IRC log of #zope3-dev for Thursday, 2005-02-17

*** mexiKON is now known as philiKON00:30
*** mgedmin has quit IRC00:36
*** srichter has quit IRC00:54
*** tarek_ is now known as tarek_pala00:55
*** tarek_pala has quit IRC01:09
*** philiKON is now known as philiZZZ01:17
*** J1m|away has quit IRC01:31
*** srichter has joined #zope3-dev01:32
*** ChanServ sets mode: +o srichter01:38
projekt01srichter, tree status --> just hilight labels in the tree pointing to the topic01:44
projekt01Independent of topic --> build the tree from a global view01:45
projekt01I tried the cookie tree, but it doesn't work, because of the not simply contextual help lookup01:46
projekt01The url of a contextual help topic view is something like: http://localhost:8080/folder/contents.html/++help++/@@contexthelp.html01:47
projekt01The cookie tree doesn't like this url and can't build the url of a child topic if you access a cpntextual help topic view01:48
*** philiKON has joined #zope3-dev01:52
srichteroh, ok02:01
projekt01Hm, yet another javascript tree02:01
projekt01Are you find if we would add package called javascipt with all javascripts in it for to share? (in skins etc)02:03
*** philiZZZ has quit IRC02:06
projekt01Are you find if we would add package02:06
projekt01Are you fine if we would add a package...02:07
srichterright, good point02:07
philiKON-1 javascripts are either part of a skin or part of a browser view; so they really always have places they belong02:08
philiKONwe don't really group other browser resources either, e.g. like icons02:08
projekt01And if I would like to use a javascript, but not like the dependency to a package?02:09
projekt01I mean not all javascripts just some libraries02:09
philiKONi think the amount of javascript in a stock zope 3 should be so limited that this is not necessary02:10
srichteractually, I am on/off, so I am not following too well right now02:10
srichteryou can do what you want ;-)02:10
srichternoone else is working on it :-)02:10
philiKONin fact, if there's a need of grouping javascript libraries etc., then zope3 has too much javascript02:11
philiKONin other words, too much skin02:11
philiKONprojekt01, write a proposal if you feel strongly about it02:11
srichtersince we decided that API doc tool is for Py developers and Online help for scripters/end users, I am going to create a online-help like book using the TXT files from the code02:11
projekt01philiKON, I agree with less then possible javascript02:12
projekt01Perhaps we can add a package later if we think we whant reuse this javascripts02:12
philiKONdo you mean "as little javascript as possible"?02:12
projekt01Stefan, online-help like book? What do you mean?02:13
*** philiKON is now known as philiZZZ02:13
projekt01See you02:13
projekt01srichter, register e.g. authentication/README.txt etc in the onlinehelp02:14
srichterprojekt01: I want to compile all the README.txt and other *.txt files to some sort of outlined book02:19
projekt01Oh, cool. Do you have a tool for that?02:19
srichterno, I am just going to do it using online help topics; I will write a new documentation module for the apidoc tool02:20
projekt01All topics which are registred with help:register?02:21
projekt01I hope I can finish my refactoring tonight and you can see how it looks like tomorrow.02:22
*** deo has quit IRC03:12
*** d2m has quit IRC04:08
srichterprojekt01: no, I am not going to reuse the help:register directive04:50
srichterjust the Python code04:50
*** hazmat has quit IRC04:59
*** `anthony has quit IRC05:08
*** mooded has quit IRC05:23
*** projekt01 has quit IRC05:39
*** `anthony has joined #zope3-dev06:24
tvonWould it be non-trivial to override or create my own version of zope/app/container/browser/ for my application?07:30
*** stub has joined #zope3-dev08:58
*** stub has quit IRC09:21
*** sashav has joined #zope3-dev09:24
*** stub has joined #zope3-dev09:48
*** Theuni has joined #zope3-dev09:52
*** bradb has joined #zope3-dev09:52
*** MalcolmC has joined #zope3-dev11:23
*** d2m has joined #zope3-dev11:25
*** jhauser has joined #zope3-dev11:41
*** hazmat has joined #zope3-dev11:46
*** Theuni has quit IRC11:50
*** apauley has joined #zope3-dev11:58
*** tziade_ has joined #zope3-dev11:59
*** tziade_ is now known as tarek_11:59
*** projekt01 has joined #zope3-dev12:03
projekt01I have problems with ftests, I use zapi.getPath(context), this ends in a TypeError in ftests lke:12:04
