- Nov 19, 2012
-
-
Ryan Fisher authored
-
- Nov 14, 2012
-
-
Duncan Macleod authored
-
- Nov 01, 2012
-
-
Chris Pankow authored
-
- Oct 24, 2012
-
-
Kipp Cannon authored
-
- Oct 22, 2012
-
-
Kipp Cannon authored
-
Kipp Cannon authored
-
- Oct 21, 2012
-
-
Kipp Cannon authored
this patch makes further progress on the problem of module imports having side effects. this removes the endElement() handlers from the content handler class to the respective element classes themselves. the startElement() handlers are now explicitly told the parent element so none of them need to access the guts of the content handler object (which, in principle, allows them to be removed from the content handler as well, but that is not done at this time). this clean-up makes plugging in support for other namespaces possible.
-
Kipp Cannon authored
ligolw_cut exercises the FilteringLIGOLWContentHandler class
-
Kipp Cannon authored
This reverts commit 9915ea92d3d443f55f8720733edb0eb5c8c73174.
-
Kipp Cannon authored
-
Kipp Cannon authored
-
- Oct 19, 2012
-
-
Kipp Cannon authored
This stress-tests glue.ligolw's ability to work with multiple namespaces.
-
Kipp Cannon authored
-
Kipp Cannon authored
-
- Oct 17, 2012
-
-
Kipp Cannon authored
-
Ian Harry authored
(cherry picked from commit 50dea3041a72580443c17c627844899a48c55e89)
-
Kipp Cannon authored
-
- Oct 16, 2012
-
-
Kipp Cannon authored
-
Kipp Cannon authored
This patch calls glue.ligolw.lsctables.instrument_set_from_ifos() to convert the "observatory" column in the CacheEntry to a sequence of instrument names.
-
- Oct 15, 2012
-
-
Ian Harry authored
Fixing some int/int issues in the multi_inspiral part of lsctables and correcting the null SNR calculation (cherry picked from commit 0071584bb3e9f482f765db008e6b12e1ab9a545b)
-
- Oct 08, 2012
-
-
Ryan Fisher authored
Changing inserts to use simplejson to decode over cjson because Debian segment database clients use updated cjson that escapes the solidus, where SL6 clients do not. Simplejson decodes both correctly.
-
- Oct 02, 2012
-
-
Tito Dal Canton authored
-
- Sep 17, 2012
-
-
Kipp Cannon authored
-
Kipp Cannon authored
-
Kipp Cannon authored
-
Kipp Cannon authored
- this patch places some meta-programing magic in place to allow backwards compatibility with code that still uses the methods. such code should get a deprecation warning informing it to switch to direct attribute access - a customized version of the property() descriptor class is implemented that wraps each return value in a once-off subclass of its own type but having a .__call__() method that returns the object itself thereby emulating the original methods. it's a mess, but it's only temporary until all the code that uses the methods can be ported over. - the motivation for doing this is to make modifying the URLs of cache entries easier by making the scheme, host, and path writable. currently the only way to modify a url is to retrieve the components and reassemble a new URL from all components from scratch. it's silly
-
Kipp Cannon authored
-
Kipp Cannon authored
-
Kipp Cannon authored
and deprecate .to_segmentlistdict() method
-
Kipp Cannon authored
- the encoder in ifos_from_instrument_set() and the decoder in instrument_set_from_ifos() are not inverses of each other when dealing with a single instrument having more than two characters. because there is just a single instrument, the encoder's output would not normally contain a delimiter which causes the decoder to fall through to the "assume input contains concatenated two-character instrument names" case, and it decodes the name incorrectly. - this patch fixes the problem by appending an explicit "," character to the encoded representation when needed to force the decoder to use the comma-delimited decoding rules. - the decoder is modified to ignore the zero-length instrument name this encoding implies is present.
-
Tito Dal Canton authored
-
- Sep 14, 2012
-
-
Kipp Cannon authored
-
Kipp Cannon authored
-
Kipp Cannon authored
- the C version of segmentlist.__abs_() doesn't check if the list's length changes each time through the loop. because the arithmetic performed in the loop involves calls out to the interpreter, it's possible for specially-crafted code to alter the list's length as a side effect of the arithmetic. this patch checks the list's length on each iteration of the loop
-
Kipp Cannon authored
-
Ryan Fisher authored
-
- Sep 13, 2012
-
-
Kipp Cannon authored
-
- Sep 11, 2012
-
-
Kipp Cannon authored
- convert the .validchildren and .validattributes class attributes of all Element classes to frozensets to increase the speed of tests that check for invalid children and attributes
-
Kipp Cannon authored
-
- Sep 04, 2012
-
-
Kipp Cannon authored
-