| Time |
S |
Nick |
Message |
| 00:09 |
|
|
abuiles left #darcs |
| 00:09 |
|
sm |
ack |
| 00:23 |
|
|
abuiles joined #darcs |
| 00:31 |
|
|
twb joined #darcs |
| 01:40 |
|
|
copumpkin left #darcs |
| 01:44 |
|
|
secorp left #darcs |
| 02:03 |
|
|
copumpkin joined #darcs |
| 02:07 |
|
|
preflex left #darcs |
| 02:15 |
|
|
FunctorSalad left #darcs |
| 02:19 |
|
|
preflex joined #darcs |
| 02:21 |
|
|
FunctorSalad joined #darcs |
| 02:40 |
|
|
abuiles left #darcs |
| 03:06 |
|
|
intripoon joined #darcs |
| 03:09 |
|
|
JaffaCake1 left #darcs |
| 03:09 |
|
|
intripoon_ left #darcs |
| 03:10 |
|
|
JaffaCake joined #darcs |
| 03:13 |
|
|
owst left #darcs |
| 03:16 |
|
|
JaffaCake left #darcs |
| 03:55 |
|
|
twb left #darcs |
| 04:09 |
|
|
sm left #darcs |
| 04:10 |
|
|
secorp joined #darcs |
| 04:14 |
|
|
twb joined #darcs |
| 04:15 |
|
|
gwern left #darcs |
| 04:28 |
|
|
gwern joined #darcs |
| 04:28 |
|
|
gwern left #darcs |
| 04:28 |
|
|
gwern joined #darcs |
| 04:46 |
|
|
lispy left #darcs |
| 04:56 |
|
|
secorp left #darcs |
| 05:21 |
|
|
dankna joined #darcs |
| 07:05 |
|
|
dreixel left #darcs |
| 07:27 |
|
|
exlevan left #darcs |
| 07:28 |
|
|
exlevan joined #darcs |
| 07:56 |
|
|
shenshei left #darcs |
| 07:59 |
|
|
raichoo joined #darcs |
| 08:01 |
|
|
darcscommitbot left #darcs |
| 08:02 |
|
|
darcswikibot left #darcs |
| 08:02 |
|
|
darcscommitbot joined #darcs |
| 08:03 |
|
|
darcswikibot joined #darcs |
| 08:20 |
|
|
lelit joined #darcs |
| 08:38 |
|
|
Weltraumschaf joined #darcs |
| 09:09 |
|
|
raichoo1 joined #darcs |
| 09:11 |
|
|
raichoo left #darcs |
| 09:26 |
|
|
copumpkin left #darcs |
| 09:26 |
|
|
copumpkin joined #darcs |
| 09:27 |
|
|
JaffaCake joined #darcs |
| 09:38 |
|
|
FunctorSalad left #darcs |
| 10:11 |
|
|
balor joined #darcs |
| 11:23 |
|
|
shenshei joined #darcs |
| 11:54 |
|
|
lelit` joined #darcs |
| 11:54 |
|
|
gh_ joined #darcs |
| 12:50 |
|
|
dankna left #darcs |
| 13:11 |
|
dixie |
hmm, I would like to use for the "darcs changes" upon the darcs source repository. |
| 13:11 |
|
dixie |
during the darcs tests. As far as I know this is prohibited and tests should use its own created testing repo. |
| 13:12 |
|
|
gal_bolle joined #darcs |
| 13:12 |
|
dixie |
the intention of the test is validation of XML output against XSD schema in contrib/ (using xmllint). |
| 13:13 |
|
dixie |
Is there way how to test in tests/ can figure out the original repository directory ? |
| 13:13 |
|
|
owst joined #darcs |
| 13:17 |
|
gal_bolle |
what do you mean by original? |
| 13:19 |
|
dixie |
sorry, I mean the shell tests are executed in /tmp/<....>/. I would like to have a path to repository with darcs being tested. |
| 13:20 |
|
gal_bolle |
you mean /home/dixie/src/darcs/screened or such? |
| 13:20 |
|
dixie |
yes |
| 13:20 |
|
mornfall |
dixie: You can't do that, because the repo may not exist (testing the release tarball, eg). |
| 13:21 |
|
gal_bolle |
you could try to see where the executable is |
| 13:21 |
|
mornfall |
dixie: Well, you could either skip the test if the repo is not there, or alternatively, put it under network tests and get the changes output from http://darcs.net. |
| 13:21 |
|
mornfall |
gal_bolle: Executables need not be in the source repository. |
| 13:21 |
|
dixie |
mornfall: that is good idea. I'll do that as a network test. |
| 13:22 |
|
gal_bolle |
no but it's a reasonable guess if you know it's fallible |
| 13:22 |
|
mornfall |
