Re: Emacs pretest 28.0.90 is out

[Po Lu]

Po Lu <luangruo@yahoo.com> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> Date: Thu, 09 Dec 2021 14:16:16 +0200
>>> From: Eli Zaretskii <eliz@gnu.org>
>>> Cc: emacs-devel@gnu.org
>>> 
>>> > After implementing the missing functions, the linker complained about
>>> > duplicate copies of mktime, one in libc.a, and the other in gnulib's
>>> > mktime module
>>> 
>>> If we have mktime from Gnulib, then why does linker pull mktime from
>>> libc.a as well?  It's supposed to be satisfied by the first function
>>> by that name that it sees in any library it links in.
>>
>> I think I know the answer: it's because we use localtime, which DJGPP
>> defines on the same source file as mktime.
>>
>> I see that in the Emacs 26.2 build, the function mktime in Gnulib's
>> mktime.c module was renamed to rpl_mktime, which is how the duplicate
>> definition problem was resolved.  The current Gnulib's mktime.c module
>> still allows that (search mktime.c for "rpl"), so all you need to do
>> is arrange for the relevant cpp macro to be defined.  This is supposed
>> to be done by sed2v2.inp, so I guess it is somehow not working now?
>
> It seems that `REPLACE_MKTIME' is 0 in gnulib.mk.  I see the issue now,
> thanks!

Now I'm really lost: temacs.exe is 14.1 MB, but trying to run it from
the Makefile or like this:

C:\DJGPP\src\emacs-28.0.90\src> temacs

Results in "This program is too large to fit in memory."

Any ideas?  Maybe I should fiddle with compiler and linker versions?  Or
perhaps there is some attribute that must be set for this to work on
MS-DOS?

Thanks in advance!