On Fri, May 14, 2021 at 12:17 PM Jan Groenewald <j...@aims.ac.za> wrote:
>
> Hi
>
> On Fri, 14 May 2021 at 13:12, Dima Pasechnik <dimp...@gmail.com> wrote:
>>
>> On Fri, May 14, 2021 at 11:53 AM Dima Pasechnik <dimp...@gmail.com> wrote:
>> >
>> > On Fri, May 14, 2021 at 11:21 AM Jan Groenewald <j...@aims.ac.za> wrote:
>> > >
>> > > Hi
>> > >
>> > > On Fri, 14 May 2021 at 11:26, François Bissey <frp.bis...@gmail.com>
>> > > wrote:
>> > >>
>> > >> It looks like maxima.fas has not been built. Last time I saw that happen
>> > >> in sage-on-gentoo was because the patch to build it was missing.
>> > >> Not sure how it could happen on your setup, the log needs further
>> > >> inspection.
>> > >>
>> > >> > On 14/05/2021, at 21:18, John Cremona <john.crem...@gmail.com> wrote:
>> > >> >
>> > >> > I have been installing sage-9.3 from the tarball on a few ubuntu
>> > >> > machines, just doing ./configure and then make. On one machine
>> > >> > running ubuntu 20.04.02 LTS maxima fails to build. I have attached
>> > >> > the log files.
>> > >> >
>> > >> > This is a machine on which I tried a few weeks ago to help the build
>> > >> > process by installing as many of the dependencies as possible. My
>> > >> > experience so far has been that this has caused so many things to go
>> > >> > wrong that I prefer the old way -- build everything -- since that just
>> > >> > takes more computer time, while the new way (so far) uses a lot more
>> > >> > of my time.
>> > >
>> > >
>> > > config.log:
>> > > ## Checking whether SageMath should install SPKG libffi... ##
>> > > ## ------------------------------------------------------- ##
>> > > configure:24887: checking for LIBFFI
>> > > configure:24894: $PKG_CONFIG --exists --print-errors "libffi"
>> > > configure:24897: $? = 0
>> > > configure:24911: $PKG_CONFIG --exists --print-errors "libffi"
>> > > configure:24914: $? = 0
>> > > configure:25100: result: yes
>> > > configure:25108: will use system package and not install SPKG libffi
>> > >
>> > > maxima log:
>> > > Condition of type: SIMPLE-ERROR
>> > > Error code 1 when executing
>> > > (EXT:RUN-PROGRAM "gcc" ("-o" "binary-ecl/maxima-package.fas"
>> > > "-L/usr/local/sage/sage-9.3/local/lib/" "/tmp/eclinit8GQbdq.o"
>> > > "binary-ecl/maxima-package.o"
>> > > "-Wl,--rpath,/usr/local/sage/sage-9.3/local/lib/" "-shared"
>> > > "-Wl,-rpath-link,/usr/local/sage/sage-9.3/local/lib"
>> > > "-L/usr/local/sage/sage-9.3/local/lib"
>> > > "-Wl,-rpath,/usr/local/sage/sage-9.3/local/lib"
>> > > "-Wl,-rpath-link,/usr/local/sage/sage-9.3/local/lib"
>> > > "-L/usr/local/sage/sage-9.3/local/lib"
>> > > "-Wl,-rpath,/usr/local/sage/sage-9.3/local/lib" "-lecl" "-lgmp" "-lgc"
>> > > ...)):
>> > > /usr/bin/ld: cannot find -lffi
>> > > collect2: error: ld returned 1 exit status
>> > > Available restarts:
>> > >
>> > > 1. (CONTINUE) Continues anyway.
>> > > 2. (RESTART-TOPLEVEL) Go back to Top-Level REPL.
>> > >
>> > > Broken at DO-COMPILE-ECL.
>> > > File:
>> > > #P"/usr/local/sage/sage-9.3/local/var/tmp/sage/build/maxima-5.44.0/src/src/maxima.system"
>> > > (Position #1601)
>> > >
>> > >
>> > >
>> > > Do you perhaps need to sudo apt install libffi-dev?
>> >
>> >
>> > One way I can explain this is a broken system-wide install of libffi-dev
>> > I don't know how this can be achieved, other than having multiple
>> > system-wide copies of libffi,
>> > or stuff manually removed.
>> >
>> > The other possibility is that the linker is told not to look for
>> > system-wide libffi - can this be achieved with -L and -Wl,-rpath ?
>> >
>> > On the other hand, it's a bug in maxima, which does not try to check
>> > for presense of libffi, just assumes it's there.
>>
>> oops, I meant to say, it's a bug in ECL, as this failing call to gcc
>> is emitted by ECL.
>
>
> Why did SageMath say this? Was it checking for runtime libffi or headers?
it's checking with pkg-config first of all.
see build/pkgs/libffi/spkg-configure.m4
If this test succeeds, it's assumed everything is fine.
Note that libffi is a pre-req for ECL, and ECL is built just fine,
as far as one can see.
I gather it must be a bug in ECL, which struggles to create a correct
incantation of C compiler calls
from a unique (Copyright JEC 2021 :-)) concostion of gmp built by Sage
and libffi provided by the system.
ECL checks for libffi carefully in its configure, as far as I know.
John, can you post logs/pkgs/ecl-21.2.1.log ?
>
> config.log:
> ## Checking whether SageMath should install SPKG libffi... ##
> ## ------------------------------------------------------- ##
> configure:24887: checking for LIBFFI
> configure:24894: $PKG_CONFIG --exists --print-errors "libffi"
> configure:24897: $? = 0
> configure:24911: $PKG_CONFIG --exists --print-errors "libffi"
> configure:24914: $? = 0
> configure:25100: result: yes
> configure:25108: will use system package and not install SPKG libffi
>
> Regards,
> Jan
> --
> .~.
> /V\ Jan Groenewald
> /( )\ www.aims.ac.za
> ^^-^^
>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sage-devel/CAAg%3Dp_3J7QyF7yTjYY3ecPz5U5uA6E2bw_iBiHYcXw4b0siBEA%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-devel/CAAWYfq008v7NJzP6Kz-Jk_06d8yseshs%2BkGS2%3DanPzucH4V64A%40mail.gmail.com.
