Re: On the dangers of PKGNAME=.

On Thursday 23 October 2003 23:10, you wrote:
> Adriaan de Groot wrote:
]> > ./share/apps/kpilot/sysinfoconduit/Template.html
> > ./share/apps/kpilot/sysinfoconduit/Template.txt
> > ./share/services/sysinfo-conduit.desktop
> >
> > (and several more).
>
> In this situation I would set $PKGNAME to "apps/kpilot" and add the
> files under it to $INSTALL_DATA.  For the other files under "share" add
> a new install target "install-local".  That's what it is for.

Heh, given that more things get installed into "services" than into 
"apps/kpilot" by this particular recipe, then I'd set PKGNAME to "sevices" -- 
which really makes no sense. It gets worse when stuff also goes into "icons" 
and "wallpapers" and "doc". Thing is, it could work so simply, if only AAP 
would let PKGNAME be "" (or ".").

> A compromise could be to make a difference between an empty $PKGNAME and
> not setting $PKGNAME.  It's subtle, but it makes sure that the user sets
> $PKGNAME, so that he (hopefully) knows what he is doing.

Well, PKGNAME= vs. PKGNAME=. is hardly more visible, both are subtle. I'm 
inclined to say that PKGNAME=. is better because no AAP code needs to change, 
just Python bugs fixed.

> I consider this a bug in os.makedirs().  Would you want to post a bug
> report for this or should I do it?

You do it please.

> The problem is that the recursive implementation of makedirs() only
> checks for existence of the dir before going recursive.  It doesn't
> check again after the recursive invocation.  It should either do another
> exists() or skip "." directories.

That would also make it susceptible to race-conditions causing makedirs() to 
throw exceptions weirdly.



-------------------------------------------------------
This SF.net email is sponsored by: The SF.net Donation Program.
Do you like what SourceForge.net is doing for the Open
Source Community?  Make a contribution, and help us add new
features and functionality. Click here: http://sourceforge.net/donate/