Re: Bug#295097: quantlib build failures due to timeout
severity 295097 normal
thanks
On 13 February 2005 at 11:31, Christopher Martin wrote:
| Package: quantlib
| Version: 0.3.8-2
| Severity: important
|
| Hello,
|
| The latest quantlib package failed to build on 3 architectures (mips,
| mipsel, and s390) due to many minutes of "inactivity" finally ended by the
| buildd. This seems to be a recurring problem with quantlib.
Yes, I am aware of this (see below). Quantlib has fairly complicated C++ code
that requires a lot of ram and cpu resources to be built (and used, being a
library). And AFAIK a lot of the autobuilders are fairly old and laughable
machines by today's standards. Quantlib's inclusion of boost created a
mismatch there as it further extended the requirements in terms of cpu and
ram.
| It would be great if you could upload a fix or workaround, ideally with an
I am unaware of how I as package maintainer can influence this. In the past,
this only ever worked by contacting the admins of the build machines who
manually edited some control tables governing the timeout patience.
I *have* in fact contacted some admins (mipsel) repeatedly over this, but
never heard anything back,
In this case, the only solution, as far as I can see, is to restrict Quantlib
from not building on mips, mipsel, s390, ...
I will cc debian-project on this. Maybe someone has a better idea, and can
teach me what else to do here.
I have reduced the severity.
Dirk
| urgency of medium or higher, to help expedite the migration of boost and a
| number of other packages into Sarge.
|
| Thanks,
| Christopher Martin
--
Better to have an approximate answer to the right question than a precise
answer to the wrong question. -- John Tukey as quoted by John Chambers
Reply to: