[rt.cpan.org #75664] Fails 09parser test in Cygwin with *** fatal error

[Sisyphus via RT]

Wed Mar 14 04:13:11 2012: Request 75664 was acted upon.
Transaction: Correspondence added by SISYPHUS
Queue: Inline
Subject: Fails 09parser test in Cygwin with *** fatal error
Broken in: 0.50
Severity: Important
Owner: Nobody
Requestors: starrychloe-***@public.gmane.org
Status: open
Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=75664 >


On Wed Mar 14 02:34:10 2012,

I updated some Cygwin packages, and then it gave me an error saying I
needed
to run "rebaseall", which I've never seen before. I had to run it from
only
ash or dash, which I've never heard of before. I ran it, it did
something,
and then I tried to install Inline::C again and it worked this time.

Thanks for that - good to know that 'rebaseall' fixes the problem.
I can actually reproduce the problem on my Cygwin installation ... and I
expect 'rebaseall' will fix the problem here, too, if I can just
discover wtf one needs to do in order to successfully run that command.

I keep getting an error which, in essence, tells me that I need to do
exactly what I just did, in order to successfully run 'rebaseall'.

Cheers,
Rob

[David Oswald via RT]

Wed Mar 14 08:03:46 2012: Request 75664 was acted upon.
Transaction: Correspondence added by daoswald
Queue: Inline
Subject: Re: [rt.cpan.org #75664] Fails 09parser test in Cygwin with *** fatal error
Broken in: 0.50
Severity: Important
Owner: Nobody
Requestors: starrychloe-***@public.gmane.org
Status: open
Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=75664 >


This shouldn't be an Inline::C bug. The issue can impact just about
anything that runs under cygwin, and that uses DLL's. It's likely
that what is getting clobbered is the C compiler. Apparently after
upgrading some packages on cygwin, a rebaseall command may be
necessary. The symptom seems to be wonky behavior across a number of
assorted tools (not just Perl, or the C compiler).

Here are some links on the topic:

http://cygwin.wikia.com/wiki/Rebaseall
http://www.heikkitoivonen.net/blog/2008/11/26/cygwin-upgrades-and-rebaseall/

My suggestion is to document the issue in a future release of Inline
by placing the following segment in the Inline FAQ, or the Inline::C
README (not sure which would be better):

Q: Why am I getting errors under cygwin after
upgrading some cygwin packages?
A: We have no idea, but it seems to be a cygwin
issue that will affect a number of
tools on your system, including the C compiler.
The solution is to execute a
'rebaseall' after terminating all long-running processes.
Here are links to a couple of articles that describe
how to do that:

...and then include the links above, or just copy and paste the
examples of how to run rebaseall.

[Chloe via RT]

Wed Mar 14 13:17:33 2012: Request 75664 was acted upon.
Transaction: Correspondence added by starrychloe
Queue: Inline
Subject: Re: [rt.cpan.org #75664] Fails 09parser test in Cygwin with *** fatal error
Broken in: 0.50
Severity: Important
Owner: Nobody
Requestors: starrychloe-***@public.gmane.org
Status: open
Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=75664 >


You also have to turn off any services with cygrunsrv.
Then run ash or dash from DOS command prompt.
Then run rebaseall.

[Sisyphus via RT]

Thu Mar 15 06:20:05 2012: Request 75664 was acted upon.
Transaction: Correspondence added by SISYPHUS
Queue: Inline
Subject: Fails 09parser test in Cygwin with *** fatal error
Broken in: 0.50
Severity: Important
Owner: Nobody
Requestors: starrychloe-***@public.gmane.org
Status: open
Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=75664 >

Post by David Oswald via RT
My suggestion is to document the issue in a future release of Inline

I generally take the view that, if it's not an Inline issue, then
there's no need to take any action at all.
But this non-Inline issue is a little different - in that it's not
immediately apparent that it has nothing to do with Inline. (At least,
it wasn't to me ... and I was damn relieved when Chloe's follow up
report appeared.)
So, I think I'll do as you suggest and create some documentation about
it in the Inline-FAQ.

Not that anyone ever checks the Inline-FAQ ... but at least it would
give us a ready-made link to which we can (indignantly ;-) refer future
questioners about this problem.
When that's done, I'll close this bug report - but I'll leave it open
for the meantime.

Thanks !!

Cheers,
Rob