[EMAIL PROTECTED] wrote:
> Author: autarch
> Date: Wed Jul  5 08:06:29 2006
> New Revision: 419247
> 
> URL: http://svn.apache.org/viewvc?rev=419247&view=rev
> Log:
> Cannot call sub to define a constant if the sub has not yet been
> defined, so just use a variable instead.

hmm, it worked ok for me (and has for quite a while)... but ok, whatever.

> 
> More refactoring to make flow of execution simpler (IMHO).

simpler, but not covering all the cases...

> 
> The test target works fine on my box without explicitly setting the
> apache binary or apxs binary. 

yes, because you have things like ~/.apache-test to tell it where they
are.  for a user that doesn't have that or TestConfigData.pm but has
Apache-Test installed you'll get this:

  $ perl Makefile.PL && make && make test
  Writing Makefile for Apache::SizeLimit
  ...

  We are now going to configure the Apache-Test framework.
  This configuration process needs to be done only once.
  ...

  Please provide a full path to 'httpd' executable:

   []

you won't have those files if you choose not to have them (by setting
$ENV{APACHE_TEST_NO_STICKY_PREFERENCES}) of if you've installed
Apache-Test from CPAN but haven't yet used it to run any tests.  or a
bunch of other reasons.

> It would be very bad form for a CPAN
> module to die in the Makefile.PL! That'd break any attempt to install
> it via a CPAN{PLUS} shell. I'll figure out how to make this optional
> in a future commit.

it was in the eval that defined HAS_APACHE_TEST, so no it wouldn't have.

I think that code needs to be re-added back in.  maybe not exactly the
way it was if you think it could be better, but the idea is so that no
combination of Apache-Test, it's configuration files, or lack thereof
causes a simple

  $ perl Makefile.PL && make && make test

to either fail or enter into an interactive process (which would also
cause the CPAN shell to timeout and not work).  I worked pretty hard to
cover all the cases, and I don't really code much without a real reason :)

--Geoff

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to