summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'master' of git://git.fedorahosted.org/git/fedora-packagerHEADmasterJochen Schmitt2010-09-150-0/+0
|\
| * Don't use log.error from within the libraryJesse Keating2010-09-101-6/+2
| | | | | | | | Just raise it, let the caller figure out what to do.
* | Don't use log.error from within the libraryJesse Keating2010-09-151-6/+2
|/ | | | | | Just raise it, let the caller figure out what to do. This reverts commit 1118325ec5f7aec7b2d50867848f4b2d51c59b41.
* fedpkg: Complete --srpm for the build commandTodd Zullinger2010-09-101-1/+2
|
* fedpkg: Use git plumbing to complete branch namesTodd Zullinger2010-09-101-3/+4
| | | | | The output of 'git branch' is not guaranteed to remain constant. Using the for-each-ref plumbing command should work reliably for longer.
* fedpkg: Remove filenames option from completionTodd Zullinger2010-09-101-1/+1
| | | | | | | | | | | | | | | The bash completion 'filenames' option prevents proper completion of local refs like tmz/f13. On systems with bash-completion >= 1.2, this should have little effect because the _filedir command sets the option as needed (or emulates it on older bash releases that lack compopt). On systems with older bash-completion (like EL-5 currently), there is a minor downside in that filename completion won't append trailing slashes to directories without the 'filenames' option. Ideally, bash-completion will be updated on EL-5 and this minor inconvenience will vanish sometime soon.
* fedpkg: Use _filedir for path completionTodd Zullinger2010-09-101-11/+16
| | | | | The _filedir function is a bit more robust than using compgen directly. A wrapper function is used to filter out .git directories from results.
* Fix up the logger for what goes whereJesse Keating2010-09-101-5/+22
| | | | | | Make sure that WARN, ERROR, and CRITICAL go to stderr, and that DEBUG and INFO go to stdout. This involves more configuration than it aught to, but that's python logging for ya.
* Make our logger print to stdoutJesse Keating2010-09-101-1/+2
| | | | | | It's of my opinion that only errors should go to stderr. For that reason I'll have all logger stuff to go stdout and any errors I'll use print() on.
* Also redirect stderr over to stdout on non-termsJesse Keating2010-09-101-1/+2
| | | | Again this is just because rpmbuild does some stupid things.
* Name the process variables right (#632097)Jesse Keating2010-09-101-2/+2
|
* Minor whitespace fixJesse Keating2010-09-081-1/+1
|
* Remove errornous dots at the end of the help messages of the commitJesse Keating2010-09-081-2/+2
| | | | | | command From Jochen@herr-schmitt.de
* Fefactoring: Use getuser to determinate fas useridJochen Schmitt2010-09-081-25/+23
| | | | | | | | | | | Hello Jesse, because I don't like duplicate code, I have create a getuser() function which determinate the userid of the fas account. Best Regards: Jochen Schmitt
* bash-completion < 1.2 compatibility (tmz@pobox.com)Jesse Keating2010-09-081-1/+8
| | | | | | | | The _get_comp_words_by_ref is provided by bash-completion >= 1.2, which EL-5 lacks. Attempting to use fedpkg completion on EL-5 results in errors: $ fedpkg <TAB>-bash: _get_comp_words_by_ref: command not found Fall back to using ${COMP_WORDS[COMP_CWORD]}.
* fedpkg: Try not to add redundant gitignore entriesTodd Zullinger2010-09-081-1/+11
| | | | | | | | | | | If there is a entry which matches the filename we're about to add already in the gitignore file, we don't need to bother adding another entry. This adds a match() method to GitIgnore which uses fnmatch. Under the hood, this is what git uses. It doesn't catch some of the special cases git has, but it should simply fail to match in those cases and will add a potentially redundant entry -- no different than previously.
