On Feb 25, 2006, at 4:55 AM, David R. Morrison wrote:
Well, upon further investigation this is a very curious story.
If I repeat the same steps as Michèle, but starting by
bootstrapping from branch_0_24, then the update goes OK. But if
the bootstrap is from HEAD, the update fails.
The error messages produced are less familiar to me, since they
come from some newer section of code in HEAD. (And I personally
don't run HEAD very often.)
Can one of the HEAD gurus enlighten us about the kind of change
which was made? Here, it looks like an error is generated as the
buildlock package is removed. I guess some new dependency checking
is done at that point which wasn't done in the past? Whatever the
cause, with branch_0_24 the updating continues and eventually all
dependencies are satisfied, whereas here, we cannot continue.
I would appreciate help from others as I try to understand what's
going on here.
Thanks,
Dave
Sorry to be sending a series of messages here, but I am starting to
understand better what is going on. The error message had misled me
into thinking that the error had to do with the buildlock package,
but actually it just seems to be an error issued by the new
Fink::SysState module.
The situation is this: foo and bar are initially splitoffs in the
same package, and foo depends on a specific version of bar. But in
the revision, foo and bar are in different packages. Due to
dependencies, fink updates bar before foo (which is correct). But
having updated foo, the versioned dependence which bar has on the old
version of foo is no longer valid, and Fink::SysState complains.
Why did this work OK in the past? Well, dpkg itself does not check
as strictly as SysState, and it allows you to upgrade bar even though
the upgrade breaks the foo dependency. Of course, by the time the
entire fink update is done, things will be fine again.
One possible fix for this would be to have a mechanism to recognize
that a group of packages need to be updated together, and that
Fink::SysState should not be expected to give a consistent answer
until they have all been updated. I'm not sure how to do this, however.
Am I on the right track, in understanding what's going on here? How
can we handle updates of the kind desired?
-- Dave
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642
_______________________________________________
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fink-devel