projekt01Like: TypeError: Not enough context to determine location root12:04
projekt01Any ideas?12:04
projekt01Zapi.getPath(context) is used in a view <browser:page> class12:05
philiZZZprojekt01, your obj doesn't have a __parent__ that goes all the way up to an IContainmentRoot12:07
projekt01Ok, I check it.12:08
projekt01Yup, you are right, I add it in the ftests directly and commit it with transaction.commit()12:09
philiZZZin ftests it's better to use the addform12:09
philiZZZvia HTTP12:09
philiZZZthat way you get all the necessary context12:10
projekt01Yes I know, but I don't have a addform for help topics. And there is no contextual topic registred as default where I can test.12:11
philiZZZah, ok12:12
philiZZZthat's a special case then12:12
apauleyhi all12:16
apauleyI just connected to a mysql db using mysqldbda12:17
apauleyAre there any docs or examples so that I can see how it is used in a product?12:17
apauleyI'd like to (for example) change the storage component of the messageboard app to mysql12:18
*** tarek_ is now known as tarek_oqp12:26
*** Theuni has joined #zope3-dev12:30
*** stub has left #zope3-dev12:31
*** `anthony has quit IRC12:35
*** `anthony has joined #zope3-dev13:07
*** niemeyer has joined #zope3-dev13:08
*** stub has joined #zope3-dev13:11
srichterapauley: you would need to use sqlos, but you are pretty much on your own13:32
*** vlado has joined #zope3-dev13:44
apauleysrichter: thanks, I'll google a bit for sqlos13:56
*** stub has quit IRC14:03
*** stub has joined #zope3-dev14:16
*** hazmat has quit IRC14:18
*** srichter has quit IRC14:22
*** vlado has quit IRC14:43
*** srichter has joined #zope3-dev15:15
*** ChanServ sets mode: +o srichter15:15
*** stub has quit IRC15:17
*** stub has joined #zope3-dev15:19
*** __gotcha has joined #zope3-dev15:21
srichterprojekt01: how is th eonline help work coming along?15:22
projekt01The onlinehelp is working, but I have trouble to define the topic structure15:24
projekt01Any sugestion about how the topic titles should be called / thema's15:25
projekt01I suggest something like:15:25
projekt01Developer info15:25
projekt01Style Guides15:25
projekt01Where should we place the README.txt files15:26
srichterI put those into an apidoc documentation module15:26
srichtercalled bookmodule15:26
srichterSo under apidoc you will have something like "Book"15:26
srichterand I'll define the structure there15:27
srichterI will create a directive called apidoc:bookchapter (which is effectively a help:register directive for the API doc; I am reusing all that code)15:27
projekt01Uhaa, then I have definitive to checkin my changes15:28
srichterwell at some point, yes15:28
projekt01Ok, I clean up the help topic registration/order and will checkin the refactoring.15:29
srichterdon't forget tests15:30
*** sashav has quit IRC15:45
srichterprojekt01: in the online help topic class, the ending *.txt is interpreted as plain text; can we change this to be ReST?15:47
*** Jim7J1AJ1 has joined #zope3-dev16:03
*** Jim7J1AJH has quit IRC16:04
*** Awayblia is now known as Workblia16:10
*** gintas has joined #zope3-dev16:15
srichterprojekt01: yipee, I have the book protoype working; I never realized it would be so easy :-)16:21
projekt01srichter, what kind of output format is it.16:31
srichterwhat do you mean?16:31
srichterIt is basically a special online help16:31
projekt01srichter, Text format: You can registrer own HelpTopic classes for rendereing your files.16:31
srichterthe txt files are oputputted as HTML16:32
srichterahh, cool16:32
projekt01There are Topic classes for .stx .rest .txt and .pt16:32
projekt01They dosn't have to provide a specific file ending16:32
srichterany Emacs gurus here?16:46
*** tvon has quit IRC16:57
srichterprojekt01: are you there?16:59
apauleysrichter: I'm not an emacs guru (I'm more used to vi), but I have recently made an effort to learn it17:04
