| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
Implement the following global fedpkg options for passing on
to rpmbuild, regardless of whether fedpkg is directly calling
rpmbuild or calling it indirectly via mock:
-D 'MACRO EXPR', --define 'MACRO EXPR'
define an rpm macro for rpmbuild and mock
--with RPMWITH enable configure option for rpmbuild and mock
--without RPMWITHOUT disable configure option for rpmbuild and mock
|
| |
|
|
|
|
|
|
|
|
|
|
| |
It doesn't make sense to explicitly pass the very same set of
multiple args.foo variables like
PackageModule(foo = args.foo, ...)
to PackageModule.__init__(), so we just pass the complete args
object.
|
| |
|
|
|
|
|
| |
Handle the case where the last set is inclusive of the last build or
not.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When doing chain builds, the last package can be treated one of two
ways.
1) It can be added to the end of the list to be built serially, it would
be the last build.
2) It can be added to the last set in the chain to be built in parallel
with the last set.
Which way depends on if sets are used or not. This code should handle
both correctly.
|
|
|
|
| |
This was added a while ago but I forgot to document it.
|
|
|
|
|
|
|
|
|
|
|
| |
This changes the parser_help command to a lambda which passes
the parser variable to the usage() function as a second
parameter.
We forgot to give the usage() function its "parser" parameter,
and after we moved the definition of the parser variable out
of the main module, the usage() function did not see the parser
variable any more.
|
|
|
|
|
| |
Turns out there could be multiple reasons for wanting to make a new srpm
even if the spec file hasn't changed, so just do it.
|
| |
|
|
|
|
|
| |
This will be handy when we start dealing with multiple remotes. Not
perfect but a good start.
|
|
|
|
| |
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
|
|
|
|
|
|
|
|
|
|
| |
Make sure that the fedpkg.1 man page file is rebuilt when any of
the files it may pull information from have been changed.
This will even cause the `fedpkg.1` rebuild (and later
'make install' the rebuilt man page) in case of an old and
outdated fedpkg.1 lying around in $(srcdir) when doing an out
of source tree build (e.g. $(top_srcdir) != $(top_builddir)).
|
| |
|
|
|
|
|
|
|
| |
Strip doesn't do what I thought it did, it will remove any of the
individual chars from the string, rather than just removing the
substring from the string. replace('substring', '') needs to be used
instead.
|
|
|
|
|
|
| |
Not only does this now work with old/new style branches, but it also
fixed up some improper regex use that could have caught too much. It
can be simplified once the branch renames are done.
|
| |
|
|
|
|
|
|
| |
This makes use of switch_branch to clone to a specific branch. This
might have the side effect of "fixing" el5 where clone --branch is not
supported by git.
|
|
|
|
|
| |
This also removes a hardcoded reference to "origin", as this can be
named differently.
|
|
|
|
|
| |
These new regexes are used to catch both styles of branching. The old
branchfilter is no longer used.
|
|
|
|
|
| |
This also catches only top level Fedora branches, without the trailing
/master
|
| |
|
| |
|
|
|
|
|
| |
Rpm can be bitchy, but also return 0, so handle that. Also try to do
something meaningful / useful with the error output.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
Move module imports of the fedora_cert, koji, and pyfedpkg
modules to the actual command runtime.
This gets rid of the requirement to have them (and their
respective imported modules) installed at build time for
generating the man page.
|
|
|
|
|
| |
Move actual man page generation call from Makefile.am
to the src/fedpkg_man_page.py __main__ section.
|
|
|
|
|
|
| |
Move location/name of man_page module (out of the
pyfedpkg module/namespace) in order to avoid
importing pyfedpkg for man page generation.
|
|
|
|
| |
This apparently disappeared in commit 582c0e69.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
usage: fedpkg verify-files [-h]
Locally run 'rpmbuild -bl' to verify the spec file's %files sections. This is
useful after a successful run of 'fedpkg install' or after a 'fedpkg local'
run which failed due to %files list inaccuracies.
|
| |
|
| |
|
| |
|
|
|
|
| |
I decided to emulate git here and not default to a rebase.
|
| |
|
|
|
|
|
|
|
|
| |
This feeds both command --help and the man page. Formatting is a little
odd. The help = on the parser itself cannot span multiple lines with \.
It'll look fine in --help but not in the man page. Multiple strings
need to be used instead. The description argument can span multiple
lines with \ just fine though.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
When dealing with the pipe stuff, we need to catch if the first command
exits with an error, because that error won't carry through to the piped
command. Fixes RHBZ:634888
|
|
|
|
| |
Want to use the COMMAND OVERVIEW option instead.
|
|
|
|
| |
Add man page generator to generate a man page at build time.
|
| |
|
| |
|