Le 01/07/2014 14:17, akhiezer a écrit :
Date: Tue, 01 Jul 2014 13:51:09 +0200
From: Pierre Labastie <[email protected]>
To: [email protected]
Subject: Re: [blfs-dev] Is Java AWT peer needed for GCC java?
.
.
Back in business,
(Solved my mail problem. My ISP had decided to SPAM some of the mails
and not others, and put them in the webmail trash instead of sending
them to me. It was my first time on their site: I do not like webmail's,
because you see more ads than mails)
Do you have imap access to the account: if so, then you could likely see
all the folders from your ordinary mail client.
Never tried imap. I've learned to use pop (long ago), and never changed.
Will give a try, thanks.
Concerning gcj, I have been pretty disappointed to see that the
development had stalled, but that may be easily understood since the
developers went to OJDK. As DJ says, the only use I see is that you can
build OJDK "from scratch" with gcj. Whether it is silly or not is a
question about where "scratch" is, when we say "linux from scratch".
Certainly coding in binary and create an assembler would be just silly,
I think you just exemplified how much it's a subjective issue. May be
silly to you, but very interesting and useful to others.
Sorry to not having been clear : I meant, in the LFS context. I'm trying
to understand what should go into the books, not what people generally
should do. Actually, I programmed a lot in assembler (and almost never
in binary), for interfacing stuff in my lab, but now, labview is used
everywhere, so nobody is doing that anymore (again, in my lab, not
generally).
but relying on too many external resources certainly defeats the goal of
LFS (for example why not use gentoo if you want to build yourself, and a
regular distro if you just want to use the software, and so on). So
"scratch" is in between... Initially, that was my proposition to not
rely on an external resource to build OJDK. Since I saw that nobody was
really enthusiastic, I lost my interest. OTOH, the page is almost ready
to be added to the book.
Fwiw, +1 from here for builds that are: from-source, from a pre-downloaded
set of tarballs, no binary installer or cfg/bld/inst-time downloads,
and that are fully-configurable prior to runtime.
Concerning Ada, it is really different: the package is mature, Ada is
used by some category of persons (in aerospace industry at least), and
it does not add any new dependencies, so I think we may have it in the
book.
I think Randy did a rather good hint a bunch'o'years back, on Ada,
full-gcc-install, and so on - maybe also including Fortran
( "Fortran From Scratch" - FFS ... ;) .
Sure, ISTR seeing that hint (and just googled to understand the FFS
part. D'you mean fortran is frustrating?:-) . At the time, full GCC with
Ada was also in BLFS, I think. Again a question of where "scratch" is:
is downloading the GNAT compiler from Adacore (which uses the GCC front
end anyway) less "from scratch" than using the GNAT Ada compiler from,
say, Debian (who bootstrapped it using the Adacore compiler)?
Also, I think some of the now BLFS pages have begun as hints.
Back to the subject: I have tried the QT AWT peer, and again, libjawt is
not built. So the only possibility seems to be gtk, which needs
libart_lgpl. OK, since the only usage of GCC-Java would be to build OJDK
from scratch, let me see if OJDK can be built without jawt...
Pierre
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
