IRC log of #zope3-dev for Tuesday, 2006-09-05

*** wrobel has quit IRC00:04
*** jhauser has quit IRC00:16
*** philiKON has joined #zope3-dev00:19
*** ktwilight has quit IRC00:32
*** philiKON has quit IRC00:49
*** ktwilight has joined #zope3-dev00:54
*** volvox has quit IRC00:55
*** rocky has quit IRC01:03
*** natea_ has joined #zope3-dev01:07
*** srichter has joined #zope3-dev01:07
*** ChanServ sets mode: +o srichter01:08
*** natea has quit IRC01:13
*** natea has joined #zope3-dev01:20
*** vlado has joined #zope3-dev01:21
*** natea_ has quit IRC01:26
ignasanyone who is knowledgable of ZODB ?01:27
*** vlado has quit IRC01:27
*** rocky has joined #zope3-dev01:30
*** natea has quit IRC01:55
*** natea has joined #zope3-dev01:56
*** jan_s has joined #zope3-dev02:18
*** rocky is now known as rocky|Zzz02:25
*** natea has quit IRC02:35
*** ignas has quit IRC02:38
*** TresEquis has quit IRC03:24
*** jinty has quit IRC03:24
*** yota has quit IRC03:33
*** natea has joined #zope3-dev03:37
*** projekt01 has left #zope3-dev03:58
*** natea_ has joined #zope3-dev03:59
*** natea has quit IRC04:02
*** natea has joined #zope3-dev04:17
*** natea_ has quit IRC04:26
*** stub has joined #zope3-dev04:48
*** natea_ has joined #zope3-dev04:58
*** natea has quit IRC05:04
*** niemeyer has quit IRC05:23
*** natea_ has quit IRC05:46
*** natea has joined #zope3-dev05:49
*** natea has quit IRC06:15
*** natea has joined #zope3-dev06:16
*** natea_ has joined #zope3-dev06:42
*** natea has quit IRC06:50
*** natea has joined #zope3-dev06:54
*** natea_ has quit IRC07:02
*** jan_s has quit IRC07:24
*** natea has quit IRC07:34
*** natea has joined #zope3-dev07:35
*** dobee has joined #zope3-dev07:54
*** timte has joined #zope3-dev08:10
*** romanofski has quit IRC08:12
*** eins has joined #zope3-dev08:15
*** natea has quit IRC08:22
*** natea has joined #zope3-dev08:22
*** batlogg has quit IRC08:27
*** batlogg has joined #zope3-dev08:37
*** zagy has joined #zope3-dev08:48
*** alecm has joined #zope3-dev08:55
*** xenru has quit IRC08:56
*** wrobel has joined #zope3-dev08:59
*** natea has quit IRC09:13
*** natea has joined #zope3-dev09:13
*** hdima has joined #zope3-dev09:16
*** kobold has joined #zope3-dev09:19
*** jhauser has joined #zope3-dev09:26
*** alecm has quit IRC09:27
*** zagy has quit IRC09:27
*** romanofski has joined #zope3-dev09:28
*** xenru has joined #zope3-dev09:32
*** zbir_ has joined #zope3-dev09:33
*** zagy has joined #zope3-dev09:35
*** philiKON has joined #zope3-dev09:42
*** zbir has quit IRC09:49
*** stu1 has joined #zope3-dev09:49
*** projekt01 has joined #zope3-dev09:56
*** natea has quit IRC10:01
*** natea has joined #zope3-dev10:02
*** stub has quit IRC10:04
*** yota has joined #zope3-dev10:07
*** stu1 is now known as stub10:42
*** MJ has quit IRC10:47
*** natea has quit IRC10:51
*** scherand has joined #zope3-dev11:10
*** MJ has joined #zope3-dev11:12
nouribrilliant: $ easy_install zc.buildout11:41
nourierror: Not a URL, existing file, or requirement spec: 'zc.buildout'11:41
nourioh well, a newer version of setuptools seems to be able to handle it, nvm11:46
*** Aiste has quit IRC12:10
*** ignas has joined #zope3-dev12:13
*** volvox has joined #zope3-dev12:17
*** gintas has joined #zope3-dev12:33
*** Aiste has joined #zope3-dev12:34
*** alga has joined #zope3-dev12:41
*** jinty has joined #zope3-dev12:41
*** batlogg_ has joined #zope3-dev12:41
*** batlogg has quit IRC12:41
*** _projekt01 has joined #zope3-dev12:43
*** _projekt01 has quit IRC12:48
*** projekt01 has quit IRC13:00
*** Aiste has quit IRC13:42
*** dunny has quit IRC13:43
*** romanofski has quit IRC13:58
*** romanofski has joined #zope3-dev14:01
*** mkerrin has joined #zope3-dev14:07
*** stainsby has joined #zope3-dev14:26
*** benji has quit IRC14:26
*** faassen has joined #zope3-dev14:39
*** rocky|Zzz is now known as rocky14:41
*** J1m has joined #zope3-dev14:45
faassenJ1m: hey14:47
*** xenru has quit IRC14:51
*** niemeyer has joined #zope3-dev14:57
*** mgedmin has joined #zope3-dev15:18
ignaswhy suddenly <include package="" /> is not working anymore ?15:21
ignasit is telling me:15:21
ignas        ZopeXMLConfigurationError: File "/home/ignas/src/schooltool/Zope3/src/zope/app/session/configure.zcml", line 12.2-15.815:21
