Re: staging broken

[David Kastrup]

Graham Percival <address@hidden> writes:

> hmm, I think I forgot to do something while reconfiguring my
> desktop.  Anyway, make doc fails.
>
> I see **tons** of messages like
>   Defined "column-markup" function in #<directory (lily) 411399f0>
> which doesn't look like it should exist in production code.  It's
> certainly not helpful when trying to debug doc build failures.

It's a pretty safe bet that this is the current incantation of the
Guilev2 patch again.  Since I am apparently the only one ever reviewing
it (let alone test it with Guilev1) and missed the latest push window...

I don't remember seeing other patches touching the module system
recently, but I might not have been looking closely enough.

> I also see errors like this:
> programming error: Parsed object should be dead: static
> scm_unused_struct* Prob::mark_smob(scm_unused_struct*)
> continuing, cross fingers
> programming error: Parsed object should be dead: static
> scm_unused_struct* Context_def::mark_smob(scm_unused_struct*)
> continuing, cross fingers
> programming error: Parsed object should be dead: static
> scm_unused_struct* Context_mod::mark_smob(scm_unused_struct*)
> continuing, cross fingers
>
> anybody recognize what's happening?

I think this one has been in for about a month already, but people
generally shrugged shoulders and said "too bad".  That was the one that
lead to the suggestion that Patchy should compile using
-ddebug-gc-assert-parsed-dead by default and refuse merging if anything
bad turned up.  We could not have done it at that time because master
was already bad.  And nobody has bothered cleaning up his code after
that discussion.

-- 
David Kastrup