Guess so. |
| 13:22 |
|
dixie |
but the second small issue is that I want to put darcs.xsd into contrib. |
| 13:22 |
|
mornfall |
darcs.xsd being? |
| 13:22 |
|
dixie |
it is not part or the production... |
| 13:23 |
|
mornfall |
A schema? |
| 13:23 |
|
dixie |
mornfall: I want to create XSD schema for XML outputs. |
| 13:24 |
|
mornfall |
Well, you *could* go for $TESTDATA/../../contrib |
| 13:24 |
|
mornfall |
(In the test.) |
| 13:25 |
|
mornfall |
(Also you can check tat $TESTDATA/../../_darcs exists and make the test local...) |
| 13:25 |
|
mornfall |
that* |
| 13:25 |
|
mornfall |
As long as it is read-only, I guess it could be OK. |
| 13:26 |
|
mornfall |
(We can export something like $top_srcdir or something, if it turns out to be useful enough.) |
| 13:28 |
|
dixie |
thanks, it works with $TESTDATA. |
| 14:12 |
|
|
dankna joined #darcs |
| 14:14 |
|
|
JaffaCake left #darcs |
| 14:32 |
|
|
JaffaCake joined #darcs |
| 14:37 |
|
|
twb` joined #darcs |
| 14:44 |
|
|
balor left #darcs |
| 14:47 |
|
|
gal_bolle left #darcs |
| 14:48 |
|
|
gal_bolle joined #darcs |
| 15:06 |
|
|
twb left #darcs |
| 15:07 |
|
|
raichoo1 left #darcs |
| 15:26 |
|
|
copumpkin left #darcs |
| 15:48 |
|
|
secorp joined #darcs |
| 15:52 |
|
|
copumpkin joined #darcs |
| 16:02 |
|
|
secorp left #darcs |
| 16:14 |
|
|
sm joined #darcs |
| 16:20 |
|
|
gal_bolle left #darcs |
| 16:21 |
|
|
gal_bolle joined #darcs |
| 16:21 |
|
|
gal_bolle left #darcs |
| 16:28 |
|
|
sm_ joined #darcs |
| 16:29 |
|
mornfall |
dcoutts: http://buildbot.darcs.net/buil[…]figure/logs/stdio -- can you look and tell me why is cabal picking mtl 2 for the build, if everything is built with mtl 1? |
| 16:30 |
|
mornfall |
(At least that's what cabal is saying there...) |
| 16:31 |
|
dcoutts |
mornfall: runghc Setup configure is dumb |
| 16:31 |
|
copumpkin |
it seems like it'd be nice for darcs to have a magic patch-squashing system in place to avoid having to unpatch and repatch things |
| 16:31 |
|
dcoutts |
it always picks the latest version of everything (that satisfies the version constraints) |
| 16:31 |
|
copumpkin |
would that be easy to add? |
| 16:31 |
|
dcoutts |
mornfall: it makes no attempt to pick consistent things |
| 16:31 |
|
|
sm left #darcs |
| 16:31 |
|
|
sm_ is now known as sm |
| 16:32 |
|
dcoutts |
mornfall: this is one of the things that the cabal program adds over runghc Setup, picking packages more sensibly |
| 16:33 |
|
mornfall |
Oh. |
| 16:33 |
|
mornfall |
So runghc Setup configure and cabal configure use different algorithms? |
| 16:33 |
|
dcoutts |
yes |
| 16:33 |
|
mornfall |
:'( |
| 16:33 |
|
dcoutts |
in both cases however you can specify what you really want |
| 16:33 |
|
dcoutts |
--constraint="foo < x" |
| 16:33 |
|
dcoutts |
or whatever |
| 16:34 |
|
mornfall |
Right, but I need a single command to run on all the slaves. |
| 16:35 |
|
mornfall |
I can try to use cabal-install and see what happens. |
| 16:35 |
|
dcoutts |
and you have no control over the package environment? |
| 16:35 |
|
mornfall |
Not really, I don't own all the slaves. |
| 16:36 |
|
dcoutts |
and don't want to specify an mtl version on any of them |
| 16:36 |
|
mornfall |
cabal-dev or capri *could* be an option |
| 16:36 |
|
dcoutts |
it's also possible to do with runghc Setup |
| 16:36 |
|
dcoutts |
since you can use a private package db |
| 16:36 |
|
mornfall |
But I was trying to keep slave setup to minimum. |
