[Ecls-list] Current status

Juan Jose Garcia-Ripoll juanjose.garciaripoll at googlemail.com
Sun Nov 7 12:09:32 UTC 2010


* I no longer have the accounts in which I performed automated builds of
Windows, FreeBSD, NedBSD and OpenBSD. It will take some time to find a
replacement which is not my own laptop.

* I test Linux/64-bits locally on two machines. Currently one of them is
showing errors in the pthread library. More precisely in the read/write
locks, which a completely different machine is not experiencing. I am really
puzzled about this.

* Solaris is being tested on the computers of the Sage project.
Solaris/Sparc builds fine, but shows up some floating point overflow in the
test phase, while Solaris/Intel simply does not build. It aborts with an
internal error. Simply weird, because it was working fine one day ago.

* To debug those "internal" or "boot time" errors I have added a function in
ECL called EXT:DUMP-C-BACKTRACE which works either using the backtrace()
function (FreeBSD, OS X, etc) or using GCC's __builtin_return_address()
combined with dladdr().

* OS X is also being regularly tested. Here the problems are obscure and I
can get perfect builds in my own terminal but the batch job is behaving
erratically.

Overall default builds seem to be working fine, but there are still corners
to be rounded and I need to get testing back online again.

Juanjo

-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/ecl-devel/attachments/20101107/ff8ad2f1/attachment.html>


More information about the ecl-devel mailing list