ignas        TypeError: Missing 'provides' attribute15:21
philiKONi did something there15:27
philiKONzope 3 trunk?15:27
philiKONignas, are on the trunk or zope 3.3?15:27
philiKONweird. that works for me15:28 shows no test failures eitehr15:29
ignasi got <include package="" /> in schooltools  site.zcml15:29
philiKONignas, did you update all of your z3 checkout?15:29 should now contain an adapts() declaration15:30
ignassvn up; make clean; make15:30
ignasit does contain adapts()15:30
ignasschooltools buildbot is spewing this too and it checks out new Zope3.3 every time i think15:31
philiKONvery weird15:33
philiKONignas, i actually jus tproduced an app with session support, so i know it works for me15:35
philiKONi can't possibly think fo a way to reproduce this15:35
philiKONignas, what if you get a blank zope 3.3 checkout, create a new instance and just start it up15:35
philiKONdoes it happen?15:35
*** benji has joined #zope3-dev15:36
ignasi'll try that15:39
ignasthough, schooltool is a bit more complicated as it is not exactly a Zope3 product it's more of an application built around Zope315:39
mgedminignas: the problem is that implements two interfaces: ISession and IPathAdapter15:42
mgedminsee the if len(p) == 1 check?15:42
mgedminsrc/zope/component/ line 17715:42
mgedminhuh15:43 has two <adapter> directives15:43
philiKONmgedmin, implements *one* interface15:44
mgedminand a <class> directove that tells session.Session that it implements IPathAdapter15:44
philiKONclass Session(object):15:44
philiKON    """See"""15:44
philiKON    implements(ISession)15:44
philiKON    adapts(IRequest)15:44
mgedmin  <class class=".session.Session">15:44
mgedmin    <allow interface=".interfaces.ISession" />15:44
mgedmin    <implements interface="zope.traversing.interfaces.IPathAdapter" />15:44
mgedmin  </class>15:44
philiKONoh shit15:44
mgedmincould it be that in schooltool the ZCML directives get reordered somehow?15:44
philiKONh mm15:44
* mgedmin always hated that zcml was order-dependent15:44
philiKONperhaps you guys have your custom <class> directive?15:44
mgedminbut random-order-dependent could be even worse :/15:45
ignasmgedmin: yes the first though was that it depends on the order of includes15:45
philiKONactions should be executed the way the original definitions were15:45
mgedminI know15:45
ignasand schooltool does it in a non standard way15:45
mgedminphiliKON: we load the ZCML file more than once in a single process15:45
mgedminin unit tests15:45
mgedminthe effects of the <implements> directive stick around15:45
mgedminplacelessTearDown doesn't clear them15:45
* mgedmin hates that the zope component architecture cannot be cleanly restored to an older state15:45
philiKONwell, test cleanup is hard15:46
philiKONthis hasn't much to do with the CA15:46
philiKONit's object's states15:46
philiKONwhat i did in five was15:46
philiKONhave the zcml directives remember all monkeyed objects15:46
philiKONand provide a cleanup hook15:46
*** vlado has joined #zope3-dev15:47
mgedminignas: simplest fix that I can think of: in our test's tearDown call implementsOnly(, ISession)15:47
* ignas eeks, shudders and goes to implement it15:48
philiKONnah, i'll add the provides="" in zcml15:48
philiKONdang, i'm out of commit range15:48
benjiignas: you could also put your tests in a layer to isolate them from the other tests15:48
mgedminwhat kind of firewall lets IRC through but not SSH?15:48
ignasbenji: can't, am using a custom testrunner15:48
philiKONmgedmin, you don't wanna know15:49
philiKONi'm not even here, officially ;)15:49
philiKONhi benji15:49
* mgedmin smells an opportunity for blackmail15:49
benjihi, philiKON15:49
*** gintas has quit IRC15:50
*** Aiste has joined #zope3-dev15:53
*** Aiste has quit IRC15:55
*** natea has joined #zope3-dev15:58
*** nathany has joined #zope3-dev16:01
* ignas just wants to show off :) upto 5:1 size ratio ~20% slower sometimes as fast as uncompressed, totally experimental16:09
ignasnot for use in any production systems16:10
ignasjust a proof of concept that shows that it is possible16:10
ignasas I still can't understand why it hasn't been done before :(16:10
mgedminignas: I committed the zcml change16:12
