Re: Automatic prerequisite generation

[Mike Capp]

Greg Chicares wrote:
> Mike Capp wrote:
>
> > [snip]
> > the MSYS shell seems too restrictive with regard to where you put your
> > files, running make from editors and capturing the output, and so on.
>
> You can do
>   make 2>&1 |less
>   make some_target >/tmp/captured-output 2>&1
>   make 2>&1 | tee captured-output
> What do you want instead?

As you doubtless guessed, I think my objections largely stemmed from 
ignorance here. I'd heard that MSYS was purely an interactive shell, 
based on the assumption that you'd always be running 'make' from the 
supplied rxvt terminal, and the docs that came with it mostly reinforce 
that impression. This would make it pretty useless as an external tool 
for an editor. If, as seems to be the case, you can also run the sh.exe 
directly or specify it as the system shell inside your makefile, MSYS 
gets a *lot* more attractive.

> And read this:
>   http://make.paulandlesley.org/autodep.html

Have done, and am even beginning to understand it ;-) Seriously, the 
paper's great, it's just the sed commands that scare the bejesus out of me.

Looking at 
http://gcc.gnu.org/onlinedocs/gcc-3.2.1/gcc/Preprocessor-Options.html , 
several of the options look as if they're providing built-in support for 
 the advanced idiom Paul describes. In particular...

[begin excerpt]
-MP
    This option instructs CPP to add a phony target for each dependency 
other than the main file, causing each to depend on nothing. These dummy 
rules work around errors make gives if you remove header files without 
updating the Makefile to match.

    This is typical output:

test.o: test.c test.h

test.h:
[end excerpt]

...which looks very much like Paul's "So, all we need to do is 
post-process the original dependency file and turn all the prerequisites 
into targets with no commands or prerequisites." Anyone have experience 
with this switch?

 >>together with a .DELETE_ON_ERROR target to handle cleanup if something
> > goes pear-shaped.
>
> Even with that directive, I think it may
> still be possible to get incomplete .d
> files, if the first command works but the
> second fails. I've seen weird things like
> that happen.

Hmm. I tried forcing various errors in the second command, and MAKE 
cleaned up properly every time, but I'll believe anything where 
COMMAND.COM is concerned...

> Here's a solution I developed when I was still
> using ms shells:
>   http://mail.gnu.org/archive/html/make-w32/2002-07/msg00008.html
> This approach requires sed, but that message
> tells how to make the perfect sed, using MSYS.

To clarify a couple of things ("|>" quoted lines are from your 
referenced post):

|> Use it to build this sed from source:
|>  ftp://alpha.gnu.org/pub/gnu/sed/sed-3.02.80.tar.gz

Is rolling your own sed still required? Even if you're using bash? The 
current MSYS already ships with GNU sed version 3.02, and I'd worry 
about a makefile that requires one particular build of one particular 
sed version to work properly.

|> # Changed 2001-12-03 by GWC: replaced single quote with double
|> # quote throughout. Required for win2k.

The sed command in your post uses single quotes throughout. Did you back 
the Windows-compatibility changes out when you switched to zsh? The 
comments look to predate your post by a fair ways.

> > However, I'm still a 'make' newbie,
> > and suspect I'm missing something here. Are there problems with this
> > approach I'm not seeing?
>
> When do dependencies get built? If you have
> to decide when to rebuild them, then you can
> forget. And when you do decide, do they all
> get rebuilt? These are the issues Paul's
> white pater deals with.

No, I was having the dependencies get built automatically using the 
remaking-makefiles feature. The GNU manual's recommendation is pretty 
much identical to Paul's paper up until the advanced section.


Many thanks for the tips and pointers; should keep me busy for a while :-)