[OE-core] FETCHCMD versus FETCHCOMMAND

[Richard Purdie]

On Sat, 2012-07-07 at 06:33 -0400, Robert P. J. Day wrote:
> just noticed the following as i was still poking around with
> fetchers:
> 
> $ grep -r FETCHCMD * | grep -v pyc
> bitbake/lib/bb/fetch2/hg.py:        basecmd = data.expand('${FETCHCMD_hg}', d)
> bitbake/lib/bb/fetch2/svn.py:        ud.basecmd = d.getVar('FETCHCMD_svn', True)
> bitbake/lib/bb/fetch2/osc.py:        basecmd = data.expand('${FETCHCMD_osc}', d)
> bitbake/lib/bb/fetch2/git.py:        ud.basecmd = data.getVar("FETCHCMD_git", d, True) or "git"
> bitbake/lib/bb/fetch2/git.py:        basecmd = data.getVar("FETCHCMD_git", d, True) or "git"
> bitbake/lib/bb/fetch2/git.py:        basecmd = data.getVar("FETCHCMD_git", d, True) or "git"
> bitbake/lib/bb/fetch2/bzr.py:        basecmd = data.expand('${FETCHCMD_bzr}', d)
> bitbake/ChangeLog:	  (requires new FETCHCMD_svn definition in bitbake.conf)
> meta/conf/bitbake.conf:FETCHCMD_svn = "/usr/bin/env svn"
> meta/conf/bitbake.conf:FETCHCMD_cvs = "/usr/bin/env cvs"
> meta/conf/bitbake.conf:FETCHCMD_wget = "/usr/bin/env wget -t 2 -T 30 -nv --no-check-certificate"
> meta/conf/bitbake.conf:FETCHCMD_bzr = "/usr/bin/env bzr"
> meta/conf/bitbake.conf:FETCHCMD_hg = "/usr/bin/env hg"
> $
> 
>   i can see that FETCHCMD represent the "base" fetch command for each
> fetcher, i also notice that there isn't a perfect correspondence
> between the fetchers and bitbake.conf.
> 
>   for example, there's no definition of FETCHCMD_osc even though
> osc.py checks for it.  conversely, bitbake.conf sets FETCHCMD_cvs even
> though the cvs.py fetcher never bothers to check it.
> 
>   nothing wrong with any of that, just an observation about
> consistency.

Thanks for the reminder. The idea was to drop the COMMAND versions in
favour of the CMD ones. I've just sent out a series of patches which
would complete that work.

We're probably heading to a point where the fetchers have sane defaults
encoded and we remove the bitbake.conf settings entirely for the default
case just allowing the user to override if ever needed.

Cheers,

Richard