* fedpkg: better bash completionJan Vcelak2010-09-081-117/+222
| | | | | - completes koji targets, sytem architectures, git branches, filenames and SRPM names after specific options
* Prep for release0.5.1.4Jesse Keating2010-08-242-2/+6
|
* Fix setting push.default when cloning with dirsJesse Keating2010-08-241-2/+4
|
* Remove build --test option in bash completionJesse Keating2010-08-231-1/+1
|
* Prepare for release0.5.1.3Jesse Keating2010-08-232-2/+16
|
* Install the bash completion script in the packageJesse Keating2010-08-231-0/+1
|
* Split fedpkg into its own subpackageJesse Keating2010-08-231-7/+27
| | | | Also sync up some other spec file changes from the pkg repo
* Error check the update call. #625679Jesse Keating2010-08-231-1/+5
|
* Use the correct remote when listing revsJeroen van Meeuwen (Ergo Project)2010-08-231-2/+5
|
* Add the bash completion fileJesse Keating2010-08-231-1/+5
|
* Send stderr over to the pipe command tooJesse Keating2010-08-231-1/+4
| | | | | | | I don't like doing this, but I'm failing at coming up with a clever way to only do it in cases we know we want it, such as rpmbuild calls. Since that's the only thing we're piping right now, call it good, but I still don't like it.
* Documented bash completion scriptLéon Keijser2010-08-231-3/+15
|
* make fedora-cvs only do anonymous chackouts since cvs is read only now.Dennis Gilmore2010-08-231-7/+1
| | | | We should just remove this at some point soon
* re-fix dist defines.Jesse Keating2010-08-231-2/+2
| | | | One of these times I'll get it right.
* Short cut the failure on repeated buildsJesse Keating2010-08-201-0/+8
| | | | | Saves a bunch of time waiting for the srpm creation and koji to figure out what n-v-r we're trying to build. Ticket #27
* Allow passing srpms to the build commandJesse Keating2010-08-201-0/+2
|
* Set the prog name right.Jesse Keating2010-08-201-0/+1
| | | | This might only effect running from a checkout.
* Add an epilog to help folks outJesse Keating2010-08-201-1/+3
|
* clone: set repo's push.default to trackingHans Ulrich Niedermann2010-08-201-0/+4
| | | | clone: set repo's push.default to tracking
* Make sure we get the right path to the fileJesse Keating2010-08-201-1/+5
|
* Make use of _run_command's cwd option.Jesse Keating2010-08-201-7/+1
|
* Add support for running a command from a specific dirJesse Keating2010-08-201-7/+15
|
* Update docstring for _run_commandJesse Keating2010-08-201-0/+2
|
* Move commit out of the PackageModule object.Jesse Keating2010-08-202-37/+42
| | | | | | Outside of clog, there is no reason why commit should need data from the spec file. This fixes a bug when marking a package as dead so that one can commit again. Ticket #54
* Avoid function calls as default param valuesHans Ulrich Niedermann2010-08-191-8/+26
| | | | | | | | | | | | | A function call used as a default parameter value is evaluated once when the module code is first read in. However, the os.getcwd() function might give different results at run time than at module reading time. This replaces those default param values with None, and has the function bodies call os.getcwd() if the param happens to be None. See also http://docs.python.org/reference/compound_stmts.html#function-definitions
* Add some new commands to the bash completionJesse Keating2010-08-191-10/+10
|
* Fix whitespaceJesse Keating2010-08-191-1/+1
|
* Add bash-completion script for fedpkgLéon Keijser2010-08-191-0/+134
|
* pull the username from fedora_cert to pass to bodhiDennis Gilmore2010-08-191-5/+11
|
* Catch double ^c's from build. RHBZ #620465Jesse Keating2010-08-171-1/+4
|
* Fix up chain buildingJesse Keating2010-08-172-10/+21
| | | | | We were not setting up the chains right, and our logging was all funky. We now handle chains properly and log it sensibly
* Add missing process call for non-pipe no tty.Jesse Keating2010-08-121-0/+5
|
* update to 0.5.1.20.5.1.2Dennis Gilmore2010-08-123-2/+10
|
* Trap attempts to make git repos. #622716Jesse Keating2010-08-121-4/+10
|
generated by cgit (git 2.34.1) at 2024-04-19 07:23:08 +0000