lp:~applet199/fluidity/python_state_extension

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

3510. By Lawrence Mitchell on 2011-06-29

Don't use hashlib module in make_{check_options,register_diagnostics}

The hashlib module only appeared in Python 2.5, superseding the now
deprecated sha module. Rather than requiring a new Python, just use
sha directly (the results are the same).

3509. By Lawrence Mitchell on 2011-06-29

Reduce spurious recompilations in clean or mostly-clean trees

Avoid touching lib/libfluidity.a unless actually necessary. This comes
in a few parts:

o Use bzr revision-info to set up Fluidity's version number (and only
update confdefs.h if the version changed)

o Calculate hashes of the check_options and register_diagnostics files
and only update them if the hash changes

o Avoid depending on .PHONY targets that just point directly at actual
targets

3508. By Cian Wilson on 2011-06-28

Trivial fix for not really a bug (unless we start requiring that special tests run).

3507. By Cian Wilson on 2011-06-28

Unperiodising a periodic mesh in parallel may produce floating nodes that are unattached to any element on that process. This leads to linked lists with zero length in make_sparsity (for the unperiodised mesh) resulting in a segfault when requesting the maxval. This commit prevents this from happening by not checking for the possible number of columns associated with floating nodes - i.e. that have a 0 length linked list.

3506. By Jon Saunders on 2011-06-28

Removing one of the two readme files in the legacy_reservoir_prototype dir to avoid confusion in non-case-sensitive environments

3505. By Tim Bond on 2011-06-27

Commit in response to bug 802426:

https://bugs.launchpad.net/fluidity/+bug/802426

This only affects the build process and is thus orthogonal to the code base.

Summary: On RHEL 5.6 with GCC 4.4 a requirement for -llapack to occur after
   -lmba2d was noted, causing a build failure. This is remedied by this commit
   by swapping the order of '$LIBS -lmba2d' in configure.in.

3504. By Cian Wilson on 2011-06-26

Bug fix for the bug fix in r3503. Thanks to Stephan for spotting this. Shouldn't affect the tests that use this as they already have dimensions of order 1.

3503. By Cian Wilson on 2011-06-22

In response to Tim's request... the test is failing on buildbot 32 bit because of precision issues in numpy lexsort. Fixing this by rounding things less than (in an absolute sense) machine precision to zero before calling lexsort in VtuMatchLocationsArbitrary. I don't have a 32 bit machine to test this on so will keep an eye on buildbot. All tests that use this function pass locally.

3502. By Simon Funke on 2011-06-22

make plot script consistent what is used for the wetting and drying paper

3501. By Patrick Farrell on 2011-06-21

An interface from libadjoint (adj_create_variable) has changed to be
more Fortran-ish. This change updates the adjoint code to match
that change in the interface.

If this doesn't compile on buildbot, just wait a day until the new
libadjoint has been built and pushed out to the builder.