I can suggest another (much less) patch, which still allows users to
play hack without path modification. But all this stuff with checking
saves creating time doesn't looks actual for me, so I keep original
patch below.
Index: hack.unix.c
On 2023/05/28 18:09:00 +0200, Marc Espie wrote:
> Here's a slightly more specific diff avoiding useless stat(2)
looks fine to me, and in ports it produces a nice error message:
% chmod 600 Makefile
% make
[...]
make: don't know how to make do-extract
Stop
Turns out that, due to the search rules we use, make
is mostly silent in case it couldn't read a Makefile
The following patch lets it track the makefilenames that
do exist, but that it wasn't able to open, so that
a post-mortem can include these.
Not sure if other use-cases could also use the
On 2023/05/28 13:16:34 +0200, Marc Espie wrote:
> Turns out that, due to the search rules we use, make
> is mostly silent in case it couldn't read a Makefile
>
> The following patch lets it track the makefilenames that
> do exist, but that it wasn't able to open, so that
> a post-mortem can
On Sun, May 28, 2023 at 04:44:37PM +0200, Omar Polo wrote:
> On 2023/05/28 13:16:34 +0200, Marc Espie wrote:
> > Turns out that, due to the search rules we use, make
> > is mostly silent in case it couldn't read a Makefile
> >
> > The following patch lets it track the makefilenames that
> > do
Here's a slightly more specific diff avoiding useless stat(2)
Index: engine.c
===
RCS file: /cvs/src/usr.bin/make/engine.c,v
retrieving revision 1.70
diff -u -p -r1.70 engine.c
--- engine.c25 Oct 2021 19:54:29 - 1.70
+++
