IRC log of #zope for Monday, 2011-03-21

*** __mac__ has quit IRC00:03
*** alexpilz1 has quit IRC00:06
*** dayne has joined #zope00:07
*** alexpilz has joined #zope00:13
*** daMaestro has joined #zope00:19
*** rogererens has quit IRC00:44
*** shastry has quit IRC00:44
*** J1m has joined #zope00:58
*** rogererens has joined #zope01:00
*** srichter has quit IRC01:10
*** dayne has quit IRC01:12
*** alexpilz has quit IRC01:14
*** alexpilz has joined #zope01:14
*** gwik1 has quit IRC01:22
*** evilbungle has quit IRC01:35
*** srichter has joined #zope01:53
*** J1m has quit IRC01:57
*** alga has quit IRC02:13
*** J1m has joined #zope02:13
*** rogererens has quit IRC02:19
*** J1m has quit IRC02:25
*** giampaolo has quit IRC02:42
*** dayne has joined #zope02:50
*** davisagli has quit IRC03:14
*** davisagli has joined #zope03:15
*** daMaestro has quit IRC03:25
*** dayne has quit IRC03:57
*** allisterb has quit IRC04:16
*** allisterb has joined #zope04:30
*** strichter has joined #zope04:36
*** srichter has quit IRC04:39
*** redir has quit IRC05:11
*** nitrogenycs has left #zope05:14
*** redir has joined #zope05:16
*** sm has quit IRC05:48
*** __mac__ has joined #zope06:46
*** __mac__ has quit IRC06:50
*** tiwula has joined #zope07:01
*** __mac__ has joined #zope07:20
*** tiwula has quit IRC07:52
*** hever has joined #zope08:31
*** yvl has joined #zope08:44
*** __mac__ has joined #zope09:04
*** gwik has joined #zope09:07
*** wosc has joined #zope09:13
*** agroszer has joined #zope09:15
*** gwik has quit IRC09:18
*** bigkevmcd has joined #zope09:25
*** planetzopebot has quit IRC09:33
*** planetzopebot has joined #zope09:34
*** d2m has joined #zope09:36
*** Wu has joined #zope10:17
*** sylvain has joined #zope10:19
*** goschtl has joined #zope10:23
*** menesis has joined #zope10:35
*** eperez has joined #zope10:39
*** humanfromearth has joined #zope10:42
*** humanfromearth has left #zope10:42
*** avoinea has joined #zope10:43
*** MrTango has joined #zope10:51
*** rogererens has joined #zope10:55
*** rogererens has quit IRC11:04
*** rogererens has joined #zope11:06
*** Gogo|gui has joined #zope11:08
*** Gogo|gui has quit IRC11:12
*** gwik has joined #zope11:15
*** Directorsppadic has joined #zope11:16
*** Directorsppadic has left #zope11:16
*** Gogo|gui has joined #zope11:17
*** Gogo|gui has quit IRC11:18
*** ccomb has joined #zope11:19
*** Gogo|gui has joined #zope11:20
*** alga has joined #zope11:24
*** Gogo|gui has quit IRC11:24
*** sunew has joined #zope11:30
*** evilbungle has joined #zope11:36
*** evilbungle has joined #zope11:36
*** zagy has joined #zope11:45
*** rogererens has quit IRC11:46
*** J1m has joined #zope11:53
*** thetet has joined #zope11:54
*** J1m has quit IRC12:03
*** nitrogenycs has joined #zope12:35
*** mr_jolly has joined #zope12:41
*** rogererens has joined #zope12:42
*** rogererens1 has joined #zope12:42
*** teix has joined #zope12:49
*** rogererens has quit IRC13:00
*** TomBlockley has joined #zope13:05
*** fredvd has joined #zope13:12
*** dayne has joined #zope13:13
*** Ariel_Calzada has joined #zope13:18
*** thetet has quit IRC13:18
*** d2m has quit IRC13:19
*** zagy has quit IRC13:25
*** thetet has joined #zope13:34
*** svenn has joined #zope13:35
*** d2m has joined #zope13:37
*** nitrogenycs has left #zope13:44
*** fredvd is now known as fredvd|lunch13:47
*** zagy has joined #zope13:52
*** zagy1 has joined #zope13:55
*** zagy has quit IRC13:55
*** nitrogenycs has joined #zope14:03
*** zagy has joined #zope14:05
*** zagy1 has quit IRC14:07
*** svenn has quit IRC14:12
*** BGaddie has joined #zope14:13
*** lucmult has joined #zope14:15
*** bobo_b has quit IRC14:18
CIA-83janwijbrand * r121053 grok/ (3 files in 3 dirs): add value index component14:22