ignasmgedmin: thanks a lot16:13
*** stub has quit IRC16:15
*** torkel_ has joined #zope3-dev16:16
philiKONmgedmin, thanks!16:23
philiKONignas, and sorry for the inconvenience16:23
*** natea has quit IRC16:35
*** batlogg_ has quit IRC16:35
*** batlogg has joined #zope3-dev16:36
*** natea has joined #zope3-dev16:37
*** zbir_ is now known as zbir16:39
*** d21 has joined #zope3-dev16:46
*** d2m has quit IRC16:48
*** stub has joined #zope3-dev16:50
*** eins has quit IRC16:52
*** whit has joined #zope3-dev16:56
*** hdima has quit IRC17:07
*** gintas has joined #zope3-dev17:08
*** d21 is now known as d2m17:10
*** stainsby has left #zope3-dev17:11
*** vlado has quit IRC17:20
*** nathany_ has joined #zope3-dev17:26
*** vlado has joined #zope3-dev17:32
*** projekt01 has joined #zope3-dev17:32
*** stub has quit IRC17:38
*** vlado has quit IRC17:40
*** nathany has quit IRC17:42
*** gumpa has joined #zope3-dev17:44
*** whit has quit IRC17:52
philiKONJ1m, what's the underscore naming scheme and how has python adopted it?17:56
benjiI (though haven't looked) don't know what naming convention formlib violates.17:57
srichterformluib violates Zope 3's style guide of using CamelCase17:57
benjiCamelCase for non-classes?17:57
rockyZope 3 style guide violates python style guide :)17:57
srichterbenji: yes17:57
J1mSee Naming Conventions.17:58
benjiwhat entities does it CamelCase?17:58
srichterrocky: but that's because Python's style guide was decided upon after ours17:58
rockysrichter: oh i totally understand, but it doesn't change my comment ;)17:58
philiKONsrichter, where's the camelcase violation?17:58
philiKONin formlib17:58
philiKONwhere's the problem17:59
J1mThe Zope 3 style guide adoped camelCase method names before the Python style guide made a choice.17:59
philiKONcamelCase for variable names?17:59
srichterphiliKON: it violates our style guide17:59
philiKONerr, i mean, camelCase for attribute names i suppose?17:59
philiKONhrrm, i actually prefer underscores17:59
philiKONand i don't think it's worth it changing formlib18:00
J1mThe Python style guide later adopted lower_case_with_underscore for method names.18:00
J1mand public attrs.18:00
philiKONoh, even for method names?18:00
philiKONthat sucks18:00
J1mAnd it explicitly says that:18:00
srichterpreference does not matter; it's uniformity that matters, because it let's allows you to remember less18:00
J1m- existing APIs should not change18:00
J1m- APIs should be internally consistent.18:00
rockynow that *does* invalidate my comment that zope 3 style guide violates python style guide ;)18:01
srichterbut Python was in Zope 2's situation and did not do a rewrite18:01
srichterI wonder whether Python 3000 will change the APIs to match the style guides18:01
philiKONsrichter, it won't18:02
philiKONguido said so18:02
benjiI don't recall that, philiKON; I thought it was still an open question18:03
philiKONhe said it at europython18:03
philiKONat least that's how i interpreted it18:03
srichterstill, we have a style guide and it has been violated; and I think there is no reason to change it; and to keep violating it defeats its purpose18:03
benjiahh, europython isn't in the email archive :)18:03
srichterI actually agree with Fred that there is no benefit in changeing the style guide18:04
J1mI think there is a benefit in changing it.18:04
philiKONrefactoring + deprecation just for style (attribute naming) isn't worth it IMO18:04
rockybenji: inferior verbal communication is always bad ;)18:04
benjiI want both consistent style within Zope, and to be consistent with PEP 8; can't have both though18:04
J1mThe the degree that packages are usable outside of Zope, they should follow the Python style guide.18:05
philiKON+1 to PEP000818:05
J1mOf course, Pep 8 says to not slavishly follow PEP 8. :)18:05
srichterJ1m: but especially things like formlib are not really usable outside18:06
philiKONyeah, i'll still callMethodsThisWay.18:06
philiKONand attributes_this_way18:06
J1mI sometimes use more whitespace than PEP 8 calls for.18:06
J1msrichter: yeah, I was anticipating changing the style guide. :)18:06
srichterbut I want much more discussion18:07
srichterI think the benefit of changing it is very small18:07
