<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
--></style>
</head>
<body class='hmmessage'>
I kind of think the original authors had in mind:<BR>
if SYSTEM_LIBS contains "X11"<BR>
<BR>
<BR>
as a useful construct..but I'm not sure..and no matter their intentions,<BR>
it seems at first promising but then actually inadequate, and that maybe<BR>
we need something to replace what it seems to promise.<BR>
<BR>
<BR>
IN PARTICULAR:<BR>
<BR>
If something is directly dependent on X11, then it seems reasonable<BR>
to do m3makefile:<BR>
<BR>
if SYSTEM_LIBS contains "X11"<BR>
<BR>
..everything..<BR>
<BR>
library("foo")<BR>
<BR>
end<BR>
<BR>
but then it seems to me that clients of library("foo") shouldn't<BR>
know or care that foo depends on X11.<BR>
<BR>
It seems like you maybe should be able to do like:<BR>
<BR>
<BR>
if SYSTEM_LIBS contains "X11"<BR>
<BR>
..everything..<BR>
<BR>
library("foo")<BR>
<BR>
else<BR>
<BR>
no_library("foo") % or something<BR>
<BR>
end<BR>
<BR>
<BR>
and then when I have:<BR>
<BR>
import("foo")<BR>
library("fooclient")<BR>
<BR>
<BR>
I should probably just a warning like:<BR>
<BR>
"not building fooclient because foo is explicitly not available"<BR>
<BR>
<BR>
"explicitly not available" as opposed to "seems to just be missing,<BR>
perhaps you forgot to build it first?"<BR>
<BR>
<BR>
In particular, "no_library" would need to be "active" and leave a mark of some sort.<BR>
<BR>
<BR>
?<BR>
<BR>
<BR>
In the meantime I'm considering some special case thing, like<BR>
configfile:<BR>
HAS_GUI = TRUE<BR>
<BR>
or<BR>
HAS_TRESTLE = TRUE<BR>
<BR>
or config.common:<BR>
if not defined("HAS_TRESTLE")<BR> HAS_TRESTLE = TRUE<BR>
end<BR>
<BR>
??<BR>
<BR>
<BR>
- Jay<BR> </body>
</html>