*** MrTango has quit IRC14:26
*** MrTango has joined #zope14:35
*** bobo_b has joined #zope14:46
*** shastry has joined #zope14:49
*** fredvd|lunch has quit IRC14:51
*** menesis has quit IRC14:57
*** mr_jolly_ has joined #zope15:00
*** mr_jolly has quit IRC15:01
*** mr_jolly_ is now known as mr_jolly15:01
*** giampaolo has joined #zope15:12
*** rogererens1 has quit IRC15:12
*** Wu has quit IRC15:13
*** Gogo|gui has joined #zope15:15
*** redir has quit IRC15:16
*** zagy has quit IRC15:17
*** fredvd has joined #zope15:18
*** d2m has quit IRC15:19
*** rogererens has joined #zope15:20
*** d2m has joined #zope15:21
*** pjfd has joined #zope15:31
*** Gogo|gui has quit IRC15:32
*** d2m has quit IRC15:33
*** menesis has joined #zope15:36
*** dayne has quit IRC15:37
*** agroszer has quit IRC15:38
*** Gogo|gui has joined #zope15:40
*** Gogo|gui has quit IRC15:40
*** runyaga has joined #zope15:42
*** runyaga has quit IRC15:42
*** runyaga has joined #zope15:42
*** J1m has joined #zope15:55
*** rogererens has quit IRC16:03
*** benji has quit IRC16:04
*** benji has joined #zope16:07
*** hever has quit IRC16:09
*** dayne has joined #zope16:09
*** rogererens has joined #zope16:22
*** hever has joined #zope16:33
*** wosc has quit IRC16:41
*** __mac__ has quit IRC16:44
*** daMaestro has joined #zope16:44
betabug"Chris Withers: the indominable Zope permatroll" hahahaha!16:46
*** redir has joined #zope16:47
*** cpf_ has quit IRC16:48
*** lucmult has quit IRC16:51
*** lucmult has joined #zope16:52
*** hever has quit IRC16:58
*** thetet has quit IRC16:59
*** tiwula has joined #zope17:02
*** alexpilz has quit IRC17:09
koshbetabug: so how is your life going?17:10
betabughmmmm, good question17:10
betabugquite fine, I'd say17:11
*** alexpilz has joined #zope17:11
*** Wu has joined #zope17:12
koshdid you ever play with my blob patch?17:18
betabugnope, sorry17:19
koshI actually updated the code with a better conversion script that you can run from the command line :)17:19
* kosh sets betabug on fire17:19
*** alexpilz has quit IRC17:19
betabugcurrently trying to fix unicode/utf-8 stuff on 2 large apps, to get them into the 2.10+ world17:19
*** bowe has joined #zope17:19
betabugkosh: well, if you keep updating the stuff, I keep delaying my tests ;-)17:19
koshI only added a script that can be run from zopectl instead of putting a python script inside the zodb17:20
koshit prevents timeout issues that way17:20
koshthe actual code has not changed17:20
betabugyeah, and probably nicer to run when you have ZEO anyway17:20
koshyup17:20
koshyeah I will have to convert stuff to unicode later also, will be an insane ammount of work17:21
betabughmm, dunno, acceptable so far17:21
koshand I will just have to tell the cjk people they can't use their own character sets anymore17:21
betabug... but then I haven't really gotten in the bug minefields yet :-)17:21
koshso many strings to convert in so many places17:21
betabugkosh: dunno, maybe they can17:21
betabugI have 1 method that does the actual converting17:22
koshno I have to say no more shift-jis for instance17:22
betabugif you can centralize stuff there, maybe you can switch stuff around17:22
betabugI don't see a reason why not17:22
koshbecause some translators are starting to use characters that don't exist in shift-jis and others are using characters that apparently are not exactly the same character (same meaning though) in shift-jis as utf-817:22
koshand the customer wants to do more combining of languages on a page17:22
betabugright17:22
koshso I need an encoding that can handle chinese and japanese for instance17:23
koshbetter to just decree utf-8 and be done with it17:23
*** nitrogenycs has quit IRC17:25
*** runyaga is now known as run|away17:27
*** alga has quit IRC17:28
betabugok then :-)17:29
betabugutf-8 is fine with me so far17:29
koshwell what I want to do is make everything unicode and see if I can get zope to encode to a chosen encoding as a last step on publishing17:30
betabughmmm, I think that should be possible17:31
betabugas "encode to a default encoding as a last step" is a given17:31
betabugso there sure is a choice too17:31
koshso I would set zope to utf-8  and then make sure my stuff was pure unicode the entire way through17:32
koshthat way if I did need it to be utf-16 instead for instance that could be done17:32
betabugI think so, yes17:33
betabugwell, as for workload, so far it's just a bunch of:17:38
betabug-        self.title = title17:38
betabug+        self.title = u(title)17:38
betabugand going through the code17:38
betabugchanging all the input parameters, REQUEST.form.get(), getattr() calls17:38
