2008-02-03 Diab Jerius <djerius@cfa.harvard.edu>
Tag 0.06
2008-02-03 14:10 -0500 djerius <djerius@cfa.harvard.edu> (63d9df4e0bbe [tip])
* Makefile.PL:
fix up dependencies
2008-02-03 14:10 -0500 djerius <djerius@cfa.harvard.edu> (5ffd8de4e518)
* lib/App/Env.pm, lib/App/Env/Example.pm, t/cache.t:
fix multi-app environments
multi-app environments had a number of fundamental problems.
* the method for merging individual application's environments was
extremely simplistic and could result in parts of an app's
environment being reset.
* application environments were not necessarily based on the current
environment and when applied could reset the user's environment.
This problem was found for environments created by shell scripts
where the user's profile or .cshrc was sourced as part of the shell
invocation, in many cases resulting in a reset of the environment to
its inital state at login.
To avoid these problems, a new restriction was placed upon the
application specific modules. They must create their environments
based upon the current one, merging their changes into it. When
creating a merged environment, the application modules are
successively invoked on a copy of the current environment and the
final environment is used. Previously the modules' environments
were "merged" in a rather inflexible manner, resulting in
overwriting changes. This also means that 1) cached application
environments can't be used during a merger and 2) application
environments can't be cached during the course of a merger.
2008-02-03 13:52 -0500 djerius <djerius@cfa.harvard.edu> (97935eccd72e)
* t/App/Env/App1.pm, t/App/Env/App2.pm, t/App/Env/Site1/App1.pm,
t/App/Env/Site1/App2.pm, t/App/Env/Site2/App1.pm:
added debug statements
2008-01-17 15:59 -0500 djerius <djerius@cfa.harvard.edu> (255e3012e783)
* .hgtags:
Added tag 0.05 for changeset e312af21854f
2008-01-17 15:54 -0500 djerius <djerius@cfa.harvard.edu> (e312af21854f [0.05])
* bin/appexec:
fixed incorrect Getopt::Long configuration
2008-01-17 15:53 -0500 djerius <djerius@cfa.harvard.edu> (c45370921fbc)
* .hgtags:
Removed tag 0.05
2008-01-17 15:21 -0500 djerius <djerius@cfa.harvard.edu> (14f8a6a44461)
* .hgtags:
Added tag 0.05 for changeset 4c9ce9315ca8
2008-01-17 15:21 -0500 djerius <djerius@cfa.harvard.edu> (4c9ce9315ca8)
* ChangeLog, Changes, README, lib/App/Env.pm:
udpate for release
2008-01-17 15:14 -0500 djerius <djerius@cfa.harvard.edu> (649b4e3e59e6)
* bin/appexec, lib/App/Env.pm:
added appexec
2008-01-17 15:13 -0500 djerius <djerius@cfa.harvard.edu> (f10f36d537d0)
* MANIFEST, MANIFEST.SKIP, Makefile.PL:
administrative cleanup; use Module::Install
2008-01-17 14:16 -0500 djerius <djerius@cfa.harvard.edu> (8ab803f914ed)
* lib/App/Env.pm:
improve documentation
2008-01-17 14:09 -0500 djerius <djerius@cfa.harvard.edu> (6a62010f989c)
* lib/App/Env/Example.pm:
fix typo
2007-09-20 21:04 -0400 djerius <djerius@cfa.harvard.edu> (9d0a1aa99422)
* .hgtags:
Added tag 0.04 for changeset 60316599e900
2007-09-20 21:04 -0400 djerius <djerius@cfa.harvard.edu> (60316599e900 [0.04])
* ChangeLog, Changes, lib/App/Env.pm:
update for release
2007-08-17 17:01 -0400 djerius <djerius@cfa.harvard.edu> (06e0a82a4657)
* lib/App/Env.pm, lib/App/Env/Example.pm, t/App/Env/Site1/App1.pm:
The AppOpts hash is now passed as the sole parameter to the envs
function.
Clean up mess from previous work where the app env module's envs()
function took a more complicated input parameter. Now just gets the
AppOpts hash directly.
2007-08-17 14:39 -0400 djerius <djerius@cfa.harvard.edu> (d28d330df856)
* MANIFEST:
remove META.yml; seems to mess up make dist
2007-08-17 14:37 -0400 djerius <djerius@cfa.harvard.edu> (3f4fb7f7eab7)
* .hgtags:
Added tag 0.03 for changeset d36b41d3f2f0
2007-08-17 14:37 -0400 djerius <djerius@cfa.harvard.edu> (d36b41d3f2f0 [0.03])
* ChangeLog, Changes, lib/App/Env.pm:
update for release
2007-08-17 12:26 -0400 djerius <djerius@cfa.harvard.edu> (7a3552e41221)
* Makefile.PL, lib/App/Env.pm:
don't use Module::Load::Conditional. clean up envs invocation.
Originally Module::Load::Conditional was used to determine whether a
module existed separately from whether there was an error during its
compilation (Module::Load doesn't distinguish this case). However
Module::Load::Conditional's dependency chain is larger than I'd
like.
Instead, Module::Load was emulated using eval "require $module".
Same results, perhaps less efficient because it requires a run time
compilation.
calling the envs function in the application environment module is
no longer done with eval (once I figured out how to get the symbolic
reference to work).
2007-08-17 12:20 -0400 djerius <djerius@cfa.harvard.edu> (d3d9720ca944)
* t/badmodule.t:
add test for non-existent module
2007-08-16 17:11 -0400 djerius <djerius@cfa.harvard.edu> (d9b5f649ec19)
* .hgtags:
Added tag 0.02 for changeset 1dc9ba777697
2007-08-16 17:11 -0400 djerius <djerius@cfa.harvard.edu> (1dc9ba777697 [0.02])
* ChangeLog, Changes, MANIFEST, lib/App/Env.pm:
update for release
2007-08-16 17:04 -0400 djerius <djerius@cfa.harvard.edu> (45d4f9205efd)
* lib/App/Env.pm:
older perl's don't have weak references; code around that
2007-08-16 16:08 -0400 djerius <djerius@cfa.harvard.edu> (4f63370307b0)
* .hgtags:
Added tag 0.01 for changeset f113452d5ba2
2007-08-16 16:08 -0400 djerius <djerius@cfa.harvard.edu> (f113452d5ba2 [0.01])
* ChangeLog, Changes, LICENSE, MANIFEST, Makefile.PL, README,
lib/App/Env.pm, lib/App/Env/Example.pm, t/App-Env.t,
t/App/Env/App1.pm, t/App/Env/App2.pm, t/App/Env/Site.pm,
t/App/Env/Site1/App1.pm, t/App/Env/Site1/App2.pm,
t/App/Env/Site2/App1.pm, t/cache.t, t/env.t, t/envstr.t, t/site1.t,
t/site2.t, t/site3.t, t/system.t, t/use1.t, t/use2.t, t/use3.t:
initial version