philiKONi think the benefit of changing formlib is also small18:08
philiKONthe damage is greater18:08
J1mI think the benefit of being different is smallter than its cost.18:08
benjiI think I can't parse that last comment.18:08
rockywell, the zope 3 style guide is yet another instance of us doing something our own way versus the python way (looking from outside in) ... it's irrelevant who came first18:08
srichterwe obviously disagree :-)18:08
philiKONbenji, J1m thinks being different isn't worth it18:09
J1mrocky: the Zope style guide goes out of it's way to follow the Python style guide.18:09
benjirocky: part of the problem is that PEP 8 changed, and we were following the pre-changed guidelines18:09
J1mOr making a commitment when before the PEP did.18:09
rockyJ1m: it *did* you mean ;)18:09
faassenyeah, those nasty python people are going out of their way to sabotage our project. :)18:09
rockyi think zope should build it's own better python-derived language  ... pull a MS C#/.net ;)18:10
srichterstyle guide is about uniformity; changing it removes that uniformity; I think part of Java's success has been this uniformity; Jason Orendorff said once to me: I can almost always anticipate the API of a Java class18:10
rockymgedmin: nice!18:10
J1meel (combination of fish and snake)18:10
philiKONsrichter, SomethingFactoryFactoryFActory?18:11
J1msrichter: no one is arguing with the benefit of uniformity.18:11
J1mWe are just trying to be uniform within a larger realm.18:12
benjiwe just need a getattr hack that translates a_method_name to aMethodName and vice versa <.87 wink>18:12
J1mphiliKON: the Python style guide uses CamelCase for class names.18:12
philiKONJ1m, i know. i was mocking on java's Factory things18:12
* mgedmin prepares for a underscoregeddon18:12
mgedminspeaking of hava, have you read The World of Nouns?18:13
philiKONi don't think we'll change getMultiADapter to get_multi_adapter18:13
philiKONmgedmin, yes!18:13
J1mIf we followed PEP 8, we would:18:13
J1m- Not change existing names18:13
J1m- Remain consistent within APIs.18:13
J1mSo zope.component would continue to use camelCase.18:14
philiKON- use PEP8 for all new packages, though18:14
J1mThat would be my vote.18:14
rockyphiliKON: remember, pep8 says you don't use camelCase for method names... you gonna be able to handle that? :)18:14
philiKONrocky, it'll be hard18:14
srichterI think the costs of having to remember which package is "new" and whcih is "old" is not worth it18:14
benjithat seems to be the best we can do18:14
philiKONlike the time when i switched from 2 spaces to 4 spaces indentation18:14
faassenmgedmin: what's the World of Nouns? I'm not sure I have or not and I can't google it up.18:15
mgedminI misremembered the name18:15
mgedminit's Execution in the Kingdom of Nouns18:15
rockyphiliKON: ugh, 2 spaces, ugh  ;)18:15
philiKONoh, take mgedmin's url. i was feeling-lucky-googling18:15
faassenah, kingdom. :)18:15
* ignas doesn't want to go back to python style naming18:15
ignaspython styling meaning18:15
ignasget_value getvalue or getValue18:16
ignasall being a possibility18:16
philiKONpep8 sez get_value18:16
ignasand python library sez what ?18:16
philiKONthat it's not gonna change18:16
ignassame with Zope3 ... PEP8 sez get_value, Zope3 sez get_value or getValue depending on the module18:17
J1mformlib is an interesting case.18:17
J1mIt started life as a zc-internal project.18:17
srichterright, and then you provide a list of "new and "old" modules :-(18:17
J1mWhen it was written, there was no onus on it to conform to the Zope 3 style.18:18
philiKONas if the python stdlib was that hard to use just because of the inconsistent naming18:18
J1mphiliKON: exactly18:18
srichterJ1m: yes, it is; I think it should have been adjusted to the Z3 style guide before going into the core, with BBB to the old API, for ZC's beenfit18:18
ignasphiliKON: well it is just that more difficult to use, at least for me18:18
srichterphiliKON: as a Python newbie, don't ask how many times I wrote: get_attr and getkey18:19
*** dobee has quit IRC18:19
philiKONgetkey must be new, i don't know that one yet ;)18:19
* ignas is a conformity nazzi, thinking that there should not be more than 1 coding style per project18:19
