171-0/+1
| | | | | | <abcd@g.o> for this patch. svn path=/main/trunk/; revision=14617
* Add back Arfrever's emerge --info change because it's good and I was justZac Medico2009-10-091-1/+1
| | | | | | confused. svn path=/main/trunk/; revision=14527
* Ask for the output of 'emerge --info' instead of 'emerge --info ↵Arfrever Frehtes Taifersar Arahesis2009-10-091-1/+1
| | | | | | =$CATEGORY/$PF'. Suggested by zmedico. svn path=/main/trunk/; revision=14526
* Ask for the output of 'emerge --info =$CATEGORY/$PF'.Arfrever Frehtes Taifersar Arahesis2009-10-091-2/+2
| | | | svn path=/main/trunk/; revision=14525
* Bug #270933 - Make the "If you need support ..." message ask for emergeZac Medico2009-10-091-2/+3
| | | | | | | --info and the build.log. Thanks to Sebastian Mingramm (few) for this patch. svn path=/main/trunk/; revision=14521
* - do not calculate source with of isolated-functions (for die function)Fabian Groffen2009-10-031-16/+16
| | | | | | | | | | | as we don't print it - move paths to the bottom and print ${S} as it's often necessary when a package fails. - move error message up directly after the message it says it failed (printing it below the code snippet usually just repeats the message) svn path=/main/trunk/; revision=14483
* only calculate spacing width for that part of the stack that we're going to ↵Fabian Groffen2009-10-031-1/+8
| | | | | | print svn path=/main/trunk/; revision=14482
* Remove ACCEPT_PROPERTIES filtering since it's filtered on the python side.Zac Medico2009-08-151-1/+1
| | | | svn path=/main/trunk/; revision=14065
* Add ACCEPT_PROPERTIES variable which has default value * and can be set inMounir Lamouri2009-08-111-1/+1
| | | | | | | make.conf or /etc/portage/package.properties. If ACCEPT_PROPERTIES doesn't match ebuild PROPERTIES, the ebuild is masked. svn path=/main/trunk/; revision=13997
* Filter PORTAGE_NONFATAL from the environment where appropriate.Zac Medico2009-08-061-1/+1
| | | | svn path=/main/trunk/; revision=13938
* Implement nonfatal().Arfrever Frehtes Taifersar Arahesis2009-08-061-0/+16
| | | | svn path=/main/trunk/; revision=13936
* Remove the code from bug #278895 since we're going to need a daemon processZac Medico2009-08-031-2/+1
| | | | | | | | | | | | in the ebuild's login session in order to pass signals from the controlling terminal to the detached session. A simple fifo-based approach in ebuild.sh does not seem to work since bash's read builtin occasionally loses the fifo data when it's 'Interrupted system call'. Maybe a python script will work better for the session leader/daemon. NOTE: The daemon will also be useful for implementing a fifo-based die helper (to replace the current signal- based approach). svn path=/main/trunk/; revision=13879
* Inside die(), only send the signal for subshell die support when we areZac Medico2009-07-301-1/+1
| | | | | | actually in a subshell. svn path=/main/trunk/; revision=13857
* Bug #278895 - Make ebuild.sh clean up orphaned processes that may have beenZac Medico2009-07-301-1/+2
| | | | | | | | | | | left behind by ebuild phases. This works by using setsid to create a new login session for the ebuild.sh process, and `kill -s SIGHUP 0` to send a SIGHUP signal to all processes in the session. The setsid is currently not done on the python side since that would cause the sandbox process in the session, and sandbox produces a warning message if it catches a SIGHUP signal. svn path=/main/trunk/; revision=13856
* Make elog functions use read -r when splitting lines, so that backslashesZac Medico2009-05-011-11/+11
| | | | | | intended for display will not be eaten. svn path=/main/trunk/; revision=13582
* Fix ACCEPT_LICENSE handling for binary packages, so that the build-time USEZac Medico2009-04-301-1/+1
| | | | | | settings are used calculate an appropriate ACCEPT_LICENSE value. svn path=/main/trunk/; revision=13560
* Revert r13390 (dosed and dohard functions) since a stack trace isn't reallyZac Medico2009-04-241-2/+1
| | | | | | | important this particular QA feature. Thanks to Tiziano Müller <dev-zero@g.o> for the suggestion. svn path=/main/trunk/; revision=13391
* Add dummy dosed and dohard functions for EAPI 3, so that a trace can beZac Medico2009-04-241-1/+2
| | | | | | displayed (not possible with the helper binaries). svn path=/main/trunk/; revision=13390
* Move the ebuild sourcing code into a new _source_ebuild() function.Zac Medico2009-04-081-0/+1
| | | | svn path=/main/trunk/; revision=13299
* Bug #264494 - In case the ebuild has enabled set -e during an ebuild phase,Zac Medico2009-04-011-0/+1
| | | | | | proactively call set +e inside qa_source(), qa_call(), and die(). svn path=/main/trunk/; revision=13273
* Update copyright headers.Zac Medico2009-03-181-1/+1
| | | | svn path=/main/trunk/; revision=13131
* Filter _sb_append_var in save_ebuild_env().Zac Medico2009-03-121-1/+1
| | | | svn path=/main/trunk/; revision=13091
* Make save_ebuild_env() filter NOCOLOR, and remove related code from ebuild.shZac Medico2009-02-181-1/+1
| | | | | | since that variable is not loaded from $T/environment anymore. svn path=/main/trunk/; revision=12625
* Filter the new hasg and hasgq functions inside save_ebuild_env().Zac Medico2009-02-071-1/+2
| | | | svn path=/main/trunk/; revision=12592
* Make die() to show a message indicating the repository that an ebuild cameZac Medico2009-01-091-1/+13
