Re: Request for recompilation: quantlib-python
On Sat, 1 Dec 2001, Dirk Eddelbuettel wrote:
> buildd.debian.org shows that the m68k build of quantlib-python dies at
>
> python2.1 setup.py build
> QuantLib/quantlib_wrap.cpp: In function `struct PyObject * _wrap_BoundaryCondition_type(PyObject *, PyObject *, PyObject *)':
> QuantLib/quantlib_wrap.cpp:2482: warning: `struct PyObject * resultobj' might be used uninitialized in this function
> QuantLib/quantlib_wrap.cpp: In function `struct PyObject * _wrap_Period_units(PyObject *, PyObject *, PyObject *)':
> QuantLib/quantlib_wrap.cpp:4099: warning: `struct PyObject * resultobj' might be used uninitialized in this function
> QuantLib/quantlib_wrap.cpp: In function `struct PyObject * _wrap_Date_weekday(PyObject *, PyObject *, PyObject *)':
> QuantLib/quantlib_wrap.cpp:4325: warning: `struct PyObject * resultobj' might be used uninitialized in this function
> QuantLib/quantlib_wrap.cpp: In function `struct PyObject * _wrap_Date_month(PyObject *, PyObject *, PyObject *)':
> QuantLib/quantlib_wrap.cpp:4475: warning: `struct PyObject * resultobj' might be used uninitialized in this function
> make: *** wait: No child processes. Stop.
> make: *** Waiting for unfinished jobs....
> make: *** wait: No child processes. Stop.
> Build killed with signal 15 after 170 minutes of inactivity
>
> That might be spurious. This packages uses SWIG wrappers to the C++ library
> libquantlib0. The resulting .so file is rather large:
>
> edd@homebud:~> ls -hlS /usr/lib/python2.1/site-packages/QuantLib/
> total 2.9M
> -rw-r--r-- 1 root root 2.4M Nov 30 00:24 QuantLibc.so
> -rw-r--r-- 1 root root 223k Nov 29 23:59 QuantLib.pyc
> -rw-r--r-- 1 root root 108k Nov 29 23:35 QuantLib.py
> -rw-r--r-- 1 root root 21k Nov 29 23:59 defaults.pyc
> -rw-r--r-- 1 root root 14k Nov 7 06:49 defaults.py
> -rw-r--r-- 1 root root 3.2k Nov 29 23:59 __init__.pyc
> -rw-r--r-- 1 root root 3.1k Oct 4 07:16 __init__.py
>
> and the compilation step for this 2.9mb baby takes a fairly long time on my
> i386 (k6-II 300, 256mb ram). A small machine with little ram might cough on
> it.
It probably didn't, but just stopped because it thought it did. Which is
not the same thing; "activity" is being measured by output to
stdout. Since large files as yours indeed need their time, they're quite
likely to fail.
> Could someone do me a favour and try to do this by hand (ie call
> "debian/rules build") and see what happens?
Luckily, this isn't necessary. What needs to be done is to change the
timelimit this particular compilation can have no output (which can be
done on a per-package basis), and retry.
Which already happened, since the default timeout is 150 minutes. If you
could give us an estimation of the time needed to compile this one, next
try will probably be successfull ;-)
BTW: note that -m68k, an alias for -68k, is a users' forum; Debian
m68k-porters can be reached on m68k-build@nocrew.org
--
wouter dot verhelst at advalvas dot be
"Human knowledge belongs to the world"
-- From the movie "Antitrust"
Reply to: