IRC log of fenfire on 2005-03-27

Timestamps are in UTC.

02:28:39 [rubberpaw]
rubberpaw has quit
02:40:19 [rubberpaw]
rubberpaw has joined #fenfire
05:07:46 [antont]
antont has quit
05:16:40 [rubberpaw]
rubberpaw has quit
06:40:31 [rubberpaw]
rubberpaw has joined #fenfire
10:04:15 [tuukkah]
ibid, did you look at the slowness of fenfeed?
10:05:09 [ibid]
10:05:17 [ibid]
(not yet, but again, no promises:)
10:08:45 [tuukkah]
I tried to look into it a bit, but I can't reproduce the slowness ;-)
10:09:18 [ibid]
10:09:36 [tuukkah]
how slow is the startup on your system?
10:09:58 [ibid]
it doesn't pause for minutes for you before showing the ui components?
10:10:41 [tuukkah]
no, with the default feeds I can read an item and quit under 20 seconds
10:10:53 [ibid]
with kaffe?
10:10:56 [tuukkah]
10:11:04 [ibid]
what kind of a system?
10:11:13 [tuukkah]
try time make run_feed FEEDS="" JAVA=kaffe
10:11:22 [tuukkah]
and quit when you see the ui components
10:11:31 [tuukkah]
oh, and something else...
10:11:44 [tuukkah]
you should remove your subscriptions too
10:11:58 [tuukkah]
edit fenfeed_conf.turtle
10:12:45 [tuukkah]
there's a list of subscriptions. you can change the property name or move the file away
10:13:10 [tuukkah]
10:14:01 [tuukkah]
time says 4.0 s real, 2.7s user
10:14:17 [tuukkah]
with a Pentium M running at 600 MHz
10:15:11 [ibid]
it's fast enough without feed
10:15:14 [ibid]
10:18:18 [tuukkah]
perhaps I can't help then
10:19:19 [ibid]
real 0m37.359s
10:19:22 [ibid]
without feeds
10:19:31 [ibid]
real 4m3.423s
10:19:34 [ibid]
with standard feeds
10:20:37 [ibid]
The requested URL /2004/12/fenfeed was not found on this server.
10:20:54 [tuukkah]
for me, it takes about 4-5 times longer with feeds too
10:21:48 [rubberpaw]
rubberpaw has quit
10:22:01 [tuukkah]
oh, that uri was just an identifier. but the server serve documentation to GET requests
10:22:12 [ibid]
model name : Pentium III (Coppermine)
10:22:19 [tuukkah]
could serve
10:22:22 [ibid]
cpu MHz : 1000.188
10:22:59 [ibid]
cache size : 256 KB
10:23:10 [ibid]
MemTotal: 256616 kB
10:23:10 [tuukkah]
I can't believe that would be so much slower
10:23:52 [tuukkah]
is it swapping or could it be doing a lot more gc?
10:25:46 [ibid]
no audible paging
10:26:37 [tuukkah]
on my older machine, with feeds the ui shows up in 30 s real
10:27:18 [tuukkah]
this is AMD-K6 300 MHz
10:27:19 [ibid]
it does some gc before the window shows up but none afterward
10:27:43 [ibid]
until the ui compnents show
10:28:07 [tuukkah]
so gc can't be the reason?
10:28:28 [ibid]
10:28:40 [tuukkah]
you mean it can be?
10:28:59 [ibid]
now it works fast enough
10:29:05 [ibid]
i have no idea what changed
10:29:30 [tuukkah]
DNS issues comes to mind
10:29:48 [tuukkah]
and you can try to rm -R _httpcache
10:30:42 [ibid]
after that, it's slow again
10:30:51 [ibid]
actually not
10:30:55 [ibid]
i just thought so
10:31:08 [ibid]
Adding to graph:
10:31:08 [ibid]
nu.xom.xslt.XSLException: Syntax error in stylesheet
10:31:16 [ibid]
Caused by: javax.xml.transform.TransformerConfigurationException: File "" not found.
10:31:27 [ibid]
Caused by: javax.xml.transform.TransformerException: File "" not found.