srichterphiliKON: I meant haskey18:20
J1mso srichter, what you're saying is that you want to maximize the number of Zope apis that are hard for Python programmers to remember. ;)18:20
philiKON(has_key is going away, by the way)18:21
srichterJ1m: no, I am saying: when using the zope namespace those rules count and if you use zc PEP8 counts (partially) and if you use stdlib PEP8 counts18:22
srichterthese are at least large enough blocks to keep apart18:22
J1mah, I can live with that.  I knew it was a mistake to move formlib to the zope namespace.18:23
srichterbut if you have to keep mappings in mind for zope.* packages, then things become very hairy18:23
* faassen sees this as a mandate to startMixing_naming_styles_ToGether_wildly.18:23
srichterJ1m: ok, so we have some common ground here: zope.* namespace follows Zope 3 style guide; other namespaces can choose, but must be consistent?18:24
faassenI think we should switch to hungarian notation, mylist_sstl (list of string, string tuples)18:24
faassensorry, I'll shut up again. :)18:24
philiKONfaassen, btw, that's an unfortunate misinterpretation of the Hungarian notation18:25
ignasfaassen: sys hungarian or apps hungarian ?18:25
philiKONfaassen, joel spolsky has a nice write up of the *good* hungarian notation18:25
philiKONignas, is sys hungarian the good kind?18:25
mgedminhey, let's prefix all names by zope_18:25
ignasapps is18:25
philiKON+1 to apps hungarian :)18:25
srichterJ1m: btw, I think this is also the fairest solution for existing projects that have adapted any style guide, such as schooltool, tiks, zc, lovely, z3c, etc18:26
ignasphiliKON: not much places it applies to Zope3 or web apps18:26
* Theuni switches to hungrian18:26
philiKONignas, true18:26
philiKONignas, or in object orientation, for tha tmatter18:26
*** whit has joined #zope3-dev18:27
ignasphiliKON: have you tried doing some graphics programming ? :) no objects can save you from relative coordinates and absolute coordinates and let's say map coordinates but apps hungarian solves that nicely ;)18:27
*** alecm has joined #zope3-dev18:28
philiKONincidentally, i used apps hungarian w/o knowing that i did already18:28
*** scherand has quit IRC18:30
*** zagy has quit IRC18:36
* benji is hungry18:41
* Theuni too, that's why i switched to hungrian (or hungryan?) notation18:41
*** WebMaven has joined #zope3-dev18:44
* philiKON is heading home soon. *mampf*18:47
*** kobold has quit IRC18:53
*** nathany_ has quit IRC18:57
*** philiKON has quit IRC18:59
*** torkel_ has quit IRC19:01
*** projekt01 has quit IRC19:03
*** rocky is now known as rocky|away19:04
*** torkel_ has joined #zope3-dev19:04
*** nathany has joined #zope3-dev19:17
*** batlogg has quit IRC19:19
*** BjornT has quit IRC19:20
*** BjornT has joined #zope3-dev19:21
*** romanofski has quit IRC19:23
*** zagy has joined #zope3-dev19:23
*** MJ has quit IRC19:24
*** ignas has quit IRC19:35
SteveAWebMaven: ping19:40
*** zagy_ has joined #zope3-dev19:50
srichterSteveA: did you notice that I checked in a feature that you wanted for years over the weekend? :-)19:51
*** projekt01 has joined #zope3-dev19:51
SteveAI didn't notice that.  What did you do?19:51
*** philiKON has joined #zope3-dev19:51
*** gintas has quit IRC19:56
srichterSteveA: it allows you to "define" component registries in ZCML and use them as bases for local sites19:59
*** zagy has quit IRC19:59
srichterSteveA: effectively, local component definitions via ZCML19:59
*** nathany has quit IRC20:01
*** batlogg has joined #zope3-dev20:03
SteveAsrichter: that's very cool.  so, whenever a piece of zcml is parsed, part of the context is what "local site" it is going to configure.20:04
srichteryou basically create a "IComponents" instance (I provide a class that effeciently pickles)20:04
srichterthen all you have to say is:20:04
srichter<registerIn registry="">20:05
srichteryou also have to register the registry as an IComponents utility20:05
srichterfor the standard local site manager I have created a view that let's you select the registries that should be available as bases to this local one20:06
*** povbot has joined #zope3-dev20:10
philiKONsrichter, i would prefer <configure registry="name_of_the_registry_as_utility">...</configure>20:10
