Commits

Author Commit Message Labels Comments Date
dr...@kaverne
merge old fix I just needed again.
dr...@kaverne
wrappers: resilience: do not die if the dstrepo is None.
Augie Fackler
layouts: add importerror ladder so hgsubversion works as an installed module
Greg Ward
layouts: fix crash when importing hgsubversion itself (issue #402) Mercurial extensions are a bit weird: they aren't normally in sys.path, so you can't assume that "import hgsubversion" works. Luckily, Mercurial sneaks a little treat into sys.modules so that "import hgext_hgsubversion" does work. In fact, to get things working *as a Mercurial extension*, all that's needed is that trivial change to two import lines, in layouts/detect.py and layouts/standard.py. U…
Siddharth Agarwal
svnexternals: import pathutil.canonpath if available hgsubversion broke with upstream Mercurial due to f962870712da. This fixes that.
David Schleimer
pull: correctly handle replacing the root of a branch with a non-copied directory Prior to this diff, we would either crash, or continue past the replacement without actually recording the change. This could lead to later failing varify if the state before and after weren't identical.
Augie Fackler
test_fetch_mappings: trim some superfluous whitespace
maugustin
authormap: allow case-insensitive authormaps for easier conversions
David Soria Parra
verify: use mercurials worker API to speedup stupid verify We use mercurial internal worker api to distribute the workload during verify. This helps with larger repositories. with patch, intel i5 dualcore hyperthreading: $ time hg svn verify hg svn 169.47s user 21.27s system 373% cpu 51.050 total before: $ time hg svn verify hg svn 100.48s user 11.85s system 99% cpu 1:52.97 total This makes stupid verify as fast as subvertpy verify. We …
David Schleimer
buildmeta: remove debugging statement
David Schleimer
layouts: add support for an infix between tbt and the hg root
David Schleimer
layouts: add support for configuring branches directory This should work for both alternately named directories (e.g. releases instead of branches), as well as subdirs (branches/releases), and combinations thereof (releases/public).
David Schleimer
layouts: add custom layout for those of us that need weird mappings This adds a config-driven custom layout, targeted at the case where you need to fetch a small subset of a large number of subversion branches, or where your subversion layout doesn't match the standard trunk/branches/tags layout very well.
David Schleimer
layouts: teach layout detection from file about alternate layouts
David Schleimer
layouts: fix remotepath for default branch in standard layout
David Schleimer
layouts: correct flubbed __unimplement call in base layout
David Schleimer
stupid: only check branch mappings for discovered files once For whatever reason, we were processing the list of discovered files for new branches every time we discovered a new file. This pulls the process-discovered-files loop out of the discover-files loop, and should reduce the running time from O(num_paths_discovered * num_files_discovered) to O(num_paths_discovered + num_files_discovered)
David Schleimer
stupid: cleanup unnecessary, always-true if statement We had an unnecessary if statement in setupid.branches_in_paths() that was checking a variable that was unconditionally set to True on the previous line. This was a remnant of a never-completed, and now mostly-cleaned up attempt to short-circuit path type detection before talking to subversion for some directories. This removes the variable assignment and if statement, and moves the body of the …
David Schleimer
stupid: kill some dead and no longer needed code While I was debugging an unrelated issue with stupid, some of the code in branches_in_paths confused me, since it was using an undeclared variable. It looked like an attempt to short-circuit the file/directory detection before talking to subversion that never quite got finished. The code is mostly unreachable, and obviously broken, so I cleaned it up. I also cleaned up some prepatory code that …
David Schleimer
tests: unbreak running test_updatemeta individually
Dan Villiom Podlaski Christiansen
test_util: include Subversion revision when drawing a pretty graph
Dan Villiom Podlaski Christiansen
subvertpy wrapper: printable path adapter
Dan Villiom Podlaski Christiansen
push: allow obtaining the traceback for Subversion exceptions
David Schleimer
tests: cd to tempdir during setUp Since we were running hg clone from the hgsubversion repo, we would often pick up config settings from hgsubversion/.hg/hgrc, including loading extensions. This caused tests to fail for me when they asserted that obsolete is disabled because I have evolve enabled in my hgsubversion checkout. This was probably decreasing the overall stability of the test suite as well, since it meant that clones …
Augie Fackler
test_util: fix missing message in SkipTest raise
Dan Villiom Podlaski Christiansen
svnmeta: factor out a property for obtaining the tag locations
Dan Villiom Podlaski Christiansen
tests: remove ability to toggle stupidity on less than class level
Dan Villiom Podlaski Christiansen
test_utility_commands: use stupid mode metaclass
Dan Villiom Podlaski Christiansen
test_tags: use stupid mode metaclass
Dan Villiom Podlaski Christiansen
test_single_dir_push: use stupid mode metaclass
  1. Prev
  2. Next