10:31:35 [ibid]
these did not come when it was slow
10:31:42 [ibid]
Caused by: org.xml.sax.SAXParseException: File "" not found.
10:31:52 [ibid]
Error while reading
10:31:52 [ibid] nu.xom.xslt.XSLException: Syntax error in stylesheet
10:33:42 [tuukkah]
these are normal on kaffe and mean that we have loaded the feed but the XSLT libs are broken and can't turn the feed into rdf
10:34:31 [ibid]
i don't know what happened, but it's fast now
10:35:15 [tuukkah]
so now nobody can reproduce it :-(
10:36:16 [ibid]
well, reasonably fast :)
10:37:49 [ibid]
well, that does corraborate the theory that it's a threading problem
10:38:37 [ibid]
those tend to be nondeterministic
10:40:36 [tuukkah]
10:40:45 [rubberpaw]
rubberpaw has joined #fenfire
10:41:23 [ibid]
i noticed
10:41:49 [tuukkah]
I had to look it up in the dictionary
10:42:05 [ibid]
* ibid too
10:42:15 [ibid]
well, to check the spelling
10:42:25 [ibid]
did that too late, though
10:43:21 [tuukkah]
I think the ui components shouldn't show before the list of subscribed feeds is loaded
10:43:52 [tuukkah]
but the list should load way faster
10:43:52 [ibid]
i find that very counterintuitive
10:44:09 [ibid]
if i see a blank window, i looks like a bug
10:44:19 [ibid]
it, to me
10:44:49 [tuukkah]
to me it looks slow :-)
10:45:13 [tuukkah]
well, it could be better to show the components with a text "loading..."
10:45:25 [ibid]
it'd look less bad if there was just a splash window
10:45:32 [ibid]
or "loading":)
10:46:43 [tuukkah]
do you think the splash window or text would go away in case of a bug ?-)
10:46:56 [ibid]
what do you mean?
10:47:32 [ibid]
though i dislike programs that take a long time to load
10:47:49 [ibid]
but that's a bigger problem than just fenfire or fenfeed:)
10:47:52 [tuukkah]
I'm trying to think how blank white window is worse to you than the text "loading..."
10:48:43 [tuukkah]
I suppose progress bar would be ideal
10:49:14 [tuukkah]
or next to ideal, if instant loading is ideal
10:49:18 [ibid]
it's a perceptional issue
10:50:22 [tuukkah]
we just have to find the code that blanks the window and change that :-)
10:51:18 [ibid]
a blank window usually is a sign of a bug in the program's handling of events ==> empty window looks buggy
10:51:39 [tuukkah]
but it's not empty, it's blank white !-)
10:51:51 [ibid]
10:57:26 [tuukkah]
perhaps org/nongnu/libvob/impl/awt/
11:08:15 [tuukkah]
yes, changing line 382 changes the startup
11:10:16 [ibid]
publish it?
11:10:44 [tuukkah]
well, changing the color from white to something else isn't worth publishing, is it?-)
11:11:11 [ibid]
11:12:07 [tuukkah]
hmm, the code that is used to draw error strings crashes the app
11:15:54 [tuukkah]
oh, now it worked - once
11:22:20 [tuukkah]
oh, nice :-) javadoc says the given scale is divided by 1000, but actually it was used as is
11:22:41 [tuukkah]
so the code tries to draw 18 point font as 18000 pixels high
11:23:58 [ibid]
a bug in kaffe?
11:24:18 [tuukkah]
no, I think this is a bug in javadoc
11:24:22 [ibid]
11:24:40 [tuukkah]
I mean, in our documentation
11:24:56 [ibid]
ah, your own routine, not a standard one :)
11:25:02 [tuukkah]
11:35:18 [tuukkah]
I think it looks nice now. I'll commit
11:38:36 [tuukkah]
12:03:15 [ibid]
better :)
12:03:46 [ibid]
and not better
12:03:51 [ibid]
it doesn't redraw
12:03:51 [tuukkah]
12:03:54 [ibid]
on expose
12:05:01 [ibid]
hm, something's not right
12:05:16 [ibid]
it has already loaded the feeds but it doesn't draw the window
12:05:43 [tuukkah]
perhaps you start to find the thread bug :-)
12:05:50 [ibid]
12:06:25 [ibid]
it works now
12:06:32 [tuukkah]
I added the text to same place where the blanking is done. without redraw, the canvas would be grey at least on kaffe
12:06:34 [ibid]
clearly a threading bug
12:07:03 [ibid]
it'll need a code review
12:07:17 [ibid]
yeah, it was gray
12:08:38 [tuukkah]
hmm, is cpu usage still at 100%?
