make bootstrap fails from fresh CVS checkout (was: GNU/Linux (Ubuntu 7.10) bootstrap from a CVS checkout fails ("Maximum buffer size exceeded"))

[Ted Zlatanov <tzz@lifelogs.com>]

Further investigation:

I traced the error to fileio.c:3867:

if (XINT (end) != st.st_size
    || ((int) st.st_size * 4) / 4 != st.st_size)
  /* extra parameters added by me */
  error ("Maximum buffer size exceeded: XINT(end)=%d, st.st_size=%d", XINT(end), st.st_size);

This shows that XINT(end) and st.st_size are both 0, by the way.  This
agrees with my observation (see the original message) that no data is
read from the file that's being loaded before the maximum buffer size is
exceeded.

If I comment out that check, the compilation proceeds until:

Dumping under names emacs and emacs-23.0.50
emacs: Can't allocate buffer for /home/tzz/source/emacs/src/temacs

in unexelf.c, line 721.  I can't avoid that error so I'm unable to build Emacs.

This is under GCC 4.1.3 on Ubuntu:

Using built-in specs.
Target: i486-linux-gnu
Configured with: ../src/configure -v --enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --enable-nls --with-gxx-include-dir=/usr/include/c++/4.1.3 --program-suffix=-4.1 --enable-__cxa_atexit --enable-clocale=gnu --enable-libstdcxx-debug --enable-mpfr --enable-checking=release i486-linux-gnu
Thread model: posix
gcc version 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu2)

I also tried GCC 3.4 with the same result, in case the 4.1.3 prerelease
was a problem.  The error was exactly the same.

It's been almost a month since I reported this problem; I'd love to
figure out the cause.  Is anyone willing to help?  My C knowledge is
insufficient to figure out what's happening.

Ted

On Fri, 30 Nov 2007 17:17:06 -0600 Ted Zlatanov <tzz@lifelogs.com> wrote: 

TZ> Is this known to be broken, or am I missing something obvious?
TZ> I am surprised a `make bootstrap' failure is happening on Ubuntu, but
TZ> even if the problem is not with Emacs itself, it certainly looks like it
TZ> (or a library interaction triggered by load-with-code-conversion) from
TZ> my investigation.

TZ> On Tue, 20 Nov 2007 14:35:46 -0600 Ted Zlatanov <tzz@lifelogs.com> wrote: 

TZ> The full log follows.  This is with a clean CVS checkout on a new
TZ> Ubuntu 7.10 install.

TZ> After looking at loadup.el, I see the problem starts with the lines:

TZ> (setq load-source-file-function 'load-with-code-conversion)
TZ> (load "files")

TZ> when I remove the load-source-file-function setq, the files.el code is
TZ> loaded fine.  The bootstrap fails later at international/utf-8.el
TZ> because it can't read a ? character, but if I have the setq the next
TZ> Lisp library loaded causes the buffer size error.

TZ> In addition, strace of the temacs process shows that no data is read()
TZ> from the .el file.  The error happens *before* which to me indicates
TZ> this is not a problem with the Lisp code.  I checked it anyhow, and
TZ> it's a clean checkout from CVS without corruption.  I suspect it's an
TZ> error at the C level but I don't know enough about the bootstrap
TZ> process to debug this.

TZ> I searched older discussions and didn't see this.  Sorry if it's an
TZ> obvious problem.