betabugand getattr() is only because there will be still a db full of utf-8 strings17:39
betabugI'll probably write a blog post about this approach tomorrow17:39
betabug(and another one, once all hell breaks loose in tests)17:39
koshmine will end up being more complex then that17:40
koshI will have the system go through the db and convert all existing data to unicode for instance17:40
betabugyeah, I thought about doing that17:41
koshuse character detection when putting data into the system17:41
betabugbut it's a big DB17:41
koshbecause the browser will NOT tell you how it encoded the data that is submitted in a form17:41
koshand it won't necessarily encode it in the same encoding as the page is sent17:41
betabugit's all utf-8 here17:41
koshhow do you know that is what it is when someone hits submit on a page to submit a form?17:41
betabugif someone screws with that, well it's their problem17:41
koshno I mean someones web browser17:42
betabugthe page is set as utf-8, all browsers I've seen so far submit forms with that17:42
betabugif they change the encoding, well, tough luck17:42
koshif someone is using IE on windows then it submits with whatever the current windows character set is17:42
koshso it will submit as shift-jis in japan most of the time for instance17:42
betabughmm, interesting17:42
betabugdoesn't work with Greek character sets like that17:43
betabugor we would have lots of trouble here17:43
koshjust like if a page is set as iso-8859-1 it won't matter because english windows IE will submit cp-125217:43
betabugbut maybe on windows they use utf-8.el_EL really17:43
koshit changes depending on the exact version of windows17:44
betabugaha17:44
betabugbut then, you can't know the difference between is-8859-1 and cp-1252 really, can't you?17:44
koshyou can use the chardet module to get it right most of the time17:45
betabughmmm, dunno17:45
koshif you are only dealing with greek as a language and only modern browsers and only modern versions of windows your solution might work17:45
betabugyeah, but I guess the few times when it really matters, it might not get it17:45
koshif that is not true you are going to run into nasty problems17:45
betabugwell, it's a mix of greek + english + rarely some other european languages17:45
betabugno asian stuff so far17:46
koshasian is pretty much where it falls apart17:46
betabugI can imagine17:46
koshwell utf-8, iso-8859-1, iso-8859-15 and a bunch of other character sets all exactly match on their lower characters which just happens to be what european languages use17:47
betabugyupp17:47
koshso in most cases if you guess wrong on the character encoding actually used it is still correct in the end17:47
betabugyeah, but the differences are going to be in weird spots17:49
TheJesterholy crap people17:50
betabughey TheJester17:50
koshgreetings lifeform17:50
TheJesterhey beta, kosh17:50
TheJesterhow's life amongst the bots d8)17:50
betabugso, what's happening?17:50
koshdoing fine here17:50
betabugamongst the bots and the occasional lost newbie17:51
betabugTheJester: I played around with Pyramid over the last few days17:51
TheJesterWas it fun?17:52
*** sunew has quit IRC17:52
betabugyeah, it's a nice thing and well documented17:52
betabugprobably gotta build something for fun17:53
betabugs/gotta/gonna/17:53
TheJesterprobably get more people with the better name d8)17:53
betabughaha, also with the merger news going around17:54
betabugwell, I was going to play with repoze.bfg "some day"17:54
koshI look at those other projects sometimes17:56
koshhowever I really like the security part of zope, access controls on methods, attributes etc17:56
betabugwell, zope gets a feel of being very "complete" after looking around17:56
betabugbut I still like to look around :-)17:57
koshI like what they did with the mailhost in zope so you can set it to send emails only of no other errors happen in the transaction17:57
koshI want to use that same code if I can to work with salesforce later17:57
koshso that data only gets synced to salesforce if no errors happened in the rest of the transaction17:58
koshit just seems that zope has so much infrastructure to make it easy to get something right17:58