12:09:01 [tuukkah]
because it might just be that a task is using the AWT thread
12:09:16 [tuukkah]
and the task takes a long time to finish
12:09:48 [ibid]
can't check
12:13:46 [tuukkah]
perhaps you can enable org.nongnu.libvob.AbstractUpdateManager.dbg to get some printout when it happens next time
12:14:14 [ibid]
how do i do that?
12:15:20 [tuukkah]
I think you need to change that attribute in code from false to true
12:15:57 [ibid]
in gzigzag there was a runtime way
12:16:03 [ibid]
can't remember what, though
12:16:35 [tuukkah]
yeah, but we always write a new main method that doesn't implement the old options =-/
12:16:55 [benja_]
12:16:57 [ibid]
12:16:59 [ibid]
12:17:03 [tuukkah]
huomenta benja_
12:17:08 [benja_]
will be here for 45 minutes
12:17:19 [tuukkah]
benja_, there's a lot of cruft in the repos =)
12:17:22 [benja_]
hope to implement at least Lob.add()
12:17:40 [benja_]
tuukkah: true -- except for the smiley :-(
12:17:48 [benja_]
should be moved to basalt/
12:18:09 [tuukkah]
and duplicates removed altogether
12:18:12 [benja_]
I recall though that once after some cruft was removed you were unhappy because you had spend a lot of time working on it :)
12:18:22 [benja_]
tuukkah: but only after checking they really are
12:18:37 [benja_]
there are some equally-named files in libvob.vobs and libvob.vobs.lava
12:18:40 [benja_]
which are different
12:18:42 [benja_]
12:19:09 [tuukkah]
yes, there's two URN5NameSpaces that are slightly different
12:19:11 [benja_]
I'd also be happy to delete stuff if I know the functionality has been duplicated in newer code
12:19:15 [benja_]
tuukkah: :-(
12:20:18 [tuukkah]
there should be a tool that looks for duplicate or unused code
12:20:40 [rubberpaw]
rubberpaw has quit
12:26:57 [tuukkah]
actually, some java compilers can start from main class and build only needed classes
12:27:18 [tuukkah]
so if some java file isn't compiled then it's waste
12:36:06 [tuukkah]
private/org/fenfire/fentwine/util/ is identic to fenfire/org/fenfire/util/ so can be removed?
12:36:28 [tuukkah]
but storm/org/nongnu/storm/util/ has minor differences that should be merged
12:36:52 [tuukkah]
benja_, fenfire code can depend on storm classes, right?
12:37:20 [benja_]
tuukkah: yes
12:37:34 [benja_]
fentwine pair can be removed, too
12:59:48 [tuukkah]
ok, done
13:02:53 [tuukkah]
the differences in URN5Namespace are that fenfire has an underscore and storm doesn't have in return "urn:urn-5:" + namespace + ":_" + num;
13:04:15 [tuukkah]
how to merge this?
13:04:44 [benja_]
tuukkah: use the one with the underscore
13:04:56 [benja_]
necessary for some rdf serializations
13:05:00 [benja_]
going now, cu
13:05:16 [tuukkah]
see you
13:24:59 [rubberpaw]
rubberpaw has joined #fenfire
14:10:16 [Gwl]
tän TIE363:n varmaan saa sisällytettyä ohjelmistotekniikan linjan maisteriopintoihin?
