<GitHub92>
artiq/master c94dafd Yann Sionneau: Revert "travis: use more recent libstdc++"...
<sb0>
ysionneau, why revert?
<ysionneau>
to make the "install instructions" easier
<ysionneau>
one less step for install
<sb0>
so you are building the package with the older libstdc++ now?
<ysionneau>
yes
<ysionneau>
I use an Ubuntu 14.04 VM instead of a Debian Jessie VM ... it's not so different for me
<ysionneau>
if it can make some early adopter life easier ...
<sb0>
how does conda deal with this sort of issue?
<sb0>
do they link everything statically? ship their own libs? assume certain versions?
<ysionneau>
I think conda does not deal with this kind of "system library compatibility" issue
<sb0>
there's no way around it
<ysionneau>
it links dynamically
<ysionneau>
except if something in build.sh tells the build system otherwise
<ysionneau>
in other words : conda does not care
<sb0>
i see that anaconda ships with qmake, that uses libstdc++.so.6
<sb0>
(the system one)
<ysionneau>
would be interesting to check which version it is linked against
cr1901_modern has joined #m-labs
* ysionneau
is having a look
cr1901_modern has quit [Ping timeout: 264 seconds]
<sb0>
rjo, does it make sense to keep the parameter and real-time result databases separated?
<ysionneau>
17:56 < sb0> do they link everything statically? ship their own libs? assume certain versions? < usually conda ships the libraries when there is a dependency, except for stuff like libc libstdc++ and maybe a few others
cr1901_modern has joined #m-labs
<ysionneau>
one need to pay attention while packaging conda stuff not to link with "host" libraries (like openssl or zlib) ... it's easy to miss those situations since it will obviously work on the developer's machine
cr1901_modern has quit [Write error: Connection reset by peer]
cr1901_modern has joined #m-labs
cr1901_modern1 has joined #m-labs
cr1901_modern1 has quit [Write error: Connection reset by peer]
cr1901_modern has quit [Ping timeout: 265 seconds]
cr1901_modern has joined #m-labs
travis-ci has joined #m-labs
<travis-ci>
m-labs/artiq#412 (master - cd199f1 : Sebastien Bourdeauducq): The build has errored.
<ysionneau>
sb0: their qmake seems to be linked against an "old" libstdc++ , it's only using GLIBCXX_3.4
cr1901_modern has quit [Ping timeout: 250 seconds]
cr1901_modern has joined #m-labs
cr1901_modern has quit [Write error: Connection reset by peer]
bentley` has joined #m-labs
<rjo>
sb0: ddb, experiments, and pdb are considered "inputs". results (realtime or not) are outputs. if the pdb gets saved, having unrelated results in there is at least expensive and confusing.
<rjo>
but if you want to merge the two, you could add a type to each entry.
<rjo>
but other than the save handling, pdb and rdb are not different AFAICT.
bhamilton has joined #m-labs
mumptai has joined #m-labs
mumptai has quit [Remote host closed the connection]
mumptai has joined #m-labs
mumptai has quit [Remote host closed the connection]
bhamilton has quit [Remote host closed the connection]
mumptai has joined #m-labs
cr1901_modern has joined #m-labs
cr1901_modern has quit [Ping timeout: 260 seconds]
<GitHub107>
[artiq] fallen pushed 1 new commit to master: http://git.io/vseL1
<GitHub107>
artiq/master 19a77b5 Yann Sionneau: conda: make aiohttp testing not depend on gunicorn on Windows platform
travis-ci has joined #m-labs
<travis-ci>
m-labs/artiq#414 (master - 19a77b5 : Yann Sionneau): The build has errored.