W. Trevor King [Tue, 1 Jun 2010 17:36:42 +0000 (13:36 -0400)]
Convert update_copyright.py to a more VCS-agnostic version.
This will make it easier for others to grab this code for their own
projects. Currently supported VCSs: Git, Mercurial, Bazaar.
W. Trevor King [Tue, 1 Jun 2010 17:33:24 +0000 (13:33 -0400)]
Use REPORT_UDIFF doctest option in test/curve_info.py
W. Trevor King [Tue, 1 Jun 2010 17:32:43 +0000 (13:32 -0400)]
Cleanups in hooke.driver.picoforce
Scaling adjustment:
* Use 'z piezo gain' in bit->m conversion (is this correct?)
Info field renames for clarity and consistency:
* 'z piezo magnification' -> 'z piezo gain' (is this correct?)
* 'z piezo scale (V/bit)' -> 'z piezo sensitivity (V/bit)'
* 'z piezo scale (V)' -> 'z piezo range (V)'
* 'raw-data' -> 'raw data'
Also:
* Stripped trailing whitespace.
* Comment typo correction 'inzrease' -> 'increase'
W. Trevor King [Tue, 1 Jun 2010 16:50:16 +0000 (12:50 -0400)]
Don't die when flat-filtering curves shorter than blind window.
This will probably only happen if the user uses an absurdly large
blind window by mistake, but that shouldn't crash Hooke.
W. Trevor King [Tue, 1 Jun 2010 15:42:53 +0000 (11:42 -0400)]
Correct relative path generation in hooke.playlist.FilePlaylist.flatten
W. Trevor King [Tue, 1 Jun 2010 14:46:26 +0000 (10:46 -0400)]
Fix /doc/build/ -> ./doc/build/ in update_copyright.py
W. Trevor King [Wed, 19 May 2010 10:51:06 +0000 (06:51 -0400)]
Add another parameter-casting hack.
W. Trevor King [Wed, 19 May 2010 10:50:41 +0000 (06:50 -0400)]
Confirm lack of peaks in test/data/vclamp_picoforce/
W. Trevor King [Wed, 19 May 2010 10:16:27 +0000 (06:16 -0400)]
Use numpy.savetxt (handles multiple columns) rather than ndarray.tofile
W. Trevor King [Wed, 19 May 2010 10:06:57 +0000 (06:06 -0400)]
Add test.flat_filter which has been driving the recent fixups
W. Trevor King [Wed, 19 May 2010 10:05:18 +0000 (06:05 -0400)]
Adjust path (if it exists) for the child playlist in FilterCommand
W. Trevor King [Wed, 19 May 2010 09:56:26 +0000 (05:56 -0400)]
Apply 'name' argument in plugin.playlist.FilterCommand
W. Trevor King [Wed, 19 May 2010 09:53:18 +0000 (05:53 -0400)]
Fix typo .add -> .append in plugin.playlist.FilterCommand
W. Trevor King [Wed, 19 May 2010 09:52:25 +0000 (05:52 -0400)]
Initialize 'min peaks' from configuration (if required) in convfilt/flatfilt
W. Trevor King [Wed, 19 May 2010 09:41:42 +0000 (05:41 -0400)]
Fix some outstanding argument issues in flatfilt.
Also adjust temporary contact point value in vclamp.scale to ensure
'blind window' fits.
W. Trevor King [Wed, 19 May 2010 09:37:34 +0000 (05:37 -0400)]
Input codec messes with readline.
Disabled in hooke.ui.commandline while I work the kinks out of
flatfilt.
W. Trevor King [Wed, 19 May 2010 08:55:53 +0000 (04:55 -0400)]
Move 'blind window' and 'min peaks' from FlatPeaksCommand to FlatFiltPlugin
W. Trevor King [Wed, 19 May 2010 08:50:39 +0000 (04:50 -0400)]
Success import and better peaks validation in flatfilt and convfilt.
W. Trevor King [Wed, 19 May 2010 08:42:20 +0000 (04:42 -0400)]
array.abs() -> numpy.abolute(array) in hooke.util.peak
W. Trevor King [Wed, 19 May 2010 08:39:32 +0000 (04:39 -0400)]
Removed trailing comma that turned 'median' into a tuple
W. Trevor King [Wed, 19 May 2010 08:39:04 +0000 (04:39 -0400)]
Added default name to hooke.util.peak.mask_to_peaks
W. Trevor King [Wed, 19 May 2010 08:08:42 +0000 (04:08 -0400)]
Fix 'min deviation' -> 'min deviations' typos.