14:18:24 [ibid]
yes, that's the intent
14:18:47 [ibid]
(i recommend english on this channel, even though the majority present understands finnish, too:)
14:18:49 [rubberpaw]
rubberpaw has quit
14:19:26 [Gwl]
ok, good.. and yeah, my question just was a bit hard to ask in english so decided to not even try :)
14:19:31 [ibid]
Gwl: of course, it cannot supplant any of the mandatory courses
14:19:33 [ibid]
14:31:09 [tuukkah]
Gwl, yeah, no problem
15:25:08 [rubberpaw]
rubberpaw has joined #fenfire
17:15:20 [tuukkah]
my email addr makes nice foaf:mbox_sha1sum =)
17:18:09 [ibid]
17:20:28 [tuukkah]
starts "bdb01db1b"
17:22:52 [tuukkah]
ibid, "Becoming a member in The Fenfire Foundation" should be shorter as it shows in navbar, browser window title etc.
17:23:29 [rubberpaw]
rubberpaw has quit
17:24:50 [tuukkah]
a related idea would be to move it to a subdirectory and limit navbar to */*.rst
17:25:15 [rubberpaw]
rubberpaw has joined #fenfire
17:25:25 [tuukkah]
maxring = 0;-)
17:27:31 [tuukkah]
"Affiliate" would be one word
17:27:41 [ibid]
tuukkah: fix it? :)
17:27:53 [tuukkah]
sure, I'm just not sure what to do
17:54:38 [tuukkah]
argh, the directories don't have g+s anymore
17:56:11 [ibid]
well, i did not anticipate the navbar thing when i first wrote it
17:56:23 [ibid]
browser window title is not a problem imho
17:56:39 [tuukkah]
yes, the titles are often long anyway
17:57:04 [ibid]
yes as in you agree or yes it is a problem? :)
17:57:13 [tuukkah]
17:57:28 [ibid]
17:58:22 [tuukkah]
benja_, majukati, please do chgrp -R himalia . in and
17:59:07 [ibid]
both or either?
17:59:12 [tuukkah]
17:59:22 [ibid]
just making sure :)
18:00:10 [tuukkah]
benja_ or majukati, and please tell the administration to put g+s on both directories
18:00:25 [ibid]
they can't do it themselves?
18:00:39 [tuukkah]
the dirs are now owned by root
18:00:43 [ibid]
18:01:24 [ibid]
* ibid is thinking doing a text editor as my project on the course :)
18:01:31 [ibid]
(with xanalogical text, of course)
18:03:11 [rubberpaw]
rubberpaw has quit
18:03:25 [tuukkah]
specifically, only benja can put changes into into effect
18:04:08 [ibid]
18:05:19 [tuukkah]
well, I was able to change the navbar on the other pages :-)
18:06:23 [ibid]
hmm, i'm not sure "affiliating" conveys the right message
18:06:38 [tuukkah]
might be
18:07:55 [tuukkah]
"Membership" might be better
18:08:06 [ibid]
not sure about that
18:08:10 [ibid]
let me ask somebody :)
18:08:17 [ibid]
leave it as it is currently for now, please
18:08:28 [tuukkah]
18:08:46 [tuukkah]
i was looking at :-)
18:14:57 [ibid]
doesn't seem like people are keen on helping
18:19:42 [tuukkah]
helping with what?
18:19:54 [ibid]
21:08 <ibid> let me ask somebody :)
18:20:02 [tuukkah]
18:20:23 [tuukkah]
I was already checking out doap stuff
18:22:32 [ibid]
21:22 <Heffalump> I would have thought the more appropriate word would have been "Joining"
18:22:58 [ibid]
21:22 <Heffalump> and you become a member _of_ things, not _in_ things
18:23:17 [tuukkah]
probably the only software that supports is Fenfire :-)
18:23:31 [ibid]
18:26:12 [tuukkah]
18:26:18 [tuukkah] Unhandled content type application/x-turtle when reading
18:26:18 [ibid]
18:27:01 [tuukkah]
which is to say that even Fenfire doesn't support browsing the url
18:38:13 [tuukkah]
hmm, the code supports application/turtle
19:25:26 [rubberpaw]
rubberpaw has joined #fenfire
20:48:26 [benja_]
20:48:41 [tuukkah]
hi benja_
20:48:41 [benja_]
tuukkah: ok, code is wrong
20:48:50 [tuukkah]
which code?
20:49:08 [benja_]
the code supporting application/turtle
20:49:17 [tuukkah]
I already committed fix
20:49:42 [benja_]
20:50:13 [tuukkah]
can you chgrp -R himalia . at and himalia?