*** projekt01 has quit IRC17:04
*** projekt01 has joined #zope3-dev17:04
apauleyI've got the oreilly book17:04
srichterI need to find all txt files in a tree, except for those that contain ".svn" in the path :-)17:05
srichterapauley: and I want to be able to click on them to open them17:05
*** Jim7J1AJ1 has quit IRC17:05
*** Jim7J1AJH has joined #zope3-dev17:05
*** sashav has joined #zope3-dev17:07
*** J1m has joined #zope3-dev17:09
projekt01srichter, I have a problem with naming in onlinehelp the defines help as a global name, butt we have also a folder called help. I need to make this folder to a package now.17:09
apauleyIs that .svn files or .svn dirs?17:09
srichterprojekt01: renaming the global varname from help to something else should be no problem, since it is only used internally17:10
srichterapauley: .svn dirs17:11
projekt01srichter, OK. Do you agree that just the help:register has to be BBB17:11
apauleyI assume you're using emacs' dired on a svn repository?17:11
srichterprojekt01: yes17:12
projekt01Oh, nice17:12
srichterapauley: yes17:12
apauleysrichter: try typing C-x C-d17:17
apauleyit should give you an ls -F listing instead of ls -l17:17
apauleynext you need to choose a dir to list17:17
apauleyo, i see. this is only a listing, not dired17:19
apauleyso you cant click on a file17:19
srichterright, and I also do not get a list of all txt files17:19
srichterI basically want a find output where I can click on the files17:19
*** sashav_ has joined #zope3-dev17:23
*** sashav_ has quit IRC17:26
*** sashav_ has joined #zope3-dev17:27
apauleysrichter: I've found something that might be of help17:30
apauleyin ~/.emacs: (setq dired-listing-switches "-l")17:30
srichterbut it still does not help me finding the files in the tree17:30
srichterapauley: there is no need to waste your time on this :-)17:31
*** sashav has quit IRC17:31
*** sashav_ is now known as sashav17:32
*** stub has quit IRC17:35
*** tvon has joined #zope3-dev17:35
*** stub has joined #zope3-dev17:36
apauleysrichter: ok, I better continue working :-)17:41
srichterapauley: yes, thanks a lot anyways! :-)17:42
*** sashav has quit IRC17:56
*** tvon|x31 has joined #zope3-dev17:57
*** tvon has quit IRC17:57
*** apauley is now known as apauley|away18:04
*** mgedmin has joined #zope3-dev18:12
*** hazmat has joined #zope3-dev18:18
*** benji_york has joined #zope3-dev18:19
*** tvon|x31 is now known as tvon18:30
*** tvon has quit IRC18:39
*** tvon has joined #zope3-dev18:39
*** regebro has joined #zope3-dev18:42
*** tarek_oqp is now known as tarek_18:48
SteveAJ1m: the issue with the python profiling library is a vindication of the zope committer's process / agreement19:15
SteveAin a schadenfreude kind of way19:16
* SteveA hopes no actual german speakers are around to correct spelling19:16
srichterit's correct :-)19:16
*** hazmat has quit IRC19:17
*** tarek_ has quit IRC19:24
*** benji_york_zope has joined #zope3-dev19:26
*** benji_york has quit IRC19:27
* mgedmin wonders what 'schadenfreude' is19:29
SteveAit is the quality of taking satisfaction in another's misfortune19:30
*** WebMaven_ has joined #zope3-dev19:41
*** MalcolmC has quit IRC19:43
*** Theuni has quit IRC19:44
* Workblia was hit with a stick, a zope 3 widget stick19:49
*** zagy has quit IRC19:59
*** __gotcha has quit IRC19:59
*** __gotcha has joined #zope3-dev20:01
*** hazmat has joined #zope3-dev20:25
*** gintas has quit IRC20:45
srichterprojekt01: are you there?20:47
*** hazmat has quit IRC20:52
*** tvon has quit IRC21:01
*** niemeyer has quit IRC21:05
*** tvon has joined #zope3-dev21:10
*** niemeyer has joined #zope3-dev21:13
J1mSteveA, interesting. I haven't been paying close attention to that. Perhaps I should take a look.21:13
projekt01srichter, yes21:17
srichterprojekt01: how is your stuff coming along?21:18
projekt01In about 1 hour21:19
projekt01What do need to do with?21:19
srichterwell, my new book module uses online help21:20
srichterand I do not want to kill your checkin by being first21:20
projekt01Ok, what do you use from the package21:20
srichterprojekt01: see
