Re: [cc65] return vs. exit()

On 2009-10-20, at 10:15, Ullrich von Bassewitz wrote:

>> silverdr$ diff -qry cc65/libsrc/ cc65_r4372/libsrc/
>> [...]
>> Only in cc65/libsrc/cbm: systime.o
>> [...]
>
> Ok. At some point, a systime module has been in cbm/, but it is no  
> longer. So
> what happened is that you got a new version via svn, and this new  
> version
> changed the location of the systime module. Since there is no longer a
> systime.s file in cbm/, doing "make zap" won't remove the  
> corresponding object
> file. So you have the old object file around, which may cause  
> problems.
> [...]
>
> "make zap" does only remove files it knows about. I've lost data  
> several times
> because people wrote software that removed anything it didn't know  
> about ("it
> doesn't belong to me, so delete it"), so the library makefile will  
> not do
> that.

I see your point. On the other hand the case here is that it doesn't  
clean what it doesn't know a thing about but still takes it into  
account when building.

As you noticed, I don't seem to have done something utterly stupid to  
get into this potentially problematic situation. Having that in mind I  
believe it is quite possible that similar situations may happen also  
in the future. Here there was a new (also to me) warning that  
triggered the investigation but similar problems may go unnoticed,  
while causing lots of bad blood later on.

/me would thinks that in this situation it would be more safe and  
consistent if things, which are not cleaned/zapped are neither taken  
into account when building or so.

Regards,

-- 
SD!

----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo@musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.