In hooke.util.peak and hooke.plugin.convfilt/flatfilt.
Also:
* Fix "std == data.std" typo in hooke.util.peak.above_noise
* Pass mean and std from noise to above_noise in find_peaks.
W. Trevor King [Wed, 19 May 2010 07:58:47 +0000 (03:58 -0400)]
hooke.util.peak.find_peaks should *not* die on noise nonconvergence
W. Trevor King [Wed, 19 May 2010 07:57:50 +0000 (03:57 -0400)]
Clean up argument handling in hooke.plugin.convfilt/flatfilt.
* Take advantage of hooke.util.peak._kwargs to split out appropriate
kwargs for find_peaks.
* Add missing arguments 'blind window' and 'median window' to
FlatFiltPlugin.
* Fix leftover convfilt code in FlatFiltPlugin._run peak naming.
* Add type casting HACK to FlatFiltPlugin._setup.
W. Trevor King [Wed, 19 May 2010 07:41:37 +0000 (03:41 -0400)]
Added input key flexibility to hooke.util.peak._kwarg
W. Trevor King [Wed, 19 May 2010 07:17:52 +0000 (03:17 -0400)]
Restore plugin setup in hooke.command.Command.__init__
W. Trevor King [Wed, 19 May 2010 07:15:23 +0000 (03:15 -0400)]
Added preliminary hooke.plugin.vclamp.scale (just def. m -> N conversion).
W. Trevor King [Wed, 19 May 2010 06:49:39 +0000 (02:49 -0400)]
Import VelocityClamp in hooke.plugin.flatfilt for ConvolutionPeakCommand._setup
W. Trevor King [Wed, 19 May 2010 06:47:37 +0000 (02:47 -0400)]
Allow negative voltages in hooke.driver.picoforce.PicoForceDriver._translate_block_info
W. Trevor King [Wed, 19 May 2010 06:46:44 +0000 (02:46 -0400)]
Fix self.hooke -> hooke typo in hooke.plugin.flatfilt.FlatFilterCommand.filter
W. Trevor King [Wed, 19 May 2010 06:40:52 +0000 (02:40 -0400)]
Fix broken hooke.plugin.playlist.playlist_name_callback
W. Trevor King [Wed, 19 May 2010 06:40:24 +0000 (02:40 -0400)]
Remove if __name__ == '__main__' section of hooke.hooke
W. Trevor King [Wed, 19 May 2010 06:31:00 +0000 (02:31 -0400)]
Add unicode-capable IO to hooke.ui.commandline.
The previous implementation produced:
$ python bin/hooke -c 'flat_filter_playlist'
...
hooke> Traceback (most recent call last):
...
UnicodeEncodeError: 'ascii' codec can't encode character u'\xec'...
The new encoding detection functions in hooke.util.encoding came from
my libbe.util.encoding module for Bugs Everywhere. From my original
BE commit (Tue 2008-11-25 15:47:19 -0500):
get_encoding() is from Trac
http://trac.edgewall.org/browser/trunk/trac/util/datefmt.py
format_datetime()
Trac has a BSD license
http://trac.edgewall.org/wiki/TracLicense
I don't know if such a small snippet requires us to "reproduce the
above copyright" or where we need to reproduce it if it is needed.
The stdout/stdin replacement code follows
http://wiki.python.org/moin/ShellRedirectionFails
References:
http://wiki.python.org/moin/Unicode
http://www.amk.ca/python/howto/unicode
http://www.python.org/dev/peps/pep-0100/
W. Trevor King [Wed, 19 May 2010 05:11:15 +0000 (01:11 -0400)]
Remove two debugging prints from hooke.ui.commandline
W. Trevor King [Wed, 19 May 2010 05:03:30 +0000 (01:03 -0400)]
Adjust update_copyright hacks to follow plugin.peakspot -> util.peak move
W. Trevor King [Wed, 19 May 2010 04:54:48 +0000 (00:54 -0400)]
Major plugin restructuring.
The restructuring makes major improvements in:
* numpy/scipy utilization (for speed)
* functional encapsulation (breaking long, complicated functions
into easily understood pieces, and sharing those pieces between
related Commands).