srichterbut I want to use your new factory code21:21
srichterso that I can always force it to use restructured text21:21
projekt01Do you mean the link Book?21:23
SteveAJ1m: the issue is that there is a module in the standard library which isn't provided under the python licence.  The licence is permissive but not "open source definition" free.21:24
SteveAThe problem is, it says "this code is free for use with python only"21:24
SteveAand an essential freedom is being able to port software to other languages and systems21:24
SteveAthere are also patents involved21:24
srichterit is basically an online help21:25
SteveAthe original author does not hold copyright.  the company he worked for was acquired quite a while ago.  the copyright probably is with Disney corp, but no one really knows where the paperwork is21:25
SteveAnasty expensive lawyer stuff, probably21:25
projekt01srichter, it looks great.21:25
srichterprojekt01: thanks21:26
DamasceneSteveA:  yeah lawyers are expensive.  ;(21:26
J1mand now, finally, they want us to sign contributor agreements.21:26
SteveAso, ubuntu and debian have cut it out of their packaging of the standard library, and put it into a separate package in the "non-free" section of the package database21:26
*** gintas has joined #zope3-dev21:26
SteveAbecause they promise their users that stuff in the main distro is properly free21:26
SteveAbut, this breaks the "you have the standard library available when you have python" ideal21:27
SteveAso, i have to get my sysadmin to install a non-free less supported package onto the integration box, because the tests depend on profile.21:27
J1mI think that idea is flawed anyway -- Python needs a packaging system21:27
SteveAalthough, it is a trivial dependency unless you run the profiling test targets21:28
J1m"Batteries included" has been an excuse/crutch for not solving the packaging problem.21:28
SteveAdon't tell ubuntu -- they'll take distutils back out of the standard python package!21:29
SteveAubuntu did, and debian does, separate out distutils from the rest of the python standard library.21:38
SteveAthis was confusing to users who thought they had what is generally known as "python 2.3", for example.21:38
SteveAthey try to run software that says "works with python 2.3", like zopeX321:38
SteveAand it gives them import errors21:39
J1mdistutils is core21:39
J1mit's one of the few things that *should* always be included21:39
SteveAthe argument from debian is "you should install software using your OS's packaging system, unless you're a software developer.  If you're a software developer, you should install the python-dev package"21:40
srichterbut how realistic is that?21:40
SteveAthe python in the ubuntu warty release lacks distutils.  the python in the forthcoming ubuntu hoary release includes it.  debian still lacks it, as far as i know.21:40
srichterDebian will never provide a DEB package for ever Python pacakge that supports distutils21:40
SteveA`anthony had an idea about that21:41
SteveAone of ubuntu's goals is to get darn close to providing a package for everything python out there21:41
srichterbut you cannot guarantee this21:42
srichterwhat if I have a proprietary Python package21:42
SteveAthen you are responsible for providing the .deb21:42
srichterit is so easy for me to write: Usign distutils do this and this and that21:42
srichterthat is an unacceptable burden to any application developer21:42
SteveAi think someone will write a "make me a .deb from this distutils-friendly python package" tool21:43
srichterdistutils works cross-platform, I just have to do it once21:43
SteveAit is already quite straightforward21:43
srichterso wat if RedHat, Mandrake, ... claim the same thing?21:43
srichterThen I have to provide a gazillion packages for one release21:43
* mgedmin is one of those folks that think "if it isn't in debian, it doesn't exist"21:44
SteveAthat's `anthony's answer to all this21:45
projekt01... I'm looking for the flag - without ZEO tests on windows21:45
J1mprojekt01, at the end of your command: '!ZEO'21:46
J1mI use this now rather than -szope21:46
J1mYou give a pattern21:46
