pkgsrc-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: pkgsrc build looking for PYPACKAGE suggestion?

On 1/13/21 10:02 AM, Greg Troxel wrote:
Bob McGowan <> writes:

OK.  The question as to if I had done a 'make clean clean-depends' is
answered with 'yes'.  However, note that this whole experience started
with trying to do cross-compiles on my Linux x86_64 system.  Once I
got qemu working and NetBSD installed, rather than download all the
files again, I copied them from the Linux system to NetBSD, which
forced the make clean; make clean-depends.

Given that the potential for cross contamination may still be there I
will clean up by deleting and downloading a complete new set of files.
You can, but really the important thing is to rm -rf */*/work, and then
do an update and make sure nothing is modified.   Redoing it amounts to
the same thing and has less opportunity for error.

By the time I saw this I had already done this:

    mv pkgsrc pkgsrc.old

    tar xzf pkgsrc.tar.gz

    cd pkgsrc/net/samba4

    make package 2>&1 | tee 2020-01-11_samba4_make_package.txt

I'm not sure how far it ran, with respect the two libraries that were a problem before, but I can say the library names do not appear in the directory .../pkgsrc/packages/All so I assume they have not be reached.

What happened this time was this error:

    waf: error: no such option: --disable-pie

As for the discussion of variables, and which should or should not be
changed by the user, I was mis-interppreting the following:

     The following variables will affect the build process of this package,
     gnome-doc-utils-0.20.10nb7.  Their current value is shown below:

             * PYTHON_VERSION_DEFAULT = 38

     Based on these variables, the following variables have been set:

             * PYPACKAGE = python27

     You may want to abort the process now with CTRL-C and change their
     before continuing.  Be sure to run `/usr/bin/make clean' after
     the changes.

The last sentence says "...change _their_ value...".  But looking
closer and at other cases of similar messages, where multiple
variables are mentioned, so I suspect the plural refers to the
variables in the upper section of the message only.
This text is misleading.  What's going on is that there are variables
set from various places (user, package, infrastructure) and via logic
that is sometimes complicated lead to variables that control things.

The text should probably make clear that only the first group of
variables is user-settable, and the second group should not be set

Don't feel bad for making  mistake here - this is very complicated and
the message you read was misleading.   I have just edited
mk/misc/ to refrain from suggesting that variables in the second
group be set.

Thanks for looking at this and for the encouragement. ;)

I suspect my wife would prefer I give up on this, but that is not going to happen.


Home | Main Index | Thread Index | Old Index