srichterphiliKON: as I said in the docs, this is a trivial change20:11
*** volvox has quit IRC20:11
srichterphiliKON: the goal was to develop a package that I could use now while it is still reviewed for core inclusion20:12
philiKONsrichter, the actionsproxy also seems hackish ;)20:12
philiKONbetter to extend handler20:12
srichterphiliKON: the reason I decided against a name is the same why we do not use layer names anymore :-)20:12
philiKONbut you also already mention taht ;20:12
*** xenru has quit IRC20:12
*** alecm has quit IRC20:12
SteveAsrichter: it looks cool.  I'll have to try it out sometime.20:13
philiKONsrichter, well, i think it'll have to be a name20:13
philiKONor, it rather should20:13
srichterextending handler has one big problem: What if someone does not use it?20:13
*** xenru has joined #zope3-dev20:13
*** ChrisW has joined #zope3-dev20:13
ChrisWhi all20:13
ChrisWzope.interface and zope.testing are seperately released, right?20:13
srichterI considered many alternatives, but at the end I decided to go with the safest20:14
philiKONmost zope.* packages are available as eggs20:14
ChrisWif so, where do you download the seperately available packages?20:14
ChrisW(I'm writing documentation for twiddler)20:14
ChrisWDNS error for me ;-)20:14
Theuniaeh .. .server not found ... :)20:14
philiKONthe last url i gave you should work20:15
srichterif we assume that all component registration is done via handler(), then we can make a lot more simplification with respect to non-registering global changes (see end of README)20:15
Theuniyup, does.20:15
ChrisWso and right?20:15
philiKONChrisW, yeah20:15
ChrisWthese just need untarring onto the python path?20:15
philiKONsrichter, i don't think we should assume handler, but i guess we could assume that directives iether use handler, or they register things with the right registry themselves20:16
ChrisWor maybe "python install"?20:16
philiKONChrisW, dude, it's eggs20:16
srichterphiliKON: I would be careful20:16
philiKONChrisW, read about installing eggs20:16
ChrisWphiliKON: don't want eggs right now20:16
ChrisWone too many things and my head will explode :-(20:16
philiKONthat's fine20:16
srichterzope.component.provideUtility() and zope.component.provideAdapter() is very tempting20:17
philiKONthat's why they should go away :)20:17
ChrisWI know "eggs iz good", but I'm just looking to write quick docs for people who want to install these two when they don't have zope around...20:17
srichterthat's a different story20:17
philiKONChrisW, getting those eggs *is* the easiest way20:17
philiKONChrisW, eggs are nothing magical20:18
philiKONChrisW, they contain the python source in them20:18
srichterI have to go with what we have now; if the requirements change, the implementation can/will change too :-)20:18
ChrisWokay, so with a normal python 2.4 install, how would I install the above?20:18
philiKONChrisW, man, just get one of these eggs, unzip it and you'll find out20:19
philiKONsometimes you're worse than the newbies you yell at all the time ;)20:19
ChrisWthere's only 2 of them20:19
philiKONcurrently ;)20:19
ChrisWI don't mind newbies, which beno professes not to be20:20
ChrisWand I have an irrational fear of eggs20:20
srichterChrisW: you are not alone; me too :-)20:20
ChrisWam I supposed to download the .tgz or the .egg?20:20
philiKONeggs are *nothing* special20:20
philiKONthey're just zip files20:20
philiKONwith python packages in them20:20
ChrisWsrichter: okay, how would you advise installing zope.interface and zope.testing in a non-zope environment?20:20
srichterChrisW: eggs, really :-)20:21
ChrisWphiliKON: then why won't you answer my question? ;_)20:21
srichterChrisW: I probably do not trust them for different reasons :-)20:21
philiKONChrisW, because i'd have to spell everything out for you that you could easily find out yourself by just download the suckers and untaring them20:21
ChrisWokay. .tgz or .egg? I'm guessing .tgz by your phrasing...20:21
philiKONdoesn't really matter.20:22
philiKONsure, .tgz will work20:22
philiKONeggs are nothing special. they're probably also not the great saviours that everyone things they are.20:22
