Re: Trunk bootstrap failure [Cygwin]

[Dan Nicolaescu]

Angelo Graziosi <address@hidden> writes:

> Il 09/07/2010 1.07, Dan Nicolaescu ha scritto:
>> Angelo Graziosi<address@hidden>  writes:
>>
>>> Il 08/07/2010 7.54, Dan Nicolaescu ha scritto:
>>>> Angelo Graziosi<address@hidden>   writes:
>>>>
>>>>> Since 'start_of_text' is not used by Cygwin build, shouldn't 'ecrt0.c'
>>>>> be completely unnecessary?
>>>>
>>>> That's exactly the reason for this discussion: getting rid of ecrt0.c.
>>>
>>> Hmm... bootstrapping rev. 100753 *without* patches (i.e. using
>>> ecrt0.c), shows that:
>>>
>>> $ grep -R start_of_text emacs/inst/usr/local/emacs
>>>
>>> is *empty* (emacs/inst/usr/local/emacs is the Emacs installation
>>> directory). This would mean that the function 'start_of_text' is not
>>> compiled any more in the Cygwin build. Right?
>>
>> Sure, before yesterday it was just sitting there unused.
>>
>
> What about 'ecrt0.c'? It seems that it happens the same for it.
>
> For completeness, the bootstrap (rev. 100755) I did applying *only*
> this patch:
>
> =====================================================
> --- emacs.orig/configure        2010-07-02 11:27:38.000000000 +0200
> +++ emacs/configure     2010-07-06 10:45:21.656250000 +0200
> @@ -5864,7 +5864,7 @@
>  case $opsys in
>    cygwin )
>      LIB_MATH=
> -    START_FILES='ecrt0.o'
> +    START_FILES='pre-crt0.o'
>      ;;
>    darwin )
>      ## Adding -lm confuses the dynamic linker, so omit it.
> =====================================================
>
> is still working.

As expected.  Nothing in this area has changed since you tried this a
couple of days ago.
I'd like to make this change.
And in order to plan for the unexpected: if in the next few months you
get weird memory errors, are you willing to test again with this patch reverted?