Renames:
* generalclamp -> fclamp
* generalvclamp -> vclamp
* generaltccd -> tccd
* peakspot -> hooke.util.peak
Splits:
* flatfilts -> flatfily
`-> convfilt
The config files convfilt.conf and flatfilts.ini were also
internalized into their respective plugins' .default_settings().
W. Trevor King [Wed, 19 May 2010 04:41:11 +0000 (00:41 -0400)]
Remove debugging print command from hooke.util.calculus.derivative
W. Trevor King [Wed, 19 May 2010 04:38:51 +0000 (00:38 -0400)]
Fix a pickling problems with hooke.plugin.playlist.FilterCommand.
Pickle doesn't like functions that aren't defined the top level of
modules (such as NoteFilterCommand's old lambda filter_fn). Add a
warning to that effect to the FilterCommand doc string, and rearrange
to encourage the use of methods (which are less likely to be defined
on the fly than functions).
Also:
* Added pickle queue debugging code to hooke.hooke. This isn't the
first tiem I've had a pickling problem. This code (when enabled),
makes it easier for me to track down the failing code.
* Fix a typo in hooke.hooke.HookeRunner.run (self.ui -> hooke.ui).
* Remove a snip of trailing whitespace in hooke.hooke.
W. Trevor King [Tue, 18 May 2010 16:35:22 +0000 (12:35 -0400)]
Allow extra arguments to NoteIndexList.filter's filter_fn
W. Trevor King [Tue, 18 May 2010 14:50:18 +0000 (10:50 -0400)]
Move argument_to_setting before plugin loading code in hooke.plugin.
Otherwise Python loads the plugins before argument_to_setting has been
defined, which makes it hard for the plugins to use the function ;).
W. Trevor King [Tue, 18 May 2010 14:44:29 +0000 (10:44 -0400)]
Add Numpy crossreferences to hooke.util.fft
W. Trevor King [Tue, 18 May 2010 14:43:20 +0000 (10:43 -0400)]
Introduce intersphinx to Numpy, Scipy, and Matplotlib
W. Trevor King [Tue, 18 May 2010 12:59:16 +0000 (08:59 -0400)]
Add hooke.plugin.argument_to_setting to convert Arguments -> Settings.
This will make it easy for Plugins to generate Settings matching
Command Arguments.
W. Trevor King [Tue, 18 May 2010 12:54:38 +0000 (08:54 -0400)]
Actually, pass plugin instance to Command.__init__.
Rather than using Plugin._setup_commands, because the commands may
wish to use some aspects of the Plugin during initialization (e.g. to
build argument lists from plugin settings). The previous
implemetation gave the commands no way to respond to the attached
plugin instance.
W. Trevor King [Tue, 18 May 2010 11:34:02 +0000 (07:34 -0400)]
Rework Plugin.commands() to include _setup_commands().
The previous implementation gave Commands no way to access the
Plugin's configuration.
W. Trevor King [Tue, 18 May 2010 02:04:06 +0000 (22:04 -0400)]
Fix syntax errors in hooke.plugin.curve
W. Trevor King [Tue, 18 May 2010 01:58:57 +0000 (21:58 -0400)]
Finish translating the hooke.plugin.curve Commands from procplots.
W. Trevor King [Mon, 17 May 2010 22:38:33 +0000 (18:38 -0400)]
Converted hooke.util.fft to Hooke coding style.
* Add numpydoc docstrings (to most functions)
* Converted hackish test suites to hackish `unittest.TestCase`s.
Still ugly, but now `nosetests` finds them.
They should really use .assertAlmostEqual and company...
* Added neccessary function definitions to doc/conf.py for the
'math' sections of fft's docstrings.
W. Trevor King [Mon, 17 May 2010 20:27:20 +0000 (16:27 -0400)]
Add intersphinx_mapping to doc/conf.py
This tell's intersphinx where to find documentation for other
projects. Currently we've only told it about the standard Python
libraries.
http://sphinx.pocoo.org/latest/ext/intersphinx.html#confval-intersphinx_mapping
W. Trevor King [Mon, 17 May 2010 20:18:46 +0000 (16:18 -0400)]
Add Sphinx extensions sphinx.ext.intersphinx sphinx.ext.pngmath.
