tech-pkg archive

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

Re: and do-*

Manuel Bouyer <> writes:

> On Mon, Jul 04, 2022 at 05:32:33PM +0100, Jonathan Perkin wrote:
>> * On 2022-07-04 at 17:29 BST, Manuel Bouyer wrote:
>> > so, since xentools415 has been converted to python/ (from
>> > we get a warning about do-build and do-install being redefined.
>> Generally we handle this by adding .if !target(do-build) style wrappers
>> around the definitions in  Is there any reason why we're not
>> doing that in this case?
> These are already conditional on PYDISTUTILSPKG, I'm just extending this
> logic. Note that PYDISTUTILSPKG also conditionally defines more stuff than
> the do-* targets; I don't know if they can cause problem for packages
> that don't want the build logic from

Starting out with the assumption that use is proper, the root
cause here is that is doing two things, and it seems best to
cleanly separate them somehow, rather than opting out of duplicates

The first question is if as it stands always intends to build as
well as define whatever is needed for an egg presence.  The second issue
is that appears to deal with one egg, and some packages have

Reading, it strikes me that the proposed approach is to disable
everything in except

So I think the right thing is just to use instead.  That is
reverting 1.9 of Makefile, and that commit does not have an
explanation of why it is correct.  I assert that it is not correct
because of:
  - adding wrong build machinery (not used, and conflicting with the
    package's build system)
  - ending up with an EGG_NAME of xen-4.15.2, which is wrong, and
    results in EGG_NAME not ending up in PLIST and thus not useful

I've appended the python PLIST entries.  It's clear that there are two
eggs: pygrub-0.6 and xen-3.0.

So I say back out 1.9 and then things are ok.  wiz and oskar copied to
explain if I'm confused.


Attachment: signature.asc
Description: PGP signature

Home | Main Index | Thread Index | Old Index