| 16:37 |
|
dcoutts |
runghc Setup configure --package-db=local |
| 16:37 |
|
dcoutts |
or whatever |
| 16:37 |
|
mornfall |
Right, but then I need to create that package db somehow. Let me try with cabal install first, maybe all slaves have it and it works as expected. (fingers crossed) |
| 16:39 |
|
dankna |
can't you just create the local package db once and sync it to all the slaves by treating it as files? |
| 16:39 |
|
dcoutts |
no |
| 16:39 |
|
dankna |
note: I don't know how the package db works |
| 16:39 |
|
dcoutts |
but you can install a certain set of packages into a local package db |
| 16:39 |
|
dankna |
well, you don't have to waste time answering my newbie question :) |
| 16:43 |
|
mornfall |
OK, so using cabal fixes the windows slaves, but breaks OSX (since there's simply no cabal-install accessible to the buildslave). |
| 16:46 |
|
dankna |
why is that? I have a recentish cabal-install binary for x86-64-apple-darwin that I could send you if you need it. |
| 16:48 |
|
darcsbb |
build #47 of 6.12.1 Debian RELEASE is complete: Failure [failed test] Build details are at http://buildbot.darcs.net/buil[…]RELEASE/builds/47 |
| 16:48 |
|
mornfall |
Well, I have no control of the OSX slaves and I think Thorkil decided it was better to keep the buildslave away from being able to run it. |
| 16:48 |
|
dankna |
oh, okay |
| 16:48 |
|
dankna |
if I had an OSX desktop I'd volunteer it as a slave, that being the case, but I don't, sorry |
| 16:52 |
|
|
gwern left #darcs |
| 16:54 |
|
|
gwern joined #darcs |
| 16:54 |
|
|
gwern left #darcs |
| 16:54 |
|
|
gwern joined #darcs |
| 16:59 |
|
sm |
requiring cabal-install on slaves seems reasonable |
| 17:04 |
|
|
lelit` left #darcs |
| 17:24 |
|
mornfall |
Well, I made it configurable per-slave for now. |
| 17:24 |
|
mornfall |
So the OSX slaves use Setup but the others use cabal. |
| 17:30 |
|
|
shenshei left #darcs |
| 17:31 |
|
darcsbb |
build #48 of 6.12.1 Debian RELEASE is complete: Success [build successful] Build details are at http://buildbot.darcs.net/buil[…]RELEASE/builds/48 |
| 17:32 |
|
mornfall |
(That previous failure was a rather mysterious one, too.) |
| 17:40 |
|
|
lelit left #darcs |
| 17:44 |
|
sm |
if you looked at only one of buildbots many screens, what would be the one to watch ? http://buildbot.darcs.net/waterfall?category=head ? |
| 17:45 |
|
mornfall |
sm: Depending on what you look for. :) |
| 17:45 |
|
mornfall |
The HEAD and RELEASE waterfalls is what I normally look at. |
| 17:45 |
|
sm |
"darcs build status" |
| 17:46 |
|
sm |
for me, head is more interesting, I guess a release manager would watch release |
| 17:46 |
|
sm |
also, from there is it possible to get to the binary that was built ? |
| 17:46 |
|
mornfall |
Nope. |
| 17:47 |
|
sm |
pity, I wonder why not |
| 17:50 |
|
mornfall |
Technical issues. :) |
| 17:50 |
|
mornfall |
The slaves themselves are not, in general, internet-accessible. |
| 17:50 |
|
mornfall |
Giving them a way to access internet accessible hosts to upload stuff seems somewhat dangerous, too. |
| 17:50 |
|
darcsbb |
build #33 of 6.12.3 Vista RELEASE is complete: Success [build successful] Build details are at http://buildbot.darcs.net/buil[…]RELEASE/builds/33 |
| 17:51 |
|
mornfall |
And last but not least, it would take a *lot* of space to store the results. |
| 17:51 |
|
sm |
it sure would be a boon to provide consistently up-to-date binaries for a wide range of platforms |