* intersphinx allows documentation linking between known Sphinx packages.
http://sphinx.pocoo.org/ext/intersphinx.html
* pngmath allows LaTeX inclusion via the math directive
http://sphinx.pocoo.org/ext/math.html
Also reorder extension list following:
sphinx.* alphabetically
everything else aphabetically
W. Trevor King [Mon, 17 May 2010 19:45:20 +0000 (15:45 -0400)]
Ran ./update_copyright.py to setup copyrights in caclulus and fft.
Also corrects a name typo in th picoforce copyright. Sometimes I miss
keystrokes and Emacs starts running aspell without me noticing :p.
W. Trevor King [Mon, 17 May 2010 19:38:31 +0000 (15:38 -0400)]
Adjusted hooke.util.calculus.derivative.__doc__ so Sphinx doesn't die.
W. Trevor King [Mon, 17 May 2010 19:08:29 +0000 (15:08 -0400)]
Merged my unitary FFT wrappers (FFT_tools) as hooke.util.fft.
Origininally versioned in Git and licensed under BSD, but I wrote
them, so I'm relicensing under the LGPL for Hooke.
The original git repository was:
http://www.physics.drexel.edu/~wking/code/git/FFT_tools.git/
W. Trevor King [Mon, 17 May 2010 19:05:09 +0000 (15:05 -0400)]
Relicensed FFT_tools.py under Hooke's LGPL
W. Trevor King [Mon, 17 May 2010 18:40:02 +0000 (14:40 -0400)]
Merged hooke.plugin.plotmanip into hooke.plugin.curve.
Also:
* Fixed a number of typos in hooke.plugin.curve.
* Rewrote derivative code for Numpy in
hooke.util.calculus.derivative.
* Removed FFT code (I'm replacing it with my FFT_tools module
shortly).
* Moved v-clamp-specific code into generalvclamp.
* Moved plotmanips to hooke.plugin.plotmanip holding area.
W. Trevor King [Mon, 17 May 2010 14:33:03 +0000 (10:33 -0400)]
Remove leftover curve.InfoCommand cruft from ExportCommand.__init__.
W. Trevor King [Mon, 17 May 2010 14:29:41 +0000 (10:29 -0400)]
Moving hooke.plugin.procplots to .curve in preparation for merge
W. Trevor King [Mon, 17 May 2010 13:18:28 +0000 (09:18 -0400)]
Started translating hooke.plugin.multifit to new Plugin/Command architecture.
W. Trevor King [Mon, 17 May 2010 12:30:20 +0000 (08:30 -0400)]
Add REAMDE and playlist.hkp for Hemingway force clamp curves.
W. Trevor King [Mon, 17 May 2010 12:29:40 +0000 (08:29 -0400)]
Added sample Hemingway force-clamp curves
W. Trevor King [Mon, 17 May 2010 12:28:43 +0000 (08:28 -0400)]
Moved test/data/titin to more descriptive test/data/vclamp_picoforce
W. Trevor King [Mon, 17 May 2010 12:18:28 +0000 (08:18 -0400)]
Add REAMDE and playlist.hkp for titin I-27 curves.
W. Trevor King [Mon, 17 May 2010 12:17:46 +0000 (08:17 -0400)]
Added 100 sample velocity-clamp curves for titin I-27
W. Trevor King [Mon, 17 May 2010 12:02:21 +0000 (08:02 -0400)]
Fix rief1997 footnote in hooke.experiment.VelocityClamp
W. Trevor King [Mon, 17 May 2010 12:00:36 +0000 (08:00 -0400)]
Updated hooke.driver.picoforce to new Driver class.
This is a fairly major rewrite of the previous PicoForce driver. I've
cleaned up header parsing and tried to break trace -> block conversion
up into compact functions. The meaning of many PicoForce file
constants is still unclear to me though, so I may have gotten some
things wrong.
Also:
* Added references for each hooke.experiment.Experiment subclass.
* TutorialDriver sets curve.info['experiment'] to an Experiment subclass.
* An adjusted test.curve_info now passes.
While we're translating the old experiment strings to the new
Experiment classes, the following conversion table may be useful:
'clamp' -> ForceClamp
'smfs' -> VelocityClamp
'smfluo' -> TwoColorCoincidenceDetection
W. Trevor King [Sun, 16 May 2010 18:35:01 +0000 (14:35 -0400)]
Fixed bugs in hooke.curve.Data.__new__ and added doctests
W. Trevor King [Sun, 16 May 2010 16:00:56 +0000 (12:00 -0400)]
Add Curve loading to hooke.playlist.Playlist.