ChrisWwtf is a .egg?20:22 file with the wrong extension?20:22
philiKONthey're just directories (or zipped direcotries) that are on your pythonpath20:22
philiKONas said, either zipped directories or just directories that can be on your pythonpath20:23
ChrisWoh yay, I love opening an INSTALL.TXT to find "now just install all this other crap before you can use the thing you downloaded" :-(20:23
philiKONi todl you to get eggs20:23
philiKONeasy_install loads all that other crap20:24
philiKONbut you wouldn't listen20:24
ChrisWI'm referring to easy_install20:24
philiKONso don't complain :)20:24
ChrisWand setuptools20:24
philiKONhuh? installing setuptools isn't any more complicated than installing another python package20:25
ChrisWis there an example of what to put in a readme.txt for "how to install zope.testing and zope.interface"?20:25
mgedmineasy install is like: just run me as root so that I can download arbitrary python code from the public internet and execute it with root privileges on your machine without performing any cryptographic authenticity checks on the downloaded code20:25
ChrisWmgedmin: exactly!20:25
philiKONmgedmin, umm... yeah, by default20:25
mgedminor maybe that's ez_setup.py20:25
ChrisWpersonally, I don't give a monkeys, 'cos I'll only be using twiddler in zope20:25
ChrisWjust trying to be a good python citizen here20:26
Theunieasy_install is the way20:27
ChrisWso what do I write?20:27
ChrisWsince I know noooothing about easy_install ;-)20:28
*** rocky|away is now known as rocky20:33
ChrisWwhere's the five channel again?20:42
* ChrisW wants to know how the work to allow Zope 2 products to live in lib/python is going...20:42
philiKONChrisW, #z3-base20:43
philiKONyou want to talk to rocky20:43
*** alecm has joined #zope3-dev20:44
*** faassen has quit IRC20:50
*** mkerrin has quit IRC20:51
*** philiKON has quit IRC20:53
*** philiKON has joined #zope3-dev20:56
* benji wonders what the #z3-base channel is for.20:56
TheuniChrisW: you download easy_install and run it once (that installs setuptools) then you run "easy_install packagename"20:56
Theuniif you point it to the egg file it should install that, if you just tell it the package name (e.g. "zope.interface") it will start wandering around on the net trying to get it20:57
benjiahh, thanks (now if that page'll load)20:57
philiKONbenji, worsk for me20:58
ChrisWheh, I jsut said "if yer not using zope, you'll need eggs, here's the instructions, you're on your own"20:58
*** ChrisW has left #zope3-dev20:58
*** gintas has joined #zope3-dev20:59
*** gumpa is now known as gumpa-gone20:59
TheuniphiliKON: i think i found that zope.component.provideUtility is *not* a convenience spelling for zope.component.getSiteManager().provideUtility ....21:01
philiKONi never said it was21:01
philiKONor did i?21:01
Theunizope.component.provideUtility(component, interface=optional) is a convenience spelling for:21:01
Theuni  zope.component.getGlobalSiteManager().registerUtility(21:01
Theuni      component, interface=optional).21:01
philiKONif i said otherwise, i was wrong21:01
TheunigetGlobalSiteManager it is?21:01
philiKONactually, i just looked at my email21:02
Theuniwhy not getSiteManager?21:02
philiKONi was right. i said getGlobalSiteManager21:02
philiKONbecause the convenience api specifically adresses the global registries21:02
Theuniexcept when it doesn't, right?21:02
philiKONit always adresses the global ones21:02
Theunibecause zope.component.getUtility _does_ retrieve local registrations when you make it do21:03
philiKONhmm, well, eyah21:03
philiKONthe get* things behave differnetly than the provide* things i guess21:03
philiKONanother reason to kill the provide* things21:03
* Theuni get's very sad again21:03
Theunithis api needs cleanup to get less confusing21:03
* philiKON is unsure what to do in his book now21:04
Theunithere you go :)21:04
Theuniif i solve this problem i have right now, i can give you some input on my experience from the last days21:04
* philiKON wonders what Theuni is baking up21:04
Theuniworld domination and two milky ways21:05
philiKONah, cool21:05
Theuni(the chocolate bar, not the huge pile of stars)21:05
philiKONJ1m, i guess i really should not use the provide* convenience API in the book, eh?21:05
