[Debburn-devel] Replacement of "cdrecord"
Christian Fromme
kaner at strace.org
Sat Sep 23 18:38:37 UTC 2006
On 23.09. 13:14, Albert Cahalan wrote:
> On 9/23/06, Christian Fromme <kaner at strace.org> wrote:
> >On 23.09. 16:40, Kai Wb. wrote:
> >>
> >> I've read the here the mailinglist and the TODO file in the repository.
> >> One point is to change the name of the directory "cdrecord" to "wodim".
> >> I'm currently working on that and have some questions:
> >> 1. Is anybody else doing this (then we should merge our work or one
> >> should stop (to save time).
> >
> >No, not that I know of.
>
> I think that the name change was a serious error.
>
> Suppose I forked the "ps" command. Well I did in fact,
> almost a decade ago. Suppose I changed the name of
> the command to "rypa". Would that break any scripts
> you might have? Might it prevent your system from
> shutting down cleanly? I hope you see the problem.
> The cdrecord wrappers are numerous and fragile, but
> the even bigger problem is scripts in people's home
> directories and build systems.
The best time for a name change is *now*, directly after the fork. It is
not totally necessary, you're right. I would also have been open to
discuss this, but the name change is done now and we should go ahead and
change all occurences of 'cdrecord' to 'wodim'. Change it back is not an
option from my pov.
Will it break scripts? Maybe[0]. But nothing a symlink couldn't fix I
think.
> >> 2. When I'm up to renaming a directory and fixing the includes I would
> >> suggest not only to change the name of the directory but also change the
> >> name of the files cdrecord.[h|c]. I think it would make sense. Or is
> >> there anything that stands against it? (apart from the work).
> >
> >Go ahead! :) But please use the cleanup/ branch for this. Thanks
>
> So that becomes the main branch. The normal default
> becomes dead and useless with time, but you can't
> just merge back because there will never be a good
> point in time when you can say that NOBODY will
> ever again want to patch the non-cleanup code.
We discussed this already and IIRC the outcome of the discussion was to
leave the main branch as it is [except bugfixes] until post-etch. Again,
there is no general rule why this must be done that way and could also
be done different if everybody [especially people with commit-access ;)]
agrees.
Best wishes,
Christian
[0] http://lkml.org/lkml/2004/8/14/44
More information about the Debburn-devel
mailing list