Highlights:
* NotRecognized.__get/setstate__ for Pickling through the Queues.
Perhaps this would be better solved in Failure.__get/setstate__?
The current fix works for the moment though.
* Use Curve.data == None to indicate unloaded curves.
* Fix TutorialDriver.is_me(self) -> is_me(self, path)
* Added ._load* to Playlist, so it only caches ._max_loaded curves
in memory.
* Fix broken Playlist.from_string doctest.
W. Trevor King [Sun, 16 May 2010 14:56:06 +0000 (10:56 -0400)]
Moved doc/hooke.jpg to doc/img/hooke.jpg
W. Trevor King [Sun, 16 May 2010 14:49:56 +0000 (10:49 -0400)]
Add hooke.plugin.curve (moved from hooke.plugin.curvetools)
Provides 'curve info' and 'export' commands. Added test.curve_info
to test the 'curve info' command, but it currently fails due to
unimplemnted Curve loading.
Also:
* Minor docstring cleanups in hooke.plugin.playlist
* Moved some functions to hooke.plugin.generalvclamp. I'm not clear
on Experiment-specific Curve interpretation yet, but these
functions (which will eventually be Commands) seem better placed
here.
* fit_interval_nm
* find_current_peak
* pickup_contact_point
* baseline_points
W. Trevor King [Sun, 16 May 2010 13:56:00 +0000 (09:56 -0400)]
Added .hooke.cfg to .hgignore.
The local config copy auto-generated by hooke.hooke.main(). We don't
want anyone adding this to the versioning system by accident.
W. Trevor King [Sun, 16 May 2010 13:53:53 +0000 (09:53 -0400)]
Broke out Hooke.run and Hooke.run_lines into HookeRunner.
This allows the main process access to the command process' possibly
altered Hooke instance. For example:
>>> h = Hooke()
>>> r = HookeRunner()
>>> h = r.run_lines(h, ['load_playlist test/data/test'])
>>> h = r.run_lines(h, ['curve_info --all=True'])
With the old implementation, second run_lines() call would fail
because the playlist was only loaded in the first command process'
Hooke instance, not the Hooke instance that the main (UI) process
held.
W. Trevor King [Sun, 16 May 2010 13:29:51 +0000 (09:29 -0400)]
Pull Hooke.run*() multiprocessing setup/teardown into ._setup/_cleanup_run.
Limits code duplication between .run() and .run_lines().
W. Trevor King [Sun, 16 May 2010 12:21:20 +0000 (08:21 -0400)]
Add hooke.playlist.NoteIndexList.__str__() method.
Now playlists etc. will flatten themselves to strings with:
'...%s...' % instance.
W. Trevor King [Sun, 16 May 2010 11:49:54 +0000 (07:49 -0400)]
Unset executible bits in curvetools.py and jumpstat.py permissions.
W. Trevor King [Sun, 16 May 2010 11:26:28 +0000 (07:26 -0400)]
Ran update_copyright.py, updating all the copyright blurbs and adding AUTHORS.
W. Trevor King [Sun, 16 May 2010 11:26:00 +0000 (07:26 -0400)]
Use ALIASES and AUTHOR_HACKS handling in update_copyright.update_authors.
The previously generated AUTHORS file was the raw list of VCS
commiters.
W. Trevor King [Sun, 16 May 2010 11:18:31 +0000 (07:18 -0400)]
Updated copyright blurbs in all files to '# Copyright'
This tag will be replaced by update_copyright.py with the full
copyright blurb for each file.
W. Trevor King [Sun, 16 May 2010 11:18:00 +0000 (07:18 -0400)]
Further author/year workarounds in update_copyright.py.
W. Trevor King [Sun, 16 May 2010 10:51:39 +0000 (06:51 -0400)]
Add AUTHOR_HACKS and YEAR_HACKS to update_copyright.py
This works around holes in the versioning where files were initially
released before being added to the versioning system, or they were
never committed under the name of one of the authors.
W. Trevor King [Sat, 15 May 2010 12:25:23 +0000 (08:25 -0400)]
Update canonical copyright names for 'illysam' and 'pancaldi.paulo'.
Also add the missing '.82' to Fabrizio's alias.
W. Trevor King [Sat, 15 May 2010 11:58:48 +0000 (07:58 -0400)]
Fix some doctests in update_copyright.py.
And make some adjustments so that update_copyright.py can update its
own copyright blurb without changing anything else inside itself.