philiKONJ1m, or do you think until we solve the site manager nomenclature, i should?21:06
Theuniwhich i consider is darn sad because it's much shorter to spell then zope.component.getSiteManager().*21:06
*** MJ has joined #zope3-dev21:06
philiKONwell, yeah21:06
philiKONusually you make multiple registrations anyhoo21:06
philiKONit's just another lime21:06
J1mphiliKON: I dunno.21:06
philiKONsm = z.c.getGlobalSiteManager()21:06
philiKONJ1m, me neither :(21:06
Theuniit's boilerplate21:06
Theuniat least i found something that should be removed, all the provide* names are declared twiced in the global registry21:07
J1mI think using the provide functions is fine for now.21:07
Theuniso the first declaration can probably just be deleted21:07
philiKONJ1m, ok. i just want this 2nd edition to stick around a bit longer :)21:07
J1mwell, maybe we should stop deprecating things so fast. :)21:08
TheuniJ1m: you noticed that provide* will do  global registrations but the get* will work against local site managers?21:08
Theunidoes it have to be that way?21:08
Theuniit's counter intuitive IMHO21:08
J1mIt does have to be that way.21:08
J1mprovide is just a shorthand21:08
J1mIt is just a convenient way to cram things into the global registry,21:09
*** mgedmin has quit IRC21:09
Theuniso can we have a convenient way to cram things into the local registry?21:09
J1mIt's fine most of the time to look up something from an implicitly defined registry.21:09
J1mI don't think you'd every want to be implicit when regsitering things.21:10
J1mThere is an api for getting the local site manager.21:10
TheuniWell, for me, I'm missing the elegance here. The spellings aren't symmetrical, so they're harder to remember.21:11
TheuniI'll try to shut up for a while on that topic and work with the current situation for a while and speak up later again.21:16
Theunii wonder if we should have multiple "severities" of backwards compatibility21:26
Theunithen we could deprecate an API but still have it around for more than two releases21:26
benjiwe could do something like Python does, for very important APIs deprecate them first without a warning, start issueing a warning at some point, then remove it after that21:29
benji(s/without a warning/with an automatically silenced warning/)21:29
*** natea_ has quit IRC21:32
*** dobee has joined #zope3-dev21:34
Theunisomething along those lines21:34
*** niemeyer_ has joined #zope3-dev21:35
*** dunny has joined #zope3-dev21:40
*** niemeyer has quit IRC21:43
*** niemeyer has joined #zope3-dev21:44
*** J1m has quit IRC21:54
*** J1m has joined #zope3-dev21:55
*** natea has joined #zope3-dev21:55
*** romanofski has joined #zope3-dev21:56
*** xenru|clone has joined #zope3-dev22:00
*** alga has quit IRC22:00
*** torkel_ has quit IRC22:03
*** niemeyer_ has quit IRC22:04
*** torkel_ has joined #zope3-dev22:08
*** philiKON has left #zope3-dev22:14
*** philiKON has joined #zope3-dev22:14
*** xenru has quit IRC22:17
*** dunny has quit IRC22:26
*** gumpa-gone is now known as gumpa22:31
*** natea_ has joined #zope3-dev22:43
*** zagy_ has quit IRC22:45
*** ktwilight has quit IRC22:45
*** ktwilight has joined #zope3-dev22:46
*** natea has quit IRC22:49
*** dobee has quit IRC23:13
*** natea has joined #zope3-dev23:22
*** Borax has joined #zope3-dev23:27
*** natea_ has quit IRC23:29
*** Borax has quit IRC23:32
*** gintas has quit IRC23:32
*** timte has quit IRC23:40
*** romanofski has quit IRC23:42
*** natea_ has joined #zope3-dev23:45
Theunidoes anybody know what zope.pagetemplate.pagetemplate.TemplateUsage is for?23:46
*** dunny has joined #zope3-dev23:47
philiKONTheuni, that shit is still in there?23:49
philiKON"usage" got ripped out ages ago23:49
Theunilooks like cruft23:49
Theunino interfaces, no docstrings, no tests, no comments23:49
Theunii just stumlbed over it because it's still in the namespace that is generated for page templates23:50
Theuniand one or two classes associated with it23:50
Theunii'll try to rip it out and see if something breaks23:50
Theunithere are tests still referring to it23:51
Theuniso yes, something breaks23:51
*** natea has quit IRC23:52
*** torkel_ has quit IRC23:56

Generated by 2.15.1 by Marius Gedminas - find it at!