| 17:51 |
|
mornfall |
sm: Well, what for? :) |
| 17:51 |
|
sm |
have you not seen people asking for binaries ? |
| 17:51 |
|
mornfall |
Of HEAD? |
| 17:52 |
|
mornfall |
I don't think. |
| 17:52 |
|
sm |
darcs would go from "painful to install on this machine, don't bother" to "easy" |
| 17:52 |
|
mornfall |
We already have a collection of binaries. |
| 17:52 |
|
sm |
better head binaries than nothing |
| 17:52 |
|
mornfall |
Of released versions. |
| 17:52 |
|
sm |
not a consistent collection, across this range of platforms, I think |
| 17:53 |
|
|
Weltraumschaf left #darcs |
| 17:53 |
|
sm |
their existence and timing is a hit and miss |
| 17:53 |
|
sm |
affair |
| 17:53 |
|
mornfall |
We have Windows, Linux, OSX, FreeBSD and OpenBSD, for x86 and some amd64. We don't have nearly as many buildslaves. |
| 17:53 |
|
mornfall |
There are no BSD ones at all, eg. |
| 17:54 |
|
mornfall |
And building *useful* binaries takes a fair bit more work than just taking what the buildbot made. |
| 17:55 |
|
mornfall |
It's same as a source release. If it was just about making a tag and running a script, we wouldn't need a release manager. |
| 17:55 |
|
sm |
I'm all for moving as far in that direction as possible |
| 17:56 |
|
mornfall |
We are all. But we all also know that it's currently unattainable. |
| 17:56 |
|
|
shenshei joined #darcs |
| 17:56 |
|
mornfall |
And the situation is the same with binaries. |
| 18:09 |
|
|
gwern left #darcs |
| 18:13 |
|
|
iago joined #darcs |
| 18:14 |
|
iago |
hi |
| 18:39 |
|
iago |
could someone help me understanding some Darcs.Test.Patch.Properties details? |
| 18:40 |
|
iago |
for some properties I have difficulties distinguish when a Nothing means "the test was ok" or when it means "the test has no sense / there is some pre-condition that fails" |
| 18:55 |
|
|
copumpkin is now known as auger |
| 19:02 |
|
|
auger is now known as copumpkin |
| 19:13 |
|
|
lispy joined #darcs |
| 19:21 |
|
|
FunctorSalad joined #darcs |
| 19:35 |
|
|
secorp joined #darcs |
| 20:16 |
|
iago |
Heffalump, please say something when you were available |
| 20:18 |
|
* dixie |
remembered pulp fiction... Heffalump will say "something" :) |
| 21:10 |
|
owst |
Is it still possible to build darcs with profiling enabled? I don't seem to be having any luck with the details on the wiki (unrecognised option --enable-profile) |
| 21:12 |
|
iago |
owst, how are you trying to do that? |
| 21:12 |
|
owst |
cabal configure --enable-profile |
| 21:12 |
|
owst |
(I also attempted --enable-library-profiling, but no luck) |
| 21:13 |
|
iago |
I have library-profiling in .cabal/config and it works fine |
| 21:13 |
|
owst |
Argh, nevermind, just found what I need: --enable-executable-profiling |
| 21:13 |
|
owst |
Sorry for the noise. |
| 21:31 |
|
|
rks left #darcs |
| 21:44 |
|
owst |
This may be an obvious question, but do all the libraries used by darcs need to have profiling enabled when they are built to use profiling on the darcs executable? |
| 21:47 |
|
sm |
owst: yes I'm afraid so |
| 21:47 |
|
owst |
Eek. |
| 21:47 |
|
owst |
Is there an easy way to accomplish that task? |
| 21:47 |
|
sm |
you might need to cabal reinstall -p a heck of a lot of libraries. Or try to install in a capri or cabal-dev sandbox |
| 21:48 |
|
sm |
cabal install --reinstall -p, I meant |
| 21:48 |
|
owst |
;) that's what I was afraid of. Ah, haven't heard of cabal-dev, I shall check it out. |