W. Trevor King [Fri, 14 May 2010 18:40:14 +0000 (14:40 -0400)]
Rearrange license text and add update_copyright.py.
Following http://www.gnu.org/licenses/gpl-howto.html
* Moved LGPLv3 text from LICENSE -> COPYING.LESSER
* Brought in GPLv3 text -> COPYING
update_copyright.py automates the maitenance of boilerplate copyright
information at the head of every file.
W. Trevor King [Fri, 14 May 2010 15:32:17 +0000 (11:32 -0400)]
Added integration test directory test/ and load_playlist test.
Run all integration tests with:
nosetests --with-doctest --doctest-tests test/
W. Trevor King [Fri, 14 May 2010 15:19:08 +0000 (11:19 -0400)]
Updated .hgignore to ignore documentation build files.
W. Trevor King [Fri, 14 May 2010 14:40:53 +0000 (10:40 -0400)]
Added .hgignore masking *.pyc files
W. Trevor King [Fri, 14 May 2010 14:39:58 +0000 (10:39 -0400)]
Added distutils-based setup.py.
W. Trevor King [Fri, 14 May 2010 13:22:52 +0000 (09:22 -0400)]
Fix git->hg and wget -o -> -O in doc/install.txt snapshot command.
Path change because of Git->Mercurial migration.
wget
-O --output-document sets the document file.
-o --output-file sets the logfile.
W. Trevor King [Fri, 14 May 2010 13:17:04 +0000 (09:17 -0400)]
Updated fork project info for Rolf and my branches in doc/install.txt.
I think I found where Rolf's code is (SVN illysam branch), and I
migrated my Git-versioned fork to Mercurial.
W. Trevor King [Thu, 13 May 2010 14:41:09 +0000 (10:41 -0400)]
Add --script and --command to bin/hooke
W. Trevor King [Thu, 13 May 2010 14:08:54 +0000 (10:08 -0400)]
Save playlist digest on loading.
The previous implementation felt that playlists were not saved
(.is_saved() -> False) when the playlists had not been altered since
loading.
We *don't* save the digest when we load from a string, because
that string is not necessarily on disk and we *do* want to save
it, even if the playlist hasn't change since loading.
I also removed the return bits, since FilePlaylist._from_xml_doc()
doesn't return anything.
W. Trevor King [Thu, 13 May 2010 11:41:46 +0000 (07:41 -0400)]
Actually set self.name in hooke.playlist.FilePlaylist.set_path
W. Trevor King [Thu, 13 May 2010 11:40:00 +0000 (07:40 -0400)]
Fix str() bug in hooke.ui.commandline.LocalExitCommand save msg.
Old implemtation crashed if Playlist.name was None.
W. Trevor King [Thu, 13 May 2010 11:34:41 +0000 (07:34 -0400)]
Move hooke/test/default.000 -> test/data/picoforce.000
+ corresponding update to test/data/test.hkp.
W. Trevor King [Thu, 13 May 2010 11:33:08 +0000 (07:33 -0400)]
Actually set self.path in hooke.playlist.FilePlaylist.set_path
W. Trevor King [Thu, 13 May 2010 11:27:58 +0000 (07:27 -0400)]
Add all_drivers_callback to hooke.plugin.playlist.LoadCommand.
I can't imagine the user not wanting to use all available playlists.
They still can, but specifiying a sublist is now optional.
W. Trevor King [Thu, 13 May 2010 11:24:52 +0000 (07:24 -0400)]
Fix typos in hooke.playlist.FilePlaylist.set_path
W. Trevor King [Thu, 13 May 2010 10:55:29 +0000 (06:55 -0400)]
Fix bug in hooke.ui.commandline.DoCommand._check_argument_length() max_arg calculation.
W. Trevor King [Thu, 13 May 2010 10:48:13 +0000 (06:48 -0400)]
Moved test and mfp_igor_scripts directories around
W. Trevor King [Thu, 13 May 2010 10:31:56 +0000 (06:31 -0400)]
Split out "Troubleshooting" into doc/bugs.txt.
This section will probably grow ;).
+ other textual cleanups.
W. Trevor King [Thu, 13 May 2010 10:16:50 +0000 (06:16 -0400)]
Fix doc/index.txt references to auto-generated pages
:ref:`~...` -> :ref:`...` following
http://sphinx.pocoo.org/_sources/contents.txt