<pnou>
au fait, maxence m'a dit que ocamlsql n'était pas utilisé, non documenté, et que ça ne vaut peut-être pas le coup de le packager
<two-face>
ah bon ?
<two-face>
ok
<pnou>
oui, apparemment c'est un vieux reste d'une ancienne version de dbforge
<two-face>
donc je l'enlève carrément ?
<two-face>
pnou: tu peux rajouter la production du manuel en html s'il te plait ?
<pnou>
je vais voir ce que je peux faire
<two-face>
pnou: il faut donc aussi que le make install n'installe pas ocamlsql
<pnou>
je m'en occupe
<two-face>
merci
<two-face>
a+
two-face has left #ocaml []
gl has quit ["do you k"]
MegaWatS has joined #ocaml
<MegaWatS>
hi
gl has joined #ocaml
two-face has joined #ocaml
TachYon25 has joined #ocaml
* gl
is away: I'm busy
smklsmkl has joined #ocaml
smkl has quit [Read error: 110 (Connection timed out)]
xmkl has joined #ocaml
smklsmkl has quit [Read error: 110 (Connection timed out)]
skylan has joined #ocaml
owll has joined #ocaml
Malkuth has left #ocaml []
<MegaWatS>
hmmm
<MegaWatS>
does anybody know whether there is any problem with my programm calling (froim caml) a c function which then, in turn, calls back a caml function?
<MegaWatS>
which was given to it as a parameter
<two-face>
hey MegaWatS
clog has joined #ocaml
electron__ has joined #ocaml
electron__ has left #ocaml []
<two-face>
pnou: je vois pas trop pourquoi on compte les versions bytecode et natif quand le natif est possible
TachYon25 has quit [Read error: 110 (Connection timed out)]
<pnou>
on compte ?
<two-face>
arf
<two-face>
on produit
<pnou>
parce que c'est plus simple je suppose
<two-face>
moi j'ai envie de ne mettre que l'opt s'il existe et de le renommer en binaire normal
<pnou>
c'est sûr que c'est plus mieux
<two-face>
je vais mettre des tests
<pnou>
ça serait bien d'uniformiser ça au niveau de debian
<two-face>
euhh
<two-face>
je ne sais pas si c'est du ressort de debian
<pnou>
bah ça serait bien que sur une architecture supporté en natif on est que des binaires natifs
<pnou>
dans les packages
<two-face>
oui mais les tests pourraient être effectués dans les makefiles de cameleon
<two-face>
avec un if
TachYon25 has joined #ocaml
<pnou>
mouais
<two-face>
ocamlsdl fait ça :)
<pnou>
oué c'est vrai que ça devrait être fait
<pnou>
en fait faudrait un makefile plus ou moins standard qui s'en occupe
<two-face>
par exemple
gl has quit [No route to host]
<two-face>
pnou: en attendant, je vais gérer ça moi-même
<two-face>
bon allez, vivement que ça soit fini :p
<pnou>
tu vois le bout du tunnel :)
<two-face>
:)
<two-face>
on sera plusieurs à s'occuper de ça après
<two-face>
quoique je doute qu'il y ait encore du boulot
<pnou>
c'est le bon plan d'être mainteneur après que t'aies fait tout le boulot :)
<two-face>
arf
<two-face>
bon, allez je vais lire
<two-face>
a+
two-face has left #ocaml []
gene9 has joined #ocaml
gene9 has quit [Client Quit]
TachYon25 has quit ["bez ki³y nie ma zaliczenia (z prawd studentek AM)"]
Yurik has joined #ocaml
<Yurik>
re
<Yurik>
anybody used bdb for the latest versions of berkeley db?
<Yurik>
oh. seems that all guys are sleeping, it's too late today
<Yurik>
:)
<MegaWatS>
I'm not sleeping
<MegaWatS>
but, no, I'm not, so I can't help you :)
<Yurik>
:)
* Yurik
is writing new wrapper
<Yurik>
to fit all functionality and latest release of bdb
<MegaWatS>
I see
<MegaWatS>
I'm currently not having much luck with my own project :[
<MegaWatS>
I wrote a wrapper for libjpeg, because the one that already exists (in camlimages) does not allow a different data source than a file
<MegaWatS>
but it produces crashes :|
<Yurik>
I'm thinking to use SWIG, however now i have a bit of code written manually
<Yurik>
and it crashes at a werid place, too :)
<MegaWatS>
and now I have already been searching for the error for 3 days straight and still no luck :|
<Yurik>
s/werid/weird/
<MegaWatS>
yeah
<MegaWatS>
it produces some kind of lingering error
<MegaWatS>
and then suddenly it crashes in the most strange places
<Yurik>
even gdb says that SIGSEGV was in ?? :))
<Yurik>
writing wrappers sucks. but I need it :-(
<MegaWatS>
I have found by now that the crash can be reproduced the most consistently when I also use arrays of weak pointers in the same program as well, and a caml string as a data source
<MegaWatS>
but neither of those circumstances seem to really be necessary for the crash
<MegaWatS>
and the annoying thingh is, the wrapper is actually kinda small and I've already gone through the code X times with an ultra fine comb and it all seems to be as it should :[
<MegaWatS>
in the beginning, I even had done some more dangerous or experimental things, because they would have been easier/smaller/faster, but by now I have refined the code to be ultra-rigoprously compliant to the specs in the caml manual's "interfacing with C code" chapter
<MegaWatS>
but nothing I can do seems to make any difference whatsoever :|
<MegaWatS>
</whine> :p
* Yurik
is back
<Yurik>
may be SWIG will save me from being tired making wrapper for bdb manually :-)
* Yurik
is going to go offline (and cries: why these cigarettes are finishing only at the night, when it is boring to go to buy them ?! ;))