| 21:48 |
|
sm |
the --reinstall method does work, just a bit tedious |
| 21:48 |
|
owst |
Yes, indeed. |
| 21:49 |
|
owst |
sm: *sigh*, thanks :) |
| 21:49 |
|
sm |
np |
| 21:56 |
|
Heffalump |
I'm here for a bit |
| 21:58 |
|
iago |
Heffalump, can I ask you for some test properties ? |
| 21:59 |
|
iago |
I just want to be sure that some Nothings mean succeeded and others mean rejected |
| 21:59 |
|
Heffalump |
sure, but I don't have any magic formula for knowing either :-) |
| 22:00 |
|
iago |
yeah but I'm pretty sure you understand better what the property is testing |
| 22:00 |
|
iago |
which helps a lot |
| 22:02 |
|
iago |
Heffalump, I think the better way is that I paste you each property and you comment me what do you thing, I have added commentaries like "succeeded?" etc |
| 22:02 |
|
iago |
say me if you prefer another way |
| 22:02 |
|
iago |
for example, this one http://paste.debian.net/107219/ (they are just 2-3) |
| 22:03 |
|
Heffalump |
I'm not quite sure what you're asking. |
| 22:03 |
|
Heffalump |
I think the Nothing cases are the uninteresting ones for this test |
| 22:03 |
|
iago |
Nothing -> rejected -- succeeded? |
| 22:03 |
|
Heffalump |
the first three Nothing cases, that is |
| 22:03 |
|
iago |
mean "I though it is a case that should be rejected, but I wonder if it should be a case succeeded" |
| 22:04 |
|
Heffalump |
rejected means "this is not a good test case", doesn't it? |
| 22:04 |
|
Heffalump |
it shouldn't be a failure. |
| 22:04 |
|
iago |
yep, rejected means quickcheck will discard that test case and generate another |
| 22:05 |
|
iago |
for quickcheck, for hunit means OK |
| 22:05 |
|
Heffalump |
that seems like the right thing to do here |
| 22:05 |
|
iago |
just 2 more |
| 22:05 |
|
iago |
http://paste.debian.net/107220/ |
| 22:05 |
|
iago |
and http://paste.debian.net/107221/ |
| 22:06 |
|
Heffalump |
what's <&&> ? |
| 22:06 |
|
iago |
oh, yeah sorry |
| 22:06 |
|
iago |
its my replacement for msum |
| 22:06 |
|
Heffalump |
I'm not quite sure I understand what partialPermutivity is trying to test. Is it run on RealPatch? |
| 22:06 |
|
iago |
mplus, sorry |
| 22:07 |
|
iago |
-- | @t <&&> s@ fails <=> t or s fails |
| 22:07 |
|
iago |
-- @t <&&> s@ succeeds <=> none fails and some succeeds |
| 22:07 |
|
iago |
-- @t <&&> s@ is rejected <=> none fails and both are rejected |
| 22:07 |
|
iago |
(<&&>) :: TestResult -> TestResult -> TestResult |
| 22:07 |
|
iago |
t@(TestFailed _) <&&> _s = t |
| 22:07 |
|
iago |
_t <&&> s@(TestFailed _) = s |
| 22:07 |
|
iago |
TestRejected <&&> s = s |
| 22:07 |
|
iago |
t <&&> TestRejected = t |
| 22:07 |
|
iago |
TestSucceeded <&&> TestSucceeded = TestSucceeded |
| 22:07 |
|
iago |
it is used here "runPrimitiveTests "real partial permutivity" (partialPermutivity commute) $ filter (notDuplicatestriple) realTriples" |
| 22:08 |
|
iago |
with realpatchs yeah |
| 22:08 |
|
Heffalump |
ok, I think that's because full permutivity isn't valid |
| 22:08 |
|
Heffalump |
I guess the same logic applies as for permutivity. I think the test is a bit dubious/weird anyway. |
| 22:08 |
|
Heffalump |
Arguably once the first pp has succeeded, the second one should not be allowed to be rejected. |
| 22:09 |
|
Heffalump |
but I don't think it's worth wasting energy on |
| 22:09 |
|
Heffalump |
I'm afraid I never quite understood joinCommute either. |
| 22:11 |
|
iago |
I'm trying to keep my brain sane |