20:50:27 [benja_]
-rw-rw-r-- 1 hastrup hastrup 3221 Mar 27 21:02 vision.html
20:50:49 [tuukkah]
hmm, when did that happen?
20:51:52 [tuukkah]
we really need to get g+s back for the root dirs
20:52:47 [benja_]
I like the "Loading..." thing, was that you?
20:52:59 [ibid]
my request, i want the credit :)
20:53:02 [tuukkah]
yes, by ibid's idea :-)
20:53:04 [benja_]
20:54:23 [tuukkah]
now I was thinking why was moved to lava and if it should be moved back
20:54:39 [benja_]
20:55:18 [tuukkah]
because I'd like to start fentwine in the most connected node of the largest component along sl:linkedTo
20:55:42 [tuukkah]
I really hate when it starts in rdfs:label or some such =-/
21:00:40 [benja_]
I haven't moved it there. Are you sure it didn't start in lava/?
21:01:43 [tuukkah]
oh right, why not
21:01:58 [tuukkah]
can it be used from there?
21:02:11 [benja_]
it's not supposed to be
21:02:42 [benja_]
it's not lava any longer if stable code calls it
21:05:31 [tuukkah]
21:05:46 [rubberpaw]
rubberpaw has quit
21:10:02 [tuukkah]
benja_, chmod g+s
21:11:17 [benja_]
21:11:38 [tuukkah]
21:13:00 [ibid]
might be a good idea to fix the affiliating
21:14:11 [benja_]
it would be nice if the navbars only showed the current part of the tree
21:14:46 [tuukkah]
benja_, I'd prefer some values for maxring and maxdepth ;-)
21:14:56 [benja_]
21:15:23 [tuukkah]
show some amount of surrounding branches too
21:16:33 [benja_]
I don't understand anything of that
21:16:53 [benja_]
oh, you're refering to the webpage?
21:17:04 [tuukkah]
yes %-)
21:17:08 [benja_]
that sounds bad :)
21:17:24 [benja_]
I mean, hard to understand for the reader
21:17:58 [tuukkah]
I like that */*.rst are visible
21:18:18 [benja_]
they used to be linked, too
21:18:26 [benja_]
not any longer, for some reason
21:18:52 [benja_]
(from the footer)
21:19:41 [ibid]
hm, somebody should do something to the course page english
21:19:45 [tuukkah]
grr :-) I mean */*.html
21:20:02 [ibid]
"Manner of Performance", right
21:20:46 [tuukkah]
I bet you can send patches, but I think otherwise it's not a priority
21:21:23 [ibid]
yeah right
21:21:53 [benja_]
argh, someone updated fenmm to use lobs, which is nice except now I have to update it to the new lob system :-o
21:22:18 [tuukkah]
21:22:26 [benja_]
mudyc apparently
21:22:35 [tuukkah]
perhaps he can do it?
21:22:48 [benja_]
then I have to move it to basalt/ for now...
21:23:05 [tuukkah]
this kind of issues were why I was wondering how Google does it
21:23:35 [ibid]
what's fenmm?
21:23:45 [benja_]
asko's mind map thing
21:23:50 [tuukkah]
perhaps fenmm should be a repo of it's own and it would state which version of fenfire it depends on
21:24:00 [ibid]
21:24:03 [tuukkah]
21:24:15 [ibid]
for that, you'd need to have versions of ff :)
21:24:19 [ibid]
and for that, releases
21:29:13 [tuukkah]
yes. we can do releases if we want to
21:29:37 [benja_]
more like if we have something useful to release
21:29:55 [tuukkah]
benja_, that's not the point. think of versions instead of releases
21:30:50 [tuukkah]
to some extent, we already have versions. we would just need to tell in each subproject, which versions of other subprojects it requires
21:31:58 [tuukkah]
then we'd write a leet darcs-using bash script that darcs gets the right patches and builds the subprojects
21:32:54 [benja_]
tuukkah: for fixing a version, it wouldn't make sense to make a release
21:36:51 [rubberpaw]
rubberpaw has joined #fenfire
21:37:30 [tuukkah]
benja_, why wouldn't it?
