Re: Building Emacs-cvs on CygwinRe

Re: Building Emacs-cvs on CygwinRe

[djh]

I stumbled upon a way to get emacs built on cygwin (at least it worked once) 
using latest cvs emacs source (as of September 21).

My previous mail mentioned the unsuccessfull build which resulted in a core dump, just as it did form another cygwin user.

While trying to isolate the problem, I changed
two items (the second one probably had no use, but I am mentioning it in case).

I tokk the last offending command which broke the makefile and snipped off
the last "   $$wins   " parameter and changed directories to ..emacs/src :

cd ../src

# Then I pasted the offending command with "$$wins" snipped off and ran it.

./bootstrap-emacs.exe -batch --no-site-file --multibyte -l autoload --eval '(setq generated-autoload-file "/cygdrive/c/emacs/cvs/emacs/lisp/loaddefs.el")' -f batch-update-autoloads
...
......
Saving file /cygdrive/c/emacs/cvs/emacs/lisp/loaddefs.el...
Wrote /cygdrive/c/emacs/cvs/emacs/lisp/loaddefs.el

----
Well as an isolated instance and without the "$$wins" it ran fine.
Then I added the two lines in the ../lisp/Makefile  to see just what $wins was.

echo $$wins
echo Above line was added for testing.

# Then ran 

  make bootstrap

# Hurry... This time the make bootstrap puzzlingly succeeded....... 

Angelo, you might also try this and see if it works for you, as a tempeoraty fix till the real problem is ironed out.

I don't know why..... Did the manual run of
./bootstrap-emacs.exe -batch --no-site-file --multibyte -l autoload --eval '(setq generated-autoload-file "/cygdrive/c/emacs/cvs/emacs/lisp/loaddefs.el")' -f batch-update-autoloads

"without the $wins parameter..
...initialize a file or something that makes subsequent runs work???

Regards,
  Darel Henman

Re: Building Emacs-cvs on CygwinRe

[Jason Rumney]

djh wrote:
> Well as an isolated instance and without the "$$wins" it ran fine.
> Then I added the two lines in the ../lisp/Makefile  to see just what $wins was.
>
> echo $$wins
> echo Above line was added for testing.
>   

$$wins was already being echoed in your original report, and there 
doesn't appear to be anything wrong with it:

    Directories: /tmp/emacs/lisp/. /tmp/emacs/lisp/./calc
    /tmp/emacs/lisp/./calendar /tmp/emacs/lisp/./emacs-lisp
    /tmp/emacs/lisp/./emulation /tmp/emacs/lisp/./erc /tmp/emacs/lisp/./eshell
    /tmp/emacs/lisp/./gnus /tmp/emacs/lisp/./international
    /tmp/emacs/lisp/./language /tmp/emacs/lisp/./mail /tmp/emacs/lisp/./mh-e
    /tmp/emacs/lisp/./net /tmp/emacs/lisp/./obsolete /tmp/emacs/lisp/./play
    /tmp/emacs/lisp/./progmodes /tmp/emacs/lisp/./term
    /tmp/emacs/lisp/./textmodes /tmp/emacs/lisp/./url


> I don't know why..... Did the manual run of
> ./bootstrap-emacs.exe -batch --no-site-file --multibyte -l autoload --eval '(setq generated-autoload-file "/cygdrive/c/emacs/cvs/emacs/lisp/loaddefs.el")' -f batch-update-autoloads
>
> "without the $wins parameter..
> ...initialize a file or something that makes subsequent runs work???
>   

It would have created an incomplete loaddefs.el file. But that should 
not make a difference to subsequent bootstraps, which will recreate the 
file from scratch.

Re: Building Emacs-cvs on CygwinRe

[Angelo Graziosi]

On Thu, 21 Sep 2006, djh wrote:

> 
> Then I added the two lines in the ../lisp/Makefile  to see just 
> what $wins was.
> 

I had yet realized (with a few tears!) what $wins means, but, as you can
see from the posts, it does not clarify the problem.

Now, mysteriously, the buid is completed to the end (even with new CVS
updates) but Emacs segment faults in a few seconds. Only the builds I did
Aug 15, 2006 seems (very) stable.


Cheers,

   Angelo 

Re: Building Emacs-cvs on CygwinRe

[Eli Zaretskii]

> From: "djh" <henman@it.to-be.co.jp>
> Date: Thu, 21 Sep 2006 18:14:45 +0900
> Cc: Angelo.Graziosi@roma1.infn.it
> 
> Angelo, you might also try this and see if it works for you, as a tempeoraty fix till the real problem is ironed out.

It's good to know there's a workaround, but, as Jason points out,
there's nothing wrong with value of $wins, so your observations don't
mean any progress in understanding the original problem.