Log of the #nice channel on irc.freenode.net

Using timezone: Central European Time
* CIA-1 leaves01:00
* CIA-1 joins
* bonniot leaves02:19
* arjanb leaves02:59
* ArtemGr joins09:27
* arjanb joins12:00
* ArtemGr leaves13:25
* Artem joins
* bonniot joins14:20
hi
<Artem>salute14:21
<bonniot>i wonder if rfe 824450 (import leaks) is not already fixed by the work on visibility14:23
<Artem>sometimes i have a feeling it does (didn't checked on purpose), but having imports declared in a separate package descriptor might change this.14:26
<bonniot>technically, there is a distinction between "depend on this package" and "make the symbols of this package available without qualification"14:29
<Artem>if rfe 824450 is fixed, perhaps it should be closed?14:31
<bonniot>yes! i'm just not 100% sure
<Artem>how about automatic generation of dependencies, does it sounds good?
i mean in package descriptor.14:32
<bonniot>isn't package dependency a fundamental architecture decision? that would mean it's a good thing to declare it (and have the compiler enforce it)14:33
<Artem>i think enforcing package dependencies will break half of my program, i have a lot of inter-dependent packages... i'm not sure this isn't normal, btw, though clear dependency tree is alluring in some way and required for proper method dispatching in Nice...14:37
<bonniot>in what way are they not enforced at the moment?14:38
<Artem>"and required for proper method dispatching in Nice" - inter-dependent packages are ok when compiled together in one sweep, though.14:39
"in what way are they not enforced at the moment?" - i can use b from a, and a from b.
<bonniot>sorry, but that's another question14:41
I did not mention circular dependencies
just declaration (as now) VS automatic generation
<Artem>all right, let's say if package have a descriptor, then descriptor dependencies are enforced.14:46
"technically, there is a distinction between "depend on this package" and "make the symbols of this package available without qualification"" - they work in opposite ways? import enlarge dependencies and package declaration constrain them..14:47
<bonniot>i just mean that you could have the first not imply the second14:49
in java the first does not really exist, because the compilation granularity is the class14:50
the first one would belong the package descriptors14:51
if it's a problem to have "open" (the second one) automatically implied by the first, then it could be done individually in each file
what do you think of automatic opens (current behaviour)?14:52
(I think import foo.* is only meaningful for importing java classes, and it definitely should be doable on a per-file basis)14:54
although the possibility of package-global import foo.* could be useful too
<Artem>"the first one would belong the package descriptors" - though now it belongs to imports14:55
"what do you think of automatic opens (current behaviour)?" - classes and methods grouped in a source file usually have some common goal, it is very probable they will need a set of imports different from classes and methods grouped in another source file.
"I think import foo.* is only meaningful for importing java classes" - java libraries are to be mentioned in (perhaps a separate) dependencies descriptor too, with some way to automatically download them in the future...14:57
if we have a single dependencies format for both Java and Nice packages, we can provide means of downloading Java jars thru Nice repository.14:58
<bonniot>it's certainly a goal ;-)14:59
unfortunately I'll have to leave...
<Artem>request to the repository: i need package a.b.c; response: either the package, or a list of download locations.
i will be out on vocation for two weeks, starting from today, without internet. FYI.15:00
<bonniot>we should discuss it further later
ok. I'll be on the road too, although with connection intermitently
<Artem>salute
<bonniot>so let's let this mature ;-)
have a good time!
<Artem>thanks15:01
<bonniot>bye to all!
* bonniot leaves
* CIA-1 leaves16:59
* CIA-1 joins
* CIA-1 leaves17:00
* CIA-1 joins
* Artem leaves18:49
* arjanb leaves20:52
* CIA-1 leaves21:54
* CIA-1 joins21:55
* CIA-1 leaves23:35
* CIA-1 joins
* CIA-1 leaves23:36
* CIA-1 joins
* CIA-1 leaves23:37
* CIA-1 joins

Generated by Sualtam