[01:32] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) joined #pypy.
[01:45] jacob22|home (~jacob@c-51c6e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: "using sirc version 2.211+KSIRC/1.3.11"
----- silence for 2 hr and 35 minutes ----- [04:20] yuuh (lhlcpsl@i3ED6B5C2.versanet.de) left irc: "utz utz utz"
[04:34] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) left irc: "http://fredrikj.net";
----- silence for 2 hr and 45 minutes ----- [07:19] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 3 hr and 28 minutes ----- [10:47] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) left irc:
[10:49] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) joined #pypy.
----- silence for 47 minutes ----- [11:36] Action: arigo @ tunes.org -- disk full :-(
----- silence for 22 minutes ----- [11:58] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) left irc:
[11:59] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) joined #pypy.
----- silence for 45 minutes ----- [12:44] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Nick collision from services.
[12:44] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 59 minutes ----- [13:43] yuuh (xqacjlpv@dsl-62-220-10-27.berlikomm.net) joined #pypy.
[13:53] stakkars (epsvpe@dsl-62-220-10-27.berlikomm.net) joined #pypy.
----- silence for 5 hr and 7 minutes ----- [19:00] sanxiyn (tinuviel@222.233.187.44) joined #pypy.
----- silence for 1 hr and 14 minutes ----- [20:14] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) joined #pypy.
----- silence for 1 hr and 8 minutes ----- [21:22] <stakkars> arigo: are you there?
[21:22] <arigo> stakkars: hi, yes
[21:22] <stakkars> hi. I made quite some changes, compactified exceptions in simplify.
[21:23] <stakkars> I now get a problem with check_graph, but I don't understand:
[21:23] <stakkars> when checking link arguments, it tries to ensure that an link argument is not last_exc_xxx,
[21:23] <stakkars> if it is not an exception link. I seem to produce the situation now.
[21:24] <stakkars> see model.py, line 356
[21:24] <stakkars> 365
[21:24] <stakkars> is it plausibility, just because it couldnot happen, or reallywrong?
[21:26] <stakkars> (I might have done a bit too much of renaming)
[21:29] <arigo> stakkars: this check is really there to be enforced, yes
[21:29] <arigo> the graph doesn't really make sense with a last_exc_xxx outside an exception handler
[21:31] <arigo> basically, check_graph() is the most precise reference around about what the rules for a valid graph are
[21:32] <stakkars> well,I might have renamed a variable too much across a link I guess.
[21:33] <stakkars> maybe in the case where an exception handler passes stuff on via raise,
[21:34] <stakkars> but I don't really see why this is an error. Anyway if you want me to fix it I'll fix it.
[21:34] <arigo> Samuele besides me discusses how the last_exc_xxx model is rather broken, in some sense
[21:35] <arigo> he also got the same problem at various points
[21:35] <arigo> you think you can propagate the constant further down some links, but no luck, it's a last_exc_xxx
[21:36] <stakkars> I never may pass that variable over a link, right?
[21:37] <stakkars> unless it is an exception link. ah
[21:40] <arigo> it's a Constant, last_exc_xxx, that's part of the problem
[21:41] <arigo> a saner model would be to use a Variable instead, from the beginning;
[21:41] <arigo> Variables that are "created" in the block that ends in the exception handler
[21:42] <stakkars> how should I act right now? I like the compactification very much
[21:42] <arigo> I'm afraid that special-casing is the only way out, for now
[21:49] <arigo> we should probably try that fix, anyway
[21:50] <stakkars> turning it into a variable
[21:50] <arigo> Samuele and me face the a choice of trying that now or continuing on something unrelated
[21:51] <stakkars> let me know :-)
[21:51] <arigo> so we'll let you decide if you think that would help you or more likely break stuff and produce svn conflicts
[21:51] <arigo> :-)
[21:52] <stakkars> I don't mind. Conflicts are unlikely, I have onlytiny corrections.
[21:52] <arigo> the idea would be to replace them with fresh variables, newly introduced by each individual link itself
[21:53] <arigo> with new .last_exception and .last_exc_value attributes on the Links to know which are those new variables
[21:53] <stakkars> umpf! changing the link protocol?
[21:53] <arigo> well, it's part of the link protocol, these Constant(last_exc_xxx), at the moment, so yes :-)
[21:54] <arigo> the alternative would be to introduce the new variables in the previous block, but that has drawbacks, like not being able to use distinct variables for the disctinct exception instances if there are multiple handlers
[21:55] <stakkars> I'm confused. But carry on, my stuff works now modulo that problem,and I'll check in.
[21:56] <arigo> ok
[21:57] <arigo> then we'll come back to that later I guess... (should be a pypy-dev mail first, anyway)
[21:57] <arigo> (not to say that what else we're working on shouldn't be a pypy-dev mail. hum.)
[21:58] <stakkars> that is, I try to special case? ok
[22:04] <arigo> yes.
[22:04] <arigo> special case... in your code, I guess -- without changing check_graph(), that is :-)
[22:05] <arigo> changing check_graph would really break stuff left and right.
[22:05] <stakkars> booh hooh what do you think of me :-]
[22:06] <arigo> ain't me, was Samuele's idea :-))) (oups)
[22:07] <stakkars> say helloto Samuele,all fine
[22:08] <stakkars> (now writing some capturing code, because, however, the problem comes at the very end of pypymain)
[22:10] <stakkars> (found the hack <==)
[22:15] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) left irc:
[22:16] dialtone (~dialtone@host111-56.pool80117.interbusiness.it) joined #pypy.
[22:19] <stakkars> got it. ciao - chris
[22:20] <arigo> bye
----- silence for 32 minutes ----- [22:52] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "[x]chat"
[22:59] <stakkars> arigo? It is your code I guess :-)
[00:00] --- Mon May 9 2005