From: Victor A. Wagner Jr. (vawjr_at_[hidden])
Date: 2004-03-05 13:51:59
At Friday 2004-03-05 00:20, you wrote:
>Victor A. Wagner Jr. wrote:
> > >So, bjam/Boost.Build would have to:
> > >1. Detect that bjam need to be rebuild, which is not hard
> > >2. Rebuild bjam
> > >3. Restart the build using the new bjam. This seems hard.
> > maybe it's hard, but gnu make managed it back in the mid '80s
> > that is, if make got re-built it managed to restart w/ the new copy.
>I was not aware of this; the only interesting behaviour of make I know was
>restarting when Makefile was changed.
>Do you know that they managed restarting with new copy? But anyway, I think
>it's easier to add explicit rebuild commands to whatever script you have.
>After all, full jam rebuild is somewhere between fast and very fast for me,
>and should be even faster on msvc.
hmmm, maybe I modified our make to have "make.exe" an intrinsic dependency
for all rules...
the speed isn't the point. I think the point is, if you have a "build
system" then one "should" (IMO) be able to use it to build anything.
30 years ago I managed to insist that ALL projects could be built by the
simple process of setting the current directory to the root of the source
tree and entering
as the sole command (this took a lot of work to create the "build.bat"
files by the programmers) Since then different build systems have come
(and gone) tho mostly variants of "make" (gnumake, nmake, polymake, etc)
and building the scripts has become "different"
this predated make by few years (certainly in the places we were). It's
distressing to think we've regressed to the point that I cannot simply go
to the "root" directory and type a single command.
>Yahoo! Groups Links
Victor A. Wagner Jr. http://rudbek.com
The five most dangerous words in the English language:
"There oughta be a law"
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk