Luca liliana.perossa at fastwebnet.it
Tue Dec 5 21:41:35 PST 2006

Luca wrote:
> Hello; trying another toolchain I've been encumbered by the "old" cannot
> compute sizeof (long double), 77.
> Just a note not hash-style related:
> When compiling gcc with java language enabled for the awt to work
> correctly it should be passed "--enable-java-awt=[gtk,xlib,qt]".
> Luca

Problems solved and a working toolchain is going to be finished in some
time so I could post a tarball of testsuites logs for chapter6 today
(probably this afternoon - timezone Europe/Rome).
A behavior I just saw: gcc-4.3.0 uses bootstrap by default unless
specified "--disable-bootstrap" in configure. I configured gcc-pass2
with "--disable-bootstrap" and while the specs tell it uses exclusively
"--hash-style=gnu" it didn't compile itself with hash-style (which
happens when bootstrapping). For the make the choices are the "standard"
ones: make, make bootstrap or make profiledbootstrap.
Bootstrap phase consists of: creating a prev-gcc (xgcc) to use to
compiling itself and the bootstrap uses stage{2,3} and compiles finally
in stage4.
Note: as I already reported to re-compile gcc after gcc-pass1 gmp/mpfr
needed otherwise it broke during configure.
Note2: For the testsuites I'll check but I didn't find yet
testsuites-results on native reiser4 file-systems so I don't have much
to compare except for results on other file-systems.

To Bruce, Randy or other of blfs team:
Note: about gcc compiled with java support: yesterday I posted the note
about the "--enable-jawa-awt=[gtk,xlib,qt]". Yesterday afternoon I
talked with Alexander and talked about a possible note in book to say,
for example:
If you would like to compile gcc-suite with java (gcj) language support,
for the awt to work correctly you have to re-install gcc after X, gtk,
qt, libart and pass to configure: "--with-x"
"--enable-java-awt=[gtk,xlib,qt]" "--enable-gtk-cairo"


More information about the lfs-dev mailing list