sage-4.0.alpha0 build failure on Atom (EeePC 901)

[Kevin Horton]

As an experiment, I attempted to build sage-4.0.alpha0 on an EeePC 901
Linux, with the stock Xandros OS. The build failed with:

checking for fdatasync... yes
configure: creating ./config.status
config.status: creating Makefile
config.status: creating sqlite3.pc
make[2]: Entering directory `/home/user/apps/sage-4.0.alpha0/spkg/
build/sqlite-3.5.3.p3/src'
sed -e s/--VERS--// ./src/sqlite.h.in | \
sed -e s/--VERSION-NUMBER--// >sqlite3.h
make[2]: *** No rule to make target `tool/lemon.c', needed by
`lemon'. Stop.
make[2]: Leaving directory `/home/user/apps/sage-4.0.alpha0/spkg/build/
sqlite-3.5.3.p3/src'
Error making sqlite

real 0m24.585s
user 0m8.730s
sys 0m5.770s
sage: An error occurred while installing sqlite-3.5.3.p3
Please email sage-devel http://groups.google.com/group/sage-devel
explaining the problem and send the relevant part of
of /home/user/apps/sage-4.0.alpha0/install.log. Describe your
computer, operating system, etc.
If you want to try to fix the problem, yourself *don't* just cd to
/home/user/apps/sage-4.0.alpha0/spkg/build/sqlite-3.5.3.p3 and type
'make'.
Instead type "/home/user/apps/sage-4.0.alpha0/sage -sh"
in order to set all environment variables correctly, then cd to
/home/user/apps/sage-4.0.alpha0/spkg/build/sqlite-3.5.3.p3
(When you are done debugging, you can type "exit" to leave the
subshell.)
make[1]: *** [installed/sqlite-3.5.3.p3] Error 1
make[1]: Leaving directory `/home/user/apps/sage-4.0.alpha0/spkg'

real 227m40.319s
user 125m21.420s
sys 4m25.560s
python: can't open file '/home/user/apps/sage-4.0.alpha0/devel/sage/
doc/common/builder.py': [Errno 2] No such file or directory

=======================
I'm not sure what other info is relevant to provide.

Note that the recorded build times are suspect, as I updated the clock
after starting the build. It was slow by about an hour.

--
Kevin Horton
Ottawa, Canada

[mabshoff]

On May 19, 6:31 am, Kevin Horton <khorto...@rogers.com> wrote:
> As an experiment, I attempted to build sage-4.0.alpha0 on an EeePC 901  
> Linux, with the stock Xandros OS.  The build failed with:

<SNIP>

> I'm not sure what other info is relevant to provide.

Look for install.log, compress it, upload it somewhere and post a link
here.

> Note that the recorded build times are suspect, as I updated the clock  
> after starting the build.  It was slow by about an hour.

Well, if you fiddle with the clock all bets are off. So for starters
restart the build with "make" and se if it fails the same time. If it
blows up check that the spkg in question is not corrupted. If it isn't
send a link to the compressed log as suggested above.

> --
> Kevin Horton
> Ottawa, Canada

Cheers,

Michael

[Kevin Horton]

On 19 May 2009, at 11:45, mabshoff wrote:

> On May 19, 6:31 am, Kevin Horton <khorto...@rogers.com> wrote:
>> As an experiment, I attempted to build sage-4.0.alpha0 on an EeePC
>> 901
>> Linux, with the stock Xandros OS. The build failed with:
>

> Well, if you fiddle with the clock all bets are off. So for starters
> restart the build with "make" and se if it fails the same time. If it
> blows up check that the spkg in question is not corrupted. If it isn't
> send a link to the compressed log as suggested above.

The tarball is corrupted. It worked fine for installs on OS X and a
Ubuntu VM, but the MD5 changed when I moved it to the EeePC.

I'll report success or failure, after I get a good tarball on the EeePC.

This is the second time I've gotten bitten by tarballs with bad MD5s.
There has got to be a better file format to use - something where the
archive validity can be checked before launching into a long install.

[mabshoff]

On May 19, 12:45 pm, Kevin Horton <khorto...@rogers.com> wrote:
> On 19 May 2009, at 11:45, mabshoff wrote:
>
> > On May 19, 6:31 am, Kevin Horton <khorto...@rogers.com> wrote:
> >> As an experiment, I attempted to build sage-4.0.alpha0 on an EeePC  
> >> 901
> >> Linux, with the stock Xandros OS.  The build failed with:
>
> > Well, if you fiddle with the clock all bets are off. So for starters
> > restart the build with "make" and se if it fails the same time. If it
> > blows up check that the spkg in question is not corrupted. If it isn't
> > send a link to the compressed log as suggested above.
>
> The tarball is corrupted.  It worked fine for installs on OS X and a  
> Ubuntu VM, but the MD5 changed when I moved it to the EeePC.

Well, that isn't really our fault, is it?

> I'll report success or failure, after I get a good tarball on the EeePC.
>
> This is the second time I've gotten bitten by tarballs with bad MD5s.  
> There has got to be a better file format to use - something where the  
> archive validity can be checked before launching into a long install.

There is a ticket to add potential corruption detection. See
http://trac.sagemath.org/sage_trac/ticket/329 - patches welcome.

But if you look in the log you ought to see that tar did not unpack
the whole tarball. And if you see any failure while building Sage you
ought to check the md5sum first.

> --
> Kevin Horton
> Ottawa, Canada

Cheers,

Michael