On 10/14/11 10:57 AM, David-Sarah Hopwood wrote:
> There is custom magic in setup.py that creates bin/tahoe using a
> shebang for the interpreter that ran setup.py;
Right, bin/tahoe is the only entry point, so it's the only script that
needs this treatment.
> but the Makefile just does this:
>
On 14/10/11 18:38, Greg Troxel wrote:
> Sounds good. It should probably be using what autoconf would call
> @PYTHON@; on pkgsrc there is no 'python', because it would end up
> referring to something possibly different than you found at configure
> time. But bin/tahoe gets this right, so it seems
On 14.10.2011 19:22, Brian Warner wrote:
> On 10/14/11 8:04 AM, Greg Troxel wrote:
>
>> Thanks for 1) making the unpacked dir match the version in the file and
>
> Glad that helps.. it's part of the tarball-generation automation, so it
> should always work that way.
>
>> I am assuming while a ne
On 14/10/11 16:04, Greg Troxel wrote:
> I did get the following complaints when building:
>
> => Checking for non-existent script interpreters in tahoe-lafs-1.9.0b1
> WARNING: [check-interpreter.mk] The interpreter "/usr/bin/env" of
> "[...]/allmydata/test/check_grid.py" is not allowed.
> WARNIN
Correct. The dependency is actually going to be driven by Twisted. Tahoe
(both current 1.8.3 and upcoming 1.9.0) is happy with both current
Foolscap-0.6.1 and upcoming Foolscap-0.6.2 . Both current Foolscap-0.6.1
and upcoming Foolscap-0.6.2 are happy with current Twisted-11.0.0 . But
cur
On 10/14/11 8:04 AM, Greg Troxel wrote:
> Thanks for 1) making the unpacked dir match the version in the file and
Glad that helps.. it's part of the tarball-generation automation, so it
should always work that way.
> I am assuming while a new foolscap release is pending, 1.9.0 will not
> require
http://tahoe-lafs.org/source/tahoe-lafs/tarballs/allmydata-tahoe-1.9.0b1.tar.gz
Thanks for 1) making the unpacked dir match the version in the file and
2) having the path be close to the release path; both of these make
pre-testing as a pkgsrc entry easier and closer to what I'll commit when
t
We fixed the last major MDMF blocker tonight (#1561), so we're finally
clear to tag beta1. This means the web-API and CLI arguments are stable,
and we're comfortable with supporting them for a while.
The changes since alpha2 are pretty minor:
* CLI: use "tahoe put --format=MDMF" to create an MD