projekt01Oh, great, I was just kidding21:46
J1mIf the patterns starts with ! then we run everything that doesn't match the pattern.21:47
projekt01Cool, thanks21:47
srichterZEO seems to have some memory leaks, every time ZEO is included 75% of my 1GB RAM is taken by Python21:47
*** GaryPoster has joined #zope3-dev21:47
SteveAhi GaryPoster21:47
srichterand it is not released until the tests are done21:47
GaryPosterHey SteveA :-)21:47
J1msrichter, I'd be surpised if there was a leak.21:51
srichterok, I just noticed it21:51
* mgedmin noticed yesterday that 'python zope' needs > 400 MB of RAM21:51
srichterI am not knowlegable enough to make statements ;-)21:51
* SteveA wonders if python will ever learn how to release memory back to the OS21:52
Damasceneso the paris sprint seems like it is geared to give zope3 a CMS framework?21:52
DamasceneSteveA:  hmmm darn it can't?  sniffles.21:52
SteveAnope.  it's an issue for long running processes that have leaks, or periodically need to load a lot of state.21:53
projekt01Ran 6775 tests in 916.570s21:53
Damascenedarn, that's really unfortunate.  you would think they would have resolved it as a key issue.21:53
J1mDamascene, It's complicated.21:54
J1mPython often does release memory21:55
* SteveA is surprised to hear that21:55
J1mThat doesn't mean the OS will actually make it available.21:55
jhauserI can support this from my scientifc working21:56
J1mHm, I may be wrong about that, come to think of it.21:56
J1mThis has changed over time.21:56
jhauserbut is a subtle thing, which malloc is used21:56
J1mIt also depends on which objects you are talking about.21:56
SteveAtim would know, i guess21:56
SteveAat least, on windows21:56
jhauseron windows it does not work21:56
SteveAand most likely on everything from crays to palms ;-)21:56
jhauserbut there is compile option to use the pymalloc21:57
J1mIt's easy with simple programs to see Python allocate and release memory.21:58
J1m>>> s = 'x' * 10000000021:58
J1m>>> s2 = 'x' * 10000000021:58
J1m>>> s=s2=''21:58
J1mI think Python has special handling for smallter objects.21:58
SteveAi thought pymalloc was the default nowadays21:59
J1mI think it manages them in free lists and these free lists don't get smaller.21:59
jhauseryes, but it's a question of the os if it can reuse the memory inbetween21:59
SteveAi guess the gc doesn't compact, as that would change memory addresses, and make the id() implementation harder21:59
J1mPython objects never move in memory22:00
projekt01I have many deprecation warnings on startup since I updated from the trunk?22:02
benji_york_zopeprojekt01, yes you do.22:03
projekt01It's not my fault?22:04
projekt01I just prepare my checkin22:04
J1mIt's benji_york_zope's fault22:04
*** niemeyer has quit IRC22:05
projekt01Ah, should I wait with commiting my changes?22:05
J1mIt would be nice to clean them up first.22:05
benji_york_zopeNo it's not!  I think it's Gary's fault.22:06
J1mIt shoukdn't be all that hard.22:06
J1mIt shouldn't be all that hard.22:06
*** jhauser has quit IRC22:07
projekt01It's hard to find, the traceback reports nothing useable22:07
projekt01Just, perhaps it's there to find22:08
J1mAre you saying that there are deprecation warnings for code that's not yours?22:09
projekt01I don't know where the deprecation warnings comes from22:09
projekt01trunk\src\ZODB\ DeprecationWarning: UtilityRegistration: This class has been moved to
projekt01I see just many of this messages22:10
J1mYou can ignore that22:10
projekt01But they where not there before, I think.22:11
J1mNo, it's because your database has pickles with deprecated class names.22:11
J1mYou don't get warnings when running the tests, right?22:12
srichterthat's odd22:12
projekt01But this must be happen the last couple hours.22:12
J1msrichter, it's not odd at all.22:12
srichterwhy? Deprecation warnings during a test run are odd22:13
J1msrichter, he isn't getting warnings when running tests. (I hope)22:14
srichteroh, then its ok22:15
srichterjust remove your Data.fs22:15
srichterand try again22:15
projekt01I removed the data.sf and it's OK now.22:17
projekt01No there are no deprecation warnings during tests22:17
