[PATCHv2] remote-helpers: build in platform independent directory

[PATCHv2] remote-helpers: build in platform independent directory

[Michael J Gruber]

The build directory which is used by distutils depends on the platform
(e.g. build/lib on Fedora 13, build/lib.linux-i686-2.6 on Ubuntu 9.04).
But test-lib.sh expects to find the build in build/lib which can cause
t5800-remote-helpers.sh to fail early.

Override distutils' choice so that the build is always in build/lib.

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
---
Sorry for the v2 so soon. With the previous version, the "make install" step
would still possibly create a platform dependent build dir. This works but
creates two lib dirs within build, which is unnecessary and confusing.
The solution with setup.cfg avoids this and is cleaner anyways.

 git_remote_helpers/setup.cfg |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)
 create mode 100644 git_remote_helpers/setup.cfg

diff --git a/git_remote_helpers/setup.cfg b/git_remote_helpers/setup.cfg
new file mode 100644
index 0000000..4bff887
--- /dev/null
+++ b/git_remote_helpers/setup.cfg
@@ -0,0 +1,3 @@
+[build]
+build_purelib = build/lib
+build_platlib = build/lib
-- 
1.7.3.rc2.221.gbf93f

Re: [PATCHv2] remote-helpers: build in platform independent directory

[Sverre Rabbelier]

Heya,

[+Jonathan again]

On Fri, Sep 17, 2010 at 23:00, Michael J Gruber
<git@drmicha.warpmail.net> wrote:
> The solution with setup.cfg avoids this and is cleaner anyways.

This does feel a lot cleaner, I like this approach better than the v1
one. I suspect Junio's question still stands though:

On Fri, Sep 17, 2010 at 20:41, Junio C Hamano <gitster@pobox.com> wrote:
> Hmm, wouldn't this interfere with the install target if you do not tell
> the "setup.py install" where your built stuff lives?

-- 
Cheers,

Sverre Rabbelier

Re: [PATCHv2] remote-helpers: build in platform independent directory

[Michael J Gruber]

Sverre Rabbelier venit, vidit, dixit 17.09.2010 23:07:
> Heya,
> 
> [+Jonathan again]
> 
> On Fri, Sep 17, 2010 at 23:00, Michael J Gruber
> <git@drmicha.warpmail.net> wrote:
>> The solution with setup.cfg avoids this and is cleaner anyways.
> 
> This does feel a lot cleaner, I like this approach better than the v1
> one. I suspect Junio's question still stands though:
> 
> On Fri, Sep 17, 2010 at 20:41, Junio C Hamano <gitster@pobox.com> wrote:
>> Hmm, wouldn't this interfere with the install target if you do not tell
>> the "setup.py install" where your built stuff lives?
> 

No, it addresses that also. Maybe my note after the boiler plate was not
clear enough, but the "build" setting in setup.cfg also governs the
build phase of "setup.py install". Install would work with v1 also but
would possibly create another dir during the build phase, with v2 both
build phases (that of the build subcommand as well as that of the
install subcommand) use the same setting.

Michael