TheJesterThat's pretty easy to setup transaction wise17:58
koshI figured it would be17:58
TheJesterI set up something similar to spawn a thread to do "some other work"17:58
koshI was talking to someone that was doing some django work recently and he said that would be a serious bitch to do that with django but he wanted to figure out a way to do it since it sounded so useful to him17:59
koshit just seems that a lot of things that are very very hard in other frameworks are kind of built in for zope18:00
TheJesterI don't know why Zope gets the bad rap it gets about being too big or too harrd18:00
koshbut this would ensure that my system and salesforce would remain completely insync on a transaction for transaction basis18:01
TheJesterThen I see people embedding HTML inside python...18:01
betabugwell, docs on pyramid make it quite easy to get started18:02
koshit just seems when I look at other projects they are always missing a few things that I use often that would be a pain in the neck to write myself and probably would not get them 100% right18:03
koshbetabug: that is what I find for most of the projects, they are very easy to get started and then that is where it ends18:03
betabughaha, it's always like that18:03
koshbetabug: you are started and now looking at having to get some very complex functionality working with no real support system for it, the simple systems just seem to be simple and stop18:03
betabuganyway, packing up here, cya18:05
koshthe biggest issue I see is that most systems put no thought into security18:06
*** __mac__ has joined #zope18:07
koshit seems like most other systems have more security exploits in a week then zope does in a year, some have more in a day18:07
koshsimple things in zope like every transaction runs as the user for that transactions seems like it is such a pain in the neck for a relational db18:10
*** Ariel_Calzada has quit IRC18:16
*** hever has joined #zope18:18
*** sylvain has quit IRC18:19
*** pjfd has quit IRC18:19
*** hathawsh has joined #zope18:19
*** sm has joined #zope18:21
*** webmaven has joined #zope18:29
*** sm_ has joined #zope18:31
*** sm has quit IRC18:31
*** sm_ is now known as sm18:31
mcdonc_kosh: pyramid is very zopey so it has declarative security, 1-request-1-transaction, mail-on-txn-commit, etc... not that you'll ever use anything but zope, just sayin18:35
mcdonc_TheJester: ahoy18:35
TheJesterHi McD18:35
mcdonc_old home day in #zope18:36
koshmcdonc_: so what are the good and bad points compared to zope then?18:36
TheJesterNormally it's just bots fighting d8)18:36
koshmcdonc_: but at least it sounds like it has a lot more features I use compared to other systems18:37
mcdonc_kosh: no zmi, less featureful traversal, no ttw code18:37
koshah those I use also18:38
*** hathawsh is now known as hath|away18:38
mcdonc_yes, i know you love you some zope and will never use anything else, was just puttin it out there mostly for benefit of others18:39
*** shastry has quit IRC18:39
koshmy problem is not moving to something else, it is something else not having features I use regularly18:40
koshI could deal with the zmi18:41
koshand ttw code could be dealt with if I had a good way to give people that don't understand security very well the ability to write little scripts and let the system take care of security18:41
koshif they write external methods their code would be insecure and it also has the problem of requiring someone to be correct all the time18:41
koshthe restricted environment in ttw code has a lot of security advantages and is a large reason of why zope stands up so well to attacks18:42
*** hath|away is now known as hathawsh18:43
smmorning zopistas18:48
*** TresEquis has joined #zope18:48
*** goschtl has quit IRC18:48
TheJestersm18:48
smkosh: do you have a lot of users writing dtml/pythonscripts ?18:49
koshsome at least, not a lot18:49
koshbut those that do write them write a  fair number of them18:49
smI used to think that zope feature would be quite widely used18:50
koshI know we use it in a major way18:50
koshI even write a fair number of them when something does not need to go on many servers18:50