tvonis zope.schema an appropriate place for an Image field?22:24
tvonmeaning if I wrote one and sent in a patch would anyone give it a second glance?22:24
*** GaryPoster has quit IRC22:25
tvonit would add a PIL dep22:25
J1mIt should be in a separate package.22:26
* tvon nods22:26
J1mProbably it's own little root project in the repo22:26
*** stub has quit IRC22:29
*** GaryPoster has joined #zope3-dev22:31
*** deo has joined #zope3-dev22:33
benji_york_zoperejj, you there?22:37
*** `anthony has quit IRC22:39
*** `anthony has joined #zope3-dev22:40
*** bskahan has quit IRC22:48
*** bskahan has joined #zope3-dev22:49
*** gintas has quit IRC22:53
projekt01srichter, you can checkout and try22:55
srichterI am already updating my code22:55
projekt01Any problems?22:56
srichterI don't think so22:56
projekt01Good news22:56
J1mprojekt01, You checked in files with dos line endings.22:57
J1mI can't start Zope now22:57
projekt01Oh, I will set the svn property22:57
J1m    raise ConfigurationError("Unknown directive", ns, n)22:57
J1mzope.configuration.xmlconfig.ZopeXMLConfigurationError: File "/home/jim/p/jicpac/1/instance/etc/site.zcml", line 7.2-7.3122:57
J1m    ZopeXMLConfigurationError: File "/home/jim/p/jicpac/1/instance/lib/python/zc/fips/configure.zcml", line 8.2-8.3422:57
J1m    ZopeXMLConfigurationError: File "/home/jim/p/jicpac/1/instance/lib/python/zc/intranet/configure.zcml", line 23.2-23.4622:57
J1m    ZopeXMLConfigurationError: File "/home/jim/p/jicpac/1/src/zope3/src/zope/app/authentication/configure.zcml", line 73.2-73.3022:57
J1m    ZopeXMLConfigurationError: File "/home/jim/p/jicpac/1/src/zope3/src/zope/app/authentication/help.zcml", line 5.222:57
J1m    ConfigurationError: ('Unknown directive', u'', u'register')22:57
srichterstrange, works for me like a charm22:58
J1mDid something get added to package-includes?22:58
projekt01But stefan was to early22:59
projekt01I change the svn props now22:59
J1mIs the help namespace new?23:00
J1mprojekt01, is the help namespace new?23:01
srichterJ1m: is help.zcml special for you?23:02
srichterbecause I do not have this file23:02
projekt01No, the help namespace was already there23:02
J1mThat's probably why you can start Zope23:03
J1mTry an svn up23:03
srichterbut I have svn upped23:03
J1mRojer just added this23:04
projekt01Ah yes, this help.zcml is new23:04
srichteroh, I have it now23:04
srichterstill started up23:04
srichterJ1m: try to svn up again23:05
J1mThe problem is that we aren't using the onlinehelp package23:05
J1mBut now depends on it.23:06
J1mIf onlinehelp is optional, then we shouldn't be added dependencies to it.23:06
srichteroh, ok, roger you should delete help.zcml23:06
J1mIf it's not optional, it shouldn't be in package includes.23:06
*** tvon has quit IRC23:06
srichterthis is going to be covered by apidoc anyways23:06
J1mI'm a bit confused23:07
projekt01Ok, should I call this file help.zcml.in23:07
srichterJ1m: we need to decide whether onlinehelp should be optional; I think it should not, which means we need to change that23:08
J1mFred and I were just talking about the fact that we need a way to say: "do this if that feature is available.23:08
srichterI have the same problem with the Book Doc Module23:08
* mgedmin nods23:08
J1mThen, these directives would only go if we had the online-help feature.23:08
srichterright now all the declarations are in apidoc23:08
srichterwhich stinks23:08
srichterprojekt01: just remove help.zcml23:08
J1mWe went through this in X3.0.23:09
projekt01Yeah, like handle help:register if the help package isn't there23:09
J1mWe had to split out zsync and index support into separate packages.23:09
srichterand that was so ugly23:09
mgedminwould something like <optional when_available="help:register"> <help:register ... /> </optional> be acceptable?23:10
J1mThis is probably a job for a less powerful zcml:condition23:11
projekt01Ok, help.zcml is removed23:11
srichterI would like to return the discussion again23:11
mgedmindoes zcml:condition exist today, or is that a reference to some piece of science fiction?23:11
srichtermgedmin: it had been removed due to some complaining (from me)23:12