| 22:11 |
|
iago |
so to be honest I did not make a big effort to try to understand joinCommute |
| 22:13 |
|
iago |
well, running tests there is not error reported so the worst can happen is that you are testing more than needed |
| 22:13 |
|
iago |
though I have to say you that the time to run test was increased |
| 22:14 |
|
iago |
it usually takes 4-5 attempts to generate an useful test case |
| 22:20 |
|
iago |
I "feel" the last "succeeded" in joinCommute should be rejected instead |
| 22:21 |
|
iago |
just because I don't see why (a :> b :>: c :>: NilFL) have to commute due the previous facts |
| 22:21 |
|
iago |
but, it is just my feeling |
| 22:23 |
|
Heffalump |
I think it could either be rejected or succeeded |
| 22:23 |
|
Heffalump |
I think 'j' will be a function that squishes two patches together. |
| 22:23 |
|
Heffalump |
So the test is that commuting before and after squishing is consistent. |
| 22:26 |
|
iago |
I see now |
| 22:26 |
|
iago |
well, rejected can always be a succeeded |
| 22:27 |
|
iago |
is like to say "succeeded but unuseful" |
| 22:27 |
|
iago |
useless* |
| 22:29 |
|
sm |
hey all.. #haskell-game is curating http://www.haskell.org/haskell[…]d_libraries/Games a bit, and wondering if droundy's original haskell bridge game is accessible anywhere - any idea ? Not the modern go version |
| 22:29 |
|
sm |
obviously. Not linking to *that* |
| 22:29 |
|
* sm |
sniffs |
| 22:30 |
|
Heffalump |
was it actually in Haskell? |
| 22:30 |
|
Heffalump |
because IIRC he worked on it before darcs, and the first darcs impl was in C++ |
| 22:30 |
|
Heffalump |
in any case, no idea |
| 22:33 |
|
sm |
not sure, good question. Anyway it's not findable, so we dropped it |
| 22:33 |
|
sm |
thanks |
| 22:38 |
|
|
Weltraumschaf joined #darcs |
| 22:39 |
|
Weltraumschaf |
alexsuraci: hi, is it possable to checkout the code from darcsden, so that i can run it local on my machine? |
| 22:41 |
|
alexsuraci |
Weltraumschaf: yes, darcs get http://darcsden.com/username/reponame |
| 22:41 |
|
alexsuraci |
or username darcsden.com:reponame if it's your repo |
| 22:43 |
|
Weltraumschaf |
sry, i was unclear. i want to run an instance of darcsden local on my machine. i'm building an browsing integration for the jenkins-darcs plugin and want to develop offline. so it wpould be nice to have a local darcsden website on my machine |
| 22:43 |
|
Weltraumschaf |
i just creat |
| 22:43 |
|
sm |
you sure can, but the dependencies can be a bit tricky |
| 22:44 |
|
sm |
you'll need couchdb and redis, possibly dev version alexsuraci ? |
| 22:44 |
|
sm |
I seem to remember some issue like that |
| 22:44 |
|
Weltraumschaf |
etoday i got a really good hint on #jenkins so i will implement darcsweb and darcsden browsing for the darcsplugin, so we can use darcs with jenkins :) |
| 22:44 |
|
Weltraumschaf |
couchdb and redis is no problem, i have it installed |
| 22:45 |
|
alexsuraci |
sm: any ol' couchdb and redis setup will do, even set to their defaults |
| 22:45 |
|
sm |
ok, it's been a while |
| 22:45 |
|
alexsuraci |
bbl though, heading to class |
| 22:45 |
|
sm |
Weltraumschaf: sounds great, what is jenkins again ? |
| 22:45 |
|
sm |
it's hard to find |
| 22:46 |
|
sm |
ah http://jenkins-ci.org/ |
| 22:46 |
|
Weltraumschaf |
jenkins aka. hudson is a continous integration server. very lightweight and easy to use |
| 22:46 |
|
Weltraumschaf |
i love it |
| 22:46 |
|
sm |
