Re: kbtin (new)
On Tue, Jun 20, 2006 at 08:55:40PM -0400, Joe Smith wrote:
> "Adam Borowski" <kilobyte@angband.pl> wrote in message
> [🔎] 20060620003809.GA10105@angband.pl">news:[🔎] 20060620003809.GA10105@angband.pl...
>
> >The ITP is #213361. The package is foo-clean; the packaging is
> >trivial (clean autotoolage).
> What is this foo package you keep talking about?
> I know of lintian, linda, puiparts, but not foo.
And pbuilder.
What I mean, the set of automated tools to run is common to every
package tested at a given time. Rare or specialized tools do exist,
but I haven't heard of them being used around d-m these days.
At this time, the fashion^Wcommon practice is to run:
* lintian
* linda
* pbuilder
* piuparts
and there is hardly ever any reason to skip any of them, so it can be
generally assumed that if a metasyntactic variable is used, it's used
as an abbreviation for all four.
This said, piuparts is pretty much a waste of CPU time for any
package that doesn't involve daemons, system-wide config or
system-wide cache -- ie, a majority of user-level software; however,
CPU time is so cheap these days that there is no excuse to skip
piuparts for us pre-DDs.
Going back to kbtin, I'm not sure whether I can claim linda-cleaness
if there's a change made solely to silence her. Linda considers
_empty_ config.{sub,guess} to be outdated; I thus put some short
placeholder contents just to make her happy. The empty
config.{sub,guess} files themselves were a hack to work around a bug
in an old version of automake, a bug that is long gone now. The
files of course should be removed from the upstream tarball, but 1.
it appears to be impossible to remove files from the source; and 2.
it makes no sense to repack the upstream tarball just to get rid of
an empty harmless file.
Of course, if you're a sponsor, forget you've ever seen the previous
paragraph :p The package is linda-clean, neither the hacks nor the
reasons for them are present in upstream svn any longer, and this is
everything you need to know :p
Cheers, regards and what not,
--
1KB // Microsoft corollary to Hanlon's razor:
// Never attribute to stupidity what can be
// adequately explained by malice.
Reply to: