Re: commit 8a796f8f is breaking a usecase

"Pierre Labastie" ([email protected] via alfs-discuss Mailing List) <[email protected]> Mon, 15 Dec 2025 14:01:41 +0100
Newsgroups gmane.linux.lfs.automated
Message-ID <[email protected]>
On Fri, 2025-12-12 at 07:47 +0800, Zhang Wen wrote:

> 
> After building the initial system, I rerun jhalfs on the host, then
> select BLFS
> book and the custom tools support, but it doesn't work. I also tried
> the
> "rebuild files" config but nothing changed.
> 
I think I understand what you do, but this has never been supported. It
became possible as a side effect of commit 200cc0151e7b from September
12, 2023, and the commit you cite in the subject reverted this side
effect.

The reason is the following: the main Makefile always creates a
BLFS_TOOL target, whether or not the BLFS tools are built. If they are
not built, the only effect of this target is to create a time-stamp
file in $(MOUNT_PT)/jhalfs, so that this target is considered
fulfilled, whenever you rerun the Makefile. The only reason why this
target was rerun in your use case is because it depended on a PHONY
target, which was always run, so that BLFS_TOOL was run also.

Now, if you insist on keeping your use case (I am not sure why you do
not want to build the BLFS tools during the first run, maybe you can
elaborate on this), I think it is better to not generate the non useful
targets if they are not needed. I'll look at this.

Regards
Pierre

-- 
http://lists.linuxfromscratch.org/sympa/info/alfs-discuss
Unsubscribe: See the above information page