jenkins *is* hudson ? |
| 22:47 |
|
Weltraumschaf |
just now its the sam. but oracle... the made trouble and now they do their 'own' hudson project and the whole community continues with jenkins |
| 22:47 |
|
sm |
ah, good to know |
| 22:47 |
|
Weltraumschaf |
it is like buildbot i gues |
| 22:48 |
|
|
secorp left #darcs |
| 22:49 |
|
Heffalump |
what's the history of hudson, then? |
| 22:53 |
|
Weltraumschaf |
khosuke, formely employee of sun, build it as integration server. then sun sayd: we're not interested anymore, we give to you, do what you want with it. and khosuke developed it with the community. it's wideyl used, because its so easy to setup and do work with it. |
| 22:53 |
|
Heffalump |
we use it at work, as it happens |
| 22:54 |
|
Weltraumschaf |
then oracle bought sun and sayd: ok, sund gives u everything, but we own the right on the name hudson! and we do not want to give the rights to the opensource community, bla bla bla advocates bla |
| 22:54 |
|
Weltraumschaf |
and then the community decided on that uncertainty to change to a name which is 'free'. anothers butler name: jenkins |
| 22:54 |
|
|
raichoo joined #darcs |
| 22:54 |
|
Heffalump |
I see :-) Is it really worth it to Oracle to maintain a fork? |
| 22:55 |
|
Weltraumschaf |
i dont thinks so. thw whole. and i mean the whole community is on jenkins now. oracle is alone. i think hudson will die soon |
| 22:56 |
|
Weltraumschaf |
Heffalump: r u using hudson with darcs at work? |
| 22:56 |
|
Heffalump |
sadly not :-) Clearcase. |
| 22:56 |
|
Weltraumschaf |
outch |
| 22:56 |
|
Weltraumschaf |
my condolence |
| 22:57 |
|
Heffalump |
I do indeed need all the sympathy I can get for that. |
| 22:57 |
|
* Weltraumschaf |
never used clercase, but only heard bad things about it |
| 22:58 |
|
Weltraumschaf |
yeah, and it's not easy to switch the scm |
| 22:58 |
|
Heffalump |
we're actually in the middle of trying to figure out what to switch to |
| 22:58 |
|
Heffalump |
it's pretty hard |
| 22:58 |
|
Weltraumschaf |
i think git is a good choice, becaus it has really good features AND really good toolintegration |
| 22:59 |
|
Heffalump |
much of the rest of the firm uses svn, but we're not sure the branching support is robust enough for us |
| 22:59 |
|
Weltraumschaf |
u can integrate it everywhere |
| 22:59 |
|
Heffalump |
IT security apparently dislikes DVCS. |
| 22:59 |
|
Heffalump |
also I'd be concerned about its shoot-yourself-in-the-foot reputation - most of our coders are mathematicians rather than software engineers. |
| 22:59 |
|
Weltraumschaf |
u can use git svn like. u have not to use it distribute. u can use it like comminiting to one trunk |
| 23:00 |
|
Heffalump |
I know that. But this is IT security we're talking about. |
| 23:00 |
|
Weltraumschaf |
svn is also good. i use it, too. but it is slow on large amount of commint. and the big scary merge is real PITA. believe me! |
| 23:01 |
|
Weltraumschaf |
i were 3 years the, were developers come: oh, sven, please!1! coudld you make the merge for me please O_o |
| 23:02 |
|
Weltraumschaf |
in my actull work we use darcs. and its much mor easier to integrate the feature 'branched' work of the developers, than with svn |
| 23:02 |
|
Weltraumschaf |
uhh, that was really bad english... sry i drunk some cocktails... |
| 23:02 |
|
Heffalump |
:-) |
| 23:03 |
|
|
twb` left #darcs |
| 23:04 |
|
Weltraumschaf |