21:38:01 [benja_]
because a release should be something someone interested in the project would download and try out
21:38:12 [benja_]
else, they won't stay interested for long
21:38:19 [tuukkah]
it would simplify things a lot for those who don't have bash and darcs and for the writer of that script
21:39:16 [tuukkah]
benja_, you don't have to advertise it as a release, just say that this is what we have today and put the jar somewhere. some people could call it an alpha release
21:39:51 [benja_]
mhm, you could do that
21:52:31 [tuukkah]
that was the point :-)
21:52:53 [tuukkah]
it could also be less frightening for the students :-)
22:02:17 [ibid]
you can just use darcs tag if versions are all that are necessary
22:06:24 [benja_]
huh, after loads of commenting out, ff compiles again =)
22:07:54 [tuukkah]
ibid, perhaps. is a tag simply a set of pathces?
22:08:28 [ibid]
i don't know how to describe it in darcs terms
22:08:47 [ibid]
but the effect is that it allows you to go back to a specific configuration
22:09:58 [rubberpaw]
rubberpaw has quit
22:10:45 [ibid]
"Tag is used to name a version of the tree. Tag differs from record in that it doesn't record any new changes, and it always depends on all patches residing in the repository when it is tagged. This means that one can later reproduce this version of the repository"
22:15:09 [ibid]
22:17:15 [tuukkah]
I suppose if I want to tag I should use urn-5 for the name ;-)
22:18:39 [ibid]
22:18:52 [tuukkah]
22:19:24 [ibid]
well, you must decide if you really need a truly distributed identity assignment authority ;)
22:21:19 [tuukkah]
identity? I though you argued that they are identifiers and not identities?
22:21:35 [ibid]
i intended to write identifier
22:21:46 [ibid]
i suppose thats a leibnizian slip :)
23:36:59 [rubberpaw]
rubberpaw has joined #fenfire
23:45:10 [tuukkah]
java file line counts after each patch on darcs repositories:
23:46:30 [benja_]
23:46:44 [tuukkah]
seconds from 1970 =)
23:46:51 [benja_]
bah :)
23:47:30 [benja_]
interesting graph
23:49:42 [tuukkah]
ibid gave the link to a page that described darcs trackdown and gave ideas about how to use it
23:50:41 [benja_]
would be nice to cull stuff in basalt/
23:50:55 [benja_]
I've just moved a lot there (not recorded yet)
23:51:04 [benja_]
(the old lob system)
23:51:25 [tuukkah]
I used wc -l $(find org -name "*.java")
23:51:31 [benja_]
23:51:43 [benja_]
that's out then :)
23:51:51 [tuukkah]
23:52:04 [benja_]
not a problem
23:52:35 [tuukkah]
otherwise, you could call it a problem:
23:52:38 [tuukkah]
darcs trackdown 'LANG=C wc -l $(find org -name "*.java") && false' | egrep -v "^Trying|java" | egrep "^[SMTWTFS]|^ *[0-9]+ +total" | sed -n 's/\([0-9]\) .*$/\1/ ; T ; N ; s/\n *\([0-9]\+\) *total$/\t\1/ ; p' | while IFS=$'\t' read date lines ; do echo -e "$(date -d "$date" +%s)\t$lines" ; done >
23:52:45 [benja_]
23:53:17 [benja_]
how about making that part a problem too by using *.rj instead of *.java where both exist? :)
23:54:11 [tuukkah]
actually, are those *.java in the repository?
23:55:52 [benja_]
23:55:56 [benja_]
perhaps that's stupid
23:56:12 [benja_]
running python being easier than compiling C++
23:56:27 [tuukkah]
at least jython
23:56:39 [benja_]
23:56:50 [benja_]
23:57:04 [tuukkah]
oh, that problem
23:57:32 [benja_]
if you want to remove the java ones, go ahead
23:58:14 [tuukkah]
I'm not sure if I want to. and libvob have taught me to have generated files in the repo
23:58:40 [benja_]
23:58:54 [benja_]
well, really is a different case
23:59:27 [tuukkah]
but perhaps the shell command should be rewritten as some proper script
23:59:31 [benja_]
I have something on the screen again! :-)