| | | | | | | | from, and do it during any phase for both ebuilds and binary packages. Now PORTAGE_REPO_NAME is initialized inside config.setcpv() for both ebuilds and binary packages, so it's always available inside ebuild.sh. svn path=/main/trunk/; revision=12402
* Source bashrc files as late as possible, giving them the opportunity toZac Medico2008-12-021-1/+2
| | | | | | | override as much as possible. Thanks to Ned Ludd <solar@g.o> for the suggestion. svn path=/main/trunk/; revision=12134
* Don't direct to /dev/null when creating $EBUILD_EXIT_STATUS_FILE inside die.Zac Medico2008-11-261-2/+1
| | | | svn path=/main/trunk/; revision=12107
* When using the `read` builtin to split newlines in e* function arguments, useZac Medico2008-11-071-12/+12
| | | | | | | $REPLY in order to ensure that whitespace in each line is correctly preserved. Thanks to Joe Peterson <lavajoe@g.o> for reporting. svn path=/main/trunk/; revision=11820
* Filter PORTAGE_VERBOSE in save_ebuild_env().Zac Medico2008-11-011-1/+1
| | | | svn path=/main/trunk/; revision=11773
* Bug #236609 - Fix columns calculation for TERM="dumb". Thanks to Ulrich MüllerZac Medico2008-10-311-1/+1
| | | | | | <ulm@g.o> for the patch. svn path=/main/trunk/; revision=11764
* Add PORTAGE_QUIET to the environment whitelist and filter it from the saved ↵Zac Medico2008-10-211-0/+1
| | | | | | | | ebuild environment. svn path=/main/trunk/; revision=11708
* Now that elog_base() uses 'echo -e' to expand escape codes prior to usingZac Medico2008-09-301-5/+1
| | | | | | | | 'read' to split on newlines, it's safe to use newlines as delimiters in the log file since 'read' is guaranteed to split any newlines contained in the arguments. svn path=/main/trunk/; revision=11601
* Use echo -e when splitting lines in e* functions, for compatibility withZac Medico2008-09-251-6/+6
| | | | | | | previous behavior wrt \e escape codes such as `elog "\e[31m***\e[0m"` for example. Thanks to Arfrever for finding this solution. svn path=/main/trunk/; revision=11545
* Bug #238251 - Use 'read' instead of $IFS for splitting newlines in e*Zac Medico2008-09-211-67/+7
| | | | | | | functions, since using $IFS causes spurious newlines to be inserted in the message when the e* function is called with more than one argument. svn path=/main/trunk/; revision=11529
* Fix typo in from previous commit.Zac Medico2008-09-191-1/+1
| | | | svn path=/main/trunk/; revision=11516
* Remove everything related to the eapi* default phase functions since theyZac Medico2008-09-181-1/+1
| | | | | | are not included in the current EAPI 2 draft. svn path=/main/trunk/; revision=11515
* Add a new src_prepare phase function which is called in-between src_unpackZac Medico2008-08-231-4/+4
| | | | | | | and src_configure (and bump EAPI to 2_pre3). Thanks to Ciaran McCreesh for the suggestion. svn path=/main/trunk/; revision=11454
* Rename the eapi2_pre2_* functions to just eapi2_* and document all theZac Medico2008-08-081-1/+1
| | | | | | new eapi* functions. svn path=/main/trunk/; revision=11351
* Implement eapi$EAPI_* default phase functions that are equivalent to theZac Medico2008-08-081-1/+1
| | | | | | | | | default_* functions for the given EAPI. For example, a function named eapi0_src_compile provides access to the default src_compile implementation that EAPI 0 provides. This feature is supported only when EAPI=2_pre2 is set. Thanks to Brian Harring for suggesting this idea. svn path=/main/trunk/; revision=11350
* * Fix the default() implementation so that it maps the ebuild argument toZac Medico2008-08-081-0/+23
| | | | | | | | | | | the correct phase function. * Add --exclude-init-phases option to save_ebuild_env() for filtering out unnecessary functions and variables when creating environment.bz2. * Add support for default() and default_* functions that die in all ebuild phases so that improper use of these functions is quickly discovered. svn path=/main/trunk/; revision=11346
* Remove redundant 'local line' definition from previous commit.Zac Medico2008-08-061-1/+0
| | | | svn path=/main/trunk/; revision=11336
* Fix breakage in elog function newline handling with causes blank/emptyZac Medico2008-08-061-5/+43
| | | | | | lines to get lost. svn path=/main/trunk/; revision=11335
* Bug #233735 - Add support for src_configure and default_* functions withZac Medico2008-08-041-3/+6
| | | | | | | new EAPI=2_pre2. Thanks to Arfrever Frehtes Taifersar Arahesis for the initial patch which I've made a few adjustments to. svn path=/main/trunk/; revision=11327
* Even though the message is split on $'\n' in elog_base(), it's stillZac Medico2008-07-281-1/+5
| | | | | | | | not entirely safe to use it as a delimiter in the log file since there can still be escaped newlines that will be expanded due to the echo -e parameter. svn path=/main/trunk/; revision=11244
* Make elog_base() just use \n as the delimiter in the file since it's safeZac Medico2008-07-271-1/+1
| | | | | | now that lines are always split. svn path=/main/trunk/; revision=11220
* Make all the elog functions split lines when displaying output, just likeZac Medico2008-07-271-5/+35
| | | | | | elog_base() does when saving it. svn path=/main/trunk/; revision=11219
* Add support to elog_base() to split messages on newlines automatically.Zac Medico2008-07-271-2/+8
| | | | | | Thanks to Arfrever for the suggestion. svn path=/main/trunk/; revision=11218
generated by cgit v1.2.3-1-g7c22 (git 2.25.1) at 2024-07-02 21:27:40 +0000