but back on topic, when u allready use svn, then svn would be better for u: u have no pain with clearcase anymore and u have only one kind of SCM to integrate in the rest of ur infrastructre |
| 23:04 |
|
Weltraumschaf |
ok, u'r not so hip then with u'r company with ol'school svn *lol* |
| 23:04 |
|
|
twb joined #darcs |
| 23:04 |
|
Heffalump |
well, the rest of the firm uses it, our infrastructure is fairly independent of theirs |
| 23:05 |
|
Heffalump |
but it'd certainly help with having them run the servers and stuff |
| 23:05 |
|
gh_ |
"and wondering if droundy's original haskell bridge game is accessible anywhere - any idea ?" hence the website named "abridgegame"… *revelation* |
| 23:05 |
|
Heffalump |
gh_: :-) |
| 23:05 |
|
twb |
Didn't it die? |
| 23:05 |
|
twb |
The game, not the site |
| 23:05 |
|
sm |
gh_: sure, the domain is still there |
| 23:06 |
|
Weltraumschaf |
but u have to maintain a clearcase infrastructure and a svn one. when u switch all to svn u have less maintanace with systems and automation. less variation in u'r business. more reuse of stuff |
| 23:06 |
|
Heffalump |
we know all that :-) |
| 23:06 |
|
Weltraumschaf |
but i know. switching the scm is not easy. |
| 23:06 |
|
Heffalump |
but if svn screws up merges or is too slow for a large codebase with lots of history, it'll burn up that saving very fast |
| 23:07 |
|
Weltraumschaf |
we thought about switching from darcs to git, because some developers wanted to use the more hip scm |
| 23:07 |
|
Heffalump |
we have to switch off Clearcase though, it's crap and expensive. We're just considering things other than svn too. |
| 23:08 |
|
Weltraumschaf |
maybe there are good alternatives in the paid market. i can only speek for darcs, svn and git, because i used it. (ok, and sourcesafe for 2 weeks *lol*) |
| 23:08 |
|
Weltraumschaf |
sourcesafe is crazy |
| 23:08 |
|
Heffalump |
we're looking at Accurev, Perforce and MS TFS as well |
| 23:09 |
|
Weltraumschaf |
iperforce and TFS i only know from the jenkins plugins, because there i stolen some code for the darcs plugin ;) |
| 23:10 |
|
Weltraumschaf |
the perforce devloper gave me the critical hint to get it work |
| 23:10 |
|
Weltraumschaf |
what i love with darcs/git is the possibility to commit my changes offline. with svn you need all the time connection to the main repo |
| 23:11 |
|
Weltraumschaf |
and did you pulled off the network cable when svn does a commit for fun? crazy. the whole workspace is screwed up after that O_o |
| 23:14 |
|
Weltraumschaf |
but what i wonder: why have u'r IT security guys have concerns about DVCS? |
| 23:17 |
|
|
dankna left #darcs |
| 23:17 |
|
Heffalump |
I didn't dig, but I guess they dislike the idea that someone can walk off with a copy of the full repo. And yes, I know. You can do that with non DVCS too. |
| 23:18 |
|
Weltraumschaf |
lol |
| 23:19 |
|
Weltraumschaf |
but ithin about the ssh features of git. i is imho more safe with the public private key stuff than svn |
| 23:19 |
|
Weltraumschaf |
but i think its only the: uuuu its new, its devel thenology thing, right? |
| 23:21 |
|
Weltraumschaf |
i'll go to sleep. cu |
| 23:22 |
|
|
Weltraumschaf left #darcs |
| 23:25 |
|
|
twb` joined #darcs |
| 23:27 |
|
|
twb` left #darcs |
| 23:33 |
|
|
secorp joined #darcs |
| 23:43 |
|
|
secorp left #darcs |
| 23:45 |
|
|
iago left #darcs |
| 23:52 |
|
|
lispy left #darcs |
| 23:54 |
|
* Heffalump |
releases 2.5.1 and goes to bed |
| 23:55 |
|
gh_ |
\o/ |
