cvs commit: ports/emulators/linux_base-gentoo-stage1 Makefile

[Trevor Johnson]

David O'Brien wrote:

> On Fri, Dec 31, 2004 at 02:52:40PM -0500, Trevor Johnson wrote:

> > It wasn't missed, but intentional.  Those files are not precious--they
> > weren't compiled (by the port), just untarred.  Unlike most ports, this
> > port doesn't do anything with them except make the packing list.  In my
> > judgment, it's better to remove them afterward, so someone who simply does
> > "make install"  will need half as much disk space.
>
> ?? That is true of 100 other ports -- they extract binary files.

There are only a few that do this purely to get a listing of the files.
Previously, this port just did a "tar tyf" to generate the list of the
files, and you didn't propose that it should actually untar the files,
just in case the user wanted to look at one.  Why has it become important
now?

> what about the 10,000 ports -- boy do they use a lot of disk space after
> 'make install' and before 'make clean'.  What should we do about them??

You seem confused--the files we're discussing aren't needed for "make
install" to proceed.  As I said, all the port does with them is get a
listing of the filenames so it can make the packing list.  During the
installation, the port extracts directly from the distfile, a second time.
Other ports typically extract from the distfile, sometimes run the
configure script therein, compile stuff by invoking a makefile from
upstream, and then invoke that makefile again when they install.  Because
it's only aspiring to stage 1, this port does not do any of that.  If we
were talking about stage 2 or 3, yes, there would be precious files that
should only be cleaned up at the user's explicit request.
-- 
Trevor Johnson