[Linux-ha-dev] Re: [Linux-ha-cvs] Linux-HA CVS: heartbeat by msoffen from

[Alan Robertson]

David Lee wrote:
> On Wed, 22 Sep 2004, Alan Robertson wrote:
> 
> 
>>linux-ha-cvs at lists.linux-ha.org wrote:
>>
>>>linux-ha CVS committal
>>>
>>>Author  : msoffen
>>>Host    :
>>>Project : linux-ha
>>>Module  : heartbeat
>>>
>>>Dir     : linux-ha/heartbeat/lib
>>>
>>>
>>>Modified Files:
>>>      Tag: STABLE_1_2
>>>	BasicSanityCheck.in TestHeartbeatComm.in
>>>
>>
>>
>>
>>This is at best a silly "patch".  At worst it is a horror.
> 
> 
> To clarify: the patch is, at source, mine: any criticism to me, not Matt.
> 
> Its intention is get things consistent with what is already in place.  In
> that sense it is an improvement (surely?).  At present, heartbeat is
> internally inconsistent and things like "BasicSanityCheck" weren't even
> getting started.
> 
> It is similar to part of a pkg/Makefile.am (1.5->1.6) patch you applied
> for me, which was a huge help (thanks!) in getting Solaris installation
> working cleanly in 1.2.3 and HEAD.
> 
> But you then raise a very valid point:
> 
> 
>>There are literally dozens of such scripts all over everywhere.  And, the
>>impact of renaming them will extend into the heart of heartbeat.
>>
>>There is to my knowledge no reason to do this anywhere at all - the sole
>>exception being *maybe* init scripts on *BSD OSes - for reasons of convention.
>>
>>No UNIX system ever existing required shell scripts to end in .sh.
> 
> 
> So what we (Matt, I, etc.) need from you is guidance on how to achieve the
> internal consistency which is currently lacking.  Either we continue with
> the INIT_EXT (which seems to have been established for some time), or the
> executive decision needs to be made to remove all trace of INIT_EXT.

> 
> Personally, I don't mind which of those two options we take.  Indeed, my
> vote would be, for the case of Solaris, for INIT_EXT to be null (or its
> equivalent: the removal option).
> 
> The intention of the patch was good, and was an improvement towards
> consistency, on the small scale, at least.  But, in the bigger picture,
> your suggested option of removal of INIT_EXT also looks fine, with the
> added virtue of simplicity, if we are reasonably happy that no likely OSes
> will need it.


The inconsistency is more in the OS than in our code.  We're trying to do 
the minimum amount to meet the OS's weirdness.  Only one script is an init 
script, and IMHO it's the only one that needs this extension.

> If you wish to remove all trace of INIT_EXT, then someone (who?) with CVS
> access needs to do it.

I don't mean that it should be removed where it's needed, but it shouldn't 
be added for things that aren't init scripts.  Since we only have one such 
init script, this seems like it should just be one script that this applies 
to.  Better comments on what it is and why it's there wouldn't hurt anything.

BasicSanityCheck is just a test program - not executed by init, but by a 
human being at the time they test a new version.  So, it doesn't follow 
under the "init script" category.


-- 
     Alan Robertson <alanr at unix.sh>

"Openness is the foundation and preservative of friendship...  Let me claim 
from you at all times your undisguised opinions." - William Wilberforce