J1mIt existed briefly. Martijn objected and we removed it.23:12
projekt01mgedmin, something like that would be nice23:12
* mgedmin would also like to have <depends_on package="" />23:12
mgedminwith the meaning "include this file, but only if it wasn't included already"23:13
J1mmgedmin, we discussed that before, but worried that it could cause some weird ordering behavior.23:13
mgedminit sucks to do package-includes dependency management manually23:13
mgedminwouldn't it be nice if zcml were purely declarative and ordering wouldn't matter?23:14
mgedminok, I'll stop making waves now23:14
projekt01But the onlinehelp needs a way to let register each package its help file23:14
J1mwe could probably resurect zcml:condition and dumb it down.23:15
projekt01The onlinehelp lives from the dependency23:15
J1mzcml:condition="have onlinehelp"23:15
projekt01It's pure dependecy to all packages23:15
J1mIt doesn't have to be a hard dependency.23:16
projekt01Yes, something like that, I think it hasn't be a general condition. Just for onlinhelp topics23:16
J1mI was thinking of something general.23:16
* mgedmin sees a parallel with C header files and #include guards23:16
J1mThe thing is, we have this working with TAL.23:17
J1mso, maybe we have:23:18
srichterhello ubuntu user accessing my Z3: currently the book module does not work due to a bug ;-)23:18
* mgedmin noticed23:18
J1m<blah zcml:condition="have onlinehelp" ...>23:18
J1mso we have a have verb that looks for things that have been provided.23:19
mgedminJ1m, I like that23:19
srichteroooh, that sounds cool23:19
J1mLater, we can add an "if" verb that takes a dotted name of a function to run:23:20
J1m<configure zcml:condition="if .config.onwindows">23:21
* mgedmin predicts that people will not like running Python code from ZCML conditions23:21
J1mwhy not?23:22
projekt01Ah, the function is a check for something.23:22
mgedminwasn't that the reason why people objected to zcml:condition?23:22
J1mWe has TAL before, but Martijn objected that it was too powerful.23:22
mgedmin"too hard to analyze with tools"23:22
projekt01I like to run code form ZCML23:22
J1mmgedmin, martijn objected to having code *in* zcml.23:23
projekt01We need to find another way for analyze the code23:23
mgedminyou mean, python: expressions?23:23
mgedminah, ok23:23
J1mor even path expressions23:23
srichterI like the verb idea23:24
* mgedmin gotta go23:24
mgedminI like the provides="foo" and condition="have foo" so much, I'd be willing to spend a weekend implementing it23:24
mgedminsee you23:24
*** mgedmin has quit IRC23:24
J1mDang, he left before I could thank him. :)23:25
srichtershoot him a mail23:25
J1mBut I heard him volunteer.23:25
J1mwill do.23:25
projekt01J1M, is your workspace running again?23:26
J1mHopefully, we'll resolve this zcml:condition thing soon, so we can put the help in the packages.23:28
srichterprojekt01: at some point we also have to support ordering of help topics23:36
J1msent a proposal to the list.23:43
srichterreplied :-)23:48
projekt01J1m, there are alphabetic now, but this would be nice23:53
srichterprojekt01: right, but this is not alwasy desirable23:53
srichteras in the case for the book23:53
projekt01Yes, having a order attribute is really hard to support. I think the tree has enough hirarchy which can be used as a order. Or not?23:55
srichtermaybe we just give it an abstract position23:56
projekt01Or we order just each level. Means each level has its own order.23:56
srichterwhich acts like weight23:56
srichterthe smaller the number the higher up it appears23:56
srichterright, each level has its own order23:56
srichterbtw, I really liek your improvements23:58
projekt01Should I add a order attribute which acts like weight?23:58
srichterthat would be great23:58
projekt01Thanks, but I'm not really happy with the tree.23:58
srichterI think we should use a static tree23:59
*** niemeyer has joined #zope3-dev23:59
srichterI do this for the apidoc book and I like it a lot23:59
projekt01It's not possible23:59
projekt01We don't access the topic as a context23:59

Generated by 2.15.1 by Marius Gedminas - find it at!