koshit is just simpler and safer18:50
koshwith external methods/product code you have to be a lot more defensive in your coding18:51
mcdonc_hey sm18:51
mcdonc_we haven't had a job where scripting delegation to untrusted users was important in many many years18:52
mcdonc_although i'm sure it's still useful to you, it's a huge amount of overhead for us18:53
*** Gogo|gui has joined #zope18:57
smmcdonc_: much the same here. I don't remember any of my clients or coworkers getting productive with scripting. I might still use it rarely myself for some quick website hack18:58
smI still think it (sandboxed scripting) is a pretty cool feature and can't think of another framework providing it18:58
smother than browser javascript I guess18:59
mcdonc_yeah it's what dragged me into zope in the first place18:59
koshit is not that the person is untrusted18:59
mcdonc_but the machinery and sacrifices required for it are outweighing its usefulness by a large margin at this point19:00
mcdonc_(for me)19:00
koshit is that expecting them to write the code with no errors and not creating more security problems is a lot to ask19:00
TresEquismcdonc_, we do have a client where scripting via the ZMI is important19:01
TresEquisalthough the only people who do it are me and their most-trusted developer ;)19:01
TresEquisrestart / reload overhead on their production machines would be fatal19:01
mcdonc_"scripting delegation to untrusted users" is a little different than that tho right19:02
TresEquissure19:02
TresEquisI'm not arguing for letting untrusted users even use the CMS, much less script it ;)19:02
mcdonc_hehe19:03
koshthe point is that expecting someone to make no security mistakes all the time is just inviting failure, people are just not that good19:03
mcdonc_well that's the definition of untrusted really19:03
mcdonc_(although "untrusted" is maybe an unkind-sounding way to say it)19:04
koshthe restricted security environment in ttw code is very useful19:04
koshmcdonc_: even your own code would be untrusted sometimes, being able to use a restricted environment for code that should not be touching any outside stuff anyways is a lot safer19:05
TheJesterBeing able to drop and reload Products properly / safely without a restart would be a great boon d8)19:06
mcdonc_given the tradeoff that ttw code cannot really be tested it's a tradeoff.... often known-to-work is more important19:06
koshttw code is useful for that also, no restarts to change code19:07
mcdonc_(at least known-to-work is more important to me on 99 jobs out of 100)19:07
mcdonc_anyway, we both have what we want, so no point in arguing about it19:07
koshwell it should be possible to cover both and do it better19:08
koshlike if we could have a way to declare a function as running in a protected execution environment19:09
koshthen you could edit on the filesystem, you could test on the filesystem and still run in restricted execution19:09
*** Gogo|gui has quit IRC19:09
TheJesterWell you can do that by writing the code to disk inside Python Scripts d8)19:09
TresEquis"oh, you want the *fuzzy* handcuffs!" :)19:09
*** hathawsh is now known as hath|away19:10
TheJesterDear Tres, I would like to sign up to your newsletter d8)19:10
*** hath|away is now known as hathawsh19:11
koshyeah the best woudl be reload on change, edit on filesystem and through the zmi, keep the restricted security, make it easier to test19:12
*** cpf_ has joined #zope19:16
*** avoinea has left #zope19:25
TresEquiskosh, as TheJester points out, that is exactly the pattern for CMF's FSPythonScripts19:26
TresEquisassuming you are running in debug mode19:26
koshhmm I did not see him point that out19:27
CIA-83faassen * r121054 martian/bootstrap.py: Newer bootstrap.19:27
CIA-83faassen * r121055 martian/src/martian/tests/test_all.py: Use standard library doctest.19:27
koshof course I would also have to run in debug mode and also install CMF that does not seem great19:27
*** pjfd has joined #zope19:30
TheJesterThose are coding problems d8)19:30
smdon't need CMF either19:32
*** fredvd has quit IRC19:33
*** evilbungle has quit IRC19:33
koshit would be nice to change the default python script that is part