Re: metaconfig tweak? (was Re: [perl #7586] STDOUT is not flushedbefore exec().)

On Mon, 16 Sep 2013 11:39:42 +0100, Nicholas Clark <nick@ccl4.org>
wrote:

> On Mon, Sep 16, 2013 at 12:27:45PM +0200, H.Merijn Brand wrote:
> > On Mon, 16 Sep 2013 11:20:24 +0100, Nicholas Clark <nick@ccl4.org>
> > wrote:
> > 
> > > On Mon, Sep 16, 2013 at 10:09:02AM +0100, Nicholas Clark wrote:
> > > 
> > > > It turns out* that $Config{fflushNULL} on HP-UX is *STILL* undef, because
> > > > fflush(NULL) wrongly closes STDIN if STDIN is a pipe (which is what the
> > > > Configure test probes for).
> > > 
> > > Is it worth documenting the specific failure case more clearly? Something
> > > like this?
> > 
> > YES!
> > 
> > Do you want to commit that yourself?
> > Follow-up steps are to regenerate the files in the perl repo and
> > Porting/glossary. Not too much work once you know the tantra. I can do
> > that now if you want
> 
> I think this time I'd prefer it if you did it. My head is rather full with
> a lot of other things.

Doing it right now, alongside the other lingering changes

> If I looks like I'm getting into a regular habit of patching Configure
> then I'll try to figure out how to do it myself (and check that I'm getting
> it right)
> 
> But in particular I'm not sure when I'm patching local units, versus when
> I would be about to make the first change to an unmodified upstream unit.

I have TWO perl source trees. One is the git repo, the other is an
rsync "clone" in which meta generates and modifies the files with
symlinks back into meta for Glossary generation (see .gitignore)

After a generation, I do «xxdiff perl{,-git}/Configure&» to see if what
is generated is now in sync. In an ideal case, that'll just show the
time stamp in the header section.

When I am done (and satisfied), I'll run a complete test in "perl", if
that passes, I'll generate the Glossary, after that, I copy all changed
files to ../perl-git and check against the git repo. I pull/rebase and
run a final make test_porting and commit and push

> Nicholas Clark

-- 
H.Merijn Brand  http://tux.nl   Perl Monger  http://amsterdam.pm.org/
using perl5.00307 .. 5.19   porting perl5 on HP-UX, AIX, and openSUSE
http://mirrors.develooper.com/hpux/        http://www.test-smoke.org/
http://qa.perl.org   http://www.goldmark.org/jeff/stupid-disclaimers/

• [ID 20010825.002] STDOUT is not flushed before exec(). by jms
• Re: [ID 20010825.002] STDOUT is not flushed before exec(). by Abhijit Menon-Sen
• Re: [ID 20010825.002] STDOUT is not flushed before exec(). by Joe Smith
• [perl #7586] STDOUT is not flushed before exec(). by Brian Fraser via RT
• [perl #7586] STDOUT is not flushed before exec(). by James E Keenan via RT
• Re: [perl #7586] STDOUT is not flushed before exec(). by Nicholas Clark
• metaconfig tweak? (was Re: [perl #7586] STDOUT is not flushed beforeexec().) by Nicholas Clark
• Re: metaconfig tweak? (was Re: [perl #7586] STDOUT is not flushedbefore exec().) by H.Merijn Brand
• Re: metaconfig tweak? (was Re: [perl #7586] STDOUT is not flushedbefore exec().) by Nicholas Clark
• Re: metaconfig tweak? (was Re: [perl #7586] STDOUT is not flushedbefore exec().) by H.Merijn Brand
• [perl #7586] STDOUT is not flushed before exec(). by James E Keenan via RT
• Re: [perl #7586] STDOUT is not flushed before exec(). by H.Merijn Brand
• [perl #7586] STDOUT is not flushed before exec(). by James E Keenan via RT