[SM-Discuss] The Great E-mails: part 3 (DEPENDS)
Eric Sandall
eric at sandall.us
Fri Aug 28 23:47:15 EDT 2009
On Sat, 29 Aug 2009 03:49:22 +0400
Vlad Glagolev <stealth at sourcemage.org> wrote:
> We have a lot of annoying things here.
>
> First of all: spaces.
>
> Things like:
>
> --
> optional_depends very-long-spell-name-or-not \
> "--enable-foo" \
> "--disable-foo" \
> "for foo feature" &&
> --
>
> are damn useless, and it's hard to fix it every time, when some more
> dependencies were added, or description of the dependency, and so on.
> It's very hard/time-consuming to go through the jungles of spaces
> while editing the spell's deps or add new ones.
>
> I'd offer to just remove (and don't add) unneeded spaces here:
>
> -
> optional_depends spell \
> "--with-feature" \
> "--without-feature" \
> "for some kind of feature support" &&
I would agree with this, but would like to continue to use two spaces
between each word/operator. Also note that "spell" should be quoted,
according to our current standard, though
http://www.sourcemage.org/DEPENDS doesn't seem to be consistent about
this and mostly (but not always) only does so for optional_depends.
e.g.
optional_depends "spell" \
"--with-feature" \
"--without-feature" \
"for ultimate access to everything" &&
> --
>
> Next, newlines.
>
> --
> depends xproto &&
> depends fontconfig &&
> depends libx11
> --
>
> is ok, but I think, depends block should be separated from
> optional_depends block by new line, and all optional_depends blocks
> should be separated by new line from each other.
I agree, and I thought this was how it was documented to being done,
but looking at http://www.sourcemage.org/DEPENDS shows otherwise. I
believe the current wiki may have mangled some of the formatting (since
the "\" and "&&" no longer line-up, but have lots of extra spacing).
>
> Also:
>
> --
> optional_depends spell \
> "--with-foo" \
> "--without-foo" \
> "for foo support" &&
> --
>
> is ok, like this:
>
> --
> optional_depends spell \
> "--with-foo" "-without-foo" \
> "for foo support" &&
I disagree, except in the case where there are *no* --with/--without
flags. Keeping them separate makes it clearer which is the enable and
which the disable, IMO.
> --
>
> and like this:
>
> --
> optional_depends spell "" "" "some pythonic spell with no actual
> configure flags" --
I belive this should be okay for the non-flagged optional dependencies,
so long as the entirety is <80 columns.
-sandalle
--
Eric Sandall | Source Mage GNU/Linux Developer
eric at sandall.us PGP: 0xA8EFDD61 | http://www.sourcemage.org/
http://eric.sandall.us/ | http://counter.li.org/ #196285
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
Url : http://lists.ibiblio.org/pipermail/sm-discuss/attachments/20090828/29b279c4/attachment.bin
More information about the SM-Discuss
mailing list