Bootstrap fails on Win32

Bootstrap fails on Win32

[B. Anyos]

Hi,

Doing a bootstrap compilation fails with file 'characters.el'
The problem reported is:

  make[1]: *** No rule to make target `../lisp/international/characters.elc', needed by `DOC'.  Stop.

When I tried to compile it manually I got the following error message:

  > emacs.exe -batch -f batch-byte-compile characters.el
  Loading d:/Personal/CVS/emacs/lisp/international/uni-category.el (source)...

  In toplevel form:
  characters.el:1136:1:Error: Symbol's value as variable is void: this-val

Checking the history this section was recently added to revision 1.76 about a week ago.
Can someone fix this please?

Thanks,
 Bela

Re: Bootstrap fails on Win32

[Francis Litterio]

B. Anyos wrote:

> Doing a bootstrap compilation fails with file 'characters.el'

> When I tried to compile it manually I got the following error message:

>   characters.el:1136:1:Error: Symbol's value as variable is void: this-val

I just updated from CVS, and the reference to that variable appears to
have been removed from characters.el.
--
Fran

Re: Bootstrap fails on Win32

[B. Anyos]

I checked the file and it's been unchanged for 13 days.
Doing a bootstrap still fails with exactly the same reason....

This has been confirmed on another machine with a separate environment.

Bela

Francis Litterio wrote:
> B. Anyos wrote:
> 
>> Doing a bootstrap compilation fails with file 'characters.el'
> 
>> When I tried to compile it manually I got the following error message:
> 
>>   characters.el:1136:1:Error: Symbol's value as variable is void: this-val
> 
> I just updated from CVS, and the reference to that variable appears to
> have been removed from characters.el.
> --
> Fran
> 
> 
> 
> 
> 

Re: Bootstrap fails on Win32

[Jason Rumney]

B. Anyos wrote:
> I checked the file and it's been unchanged for 13 days.
> Doing a bootstrap still fails with exactly the same reason....
>   
Are the lisp/international/uni-*.el files checked out as binary? (cvs 
update -kb lisp/international/uni-*.el)

Re: Bootstrap fails on Win32

[B. Anyos]

That was it! Thanks!
Code compiles w/o errors now.

Jason Rumney wrote:
> B. Anyos wrote:
>> I checked the file and it's been unchanged for 13 days.
>> Doing a bootstrap still fails with exactly the same reason....
>>   
> Are the lisp/international/uni-*.el files checked out as binary? (cvs
> update -kb lisp/international/uni-*.el)
> 
> 
> 
> 
> 

Re: Bootstrap fails on Win32

[Juanma Barranquero]

On Thu, Sep 11, 2008 at 12:00, B. Anyos <banyos@freemail.hu> wrote:

> That was it! Thanks!
> Code compiles w/o errors now.

Then, I'd recommend reading nt/INSTALL. The *Preliminaries* section
talks about the files that must be checked out with -kb. Perhaps it
will save you trouble in the future.

Happy compiling,

   Juanma

Re: Bootstrap fails on Win32

[Francis Litterio]

Juanma Barranquero wrote:

> B. Anyos wrote:
>
>> That was it! Thanks!
>> Code compiles w/o errors now.
>
> Then, I'd recommend reading nt/INSTALL. The *Preliminaries* section
> talks about the files that must be checked out with -kb.

I've found that if the entire source tree is checked out on Windows
using -kb, the build works just fine.  I don't think there's a need to
check out just some files using -kb.
--
Fran

Re: Bootstrap fails on Win32

[Juanma Barranquero]

On Thu, Sep 11, 2008 at 16:31, Francis Litterio <flitterio@gmail.com> wrote:

> I've found that if the entire source tree is checked out on Windows
> using -kb, the build works just fine.  I don't think there's a need to
> check out just some files using -kb.

Yes.

But IIRC, when you commit changes to the CVS, -kb files are sent
entirely, while normal non-kb text files are patched (only differences
are sent). That makes a difference for big files like ChangeLogs.

   Juanma

Re: Bootstrap fails on Win32

[Andreas Schwab]

"Juanma Barranquero" <lekktu@gmail.com> writes:

> But IIRC, when you commit changes to the CVS, -kb files are sent
> entirely, while normal non-kb text files are patched (only differences
> are sent). That makes a difference for big files like ChangeLogs.

It's the other way round: patches are received (or not) when updating.
When committing (or diffing) the whole contents are always sent, since
CVS does not have the concept of pristine sources, so it cannot compute
the changes locally.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Bootstrap fails on Win32

[Juanma Barranquero]

On Thu, Sep 11, 2008 at 18:20, Andreas Schwab <schwab@suse.de> wrote:

> It's the other way round: patches are received (or not) when updating.
> When committing (or diffing) the whole contents are always sent, [...]

Aha.

Then checking out all the sources with -kb has an even greater impact,
because it affects all updates even if you don't have write access...

 Juanma

Re: Bootstrap fails on Win32

[Stefan Monnier]

> But IIRC, when you commit changes to the CVS, -kb files are sent
> entirely, while normal non-kb text files are patched (only differences
> are sent). That makes a difference for big files like ChangeLogs.

Actually, it makes no difference for those cases.  It does have some
performance impacts, but the more problematic effet is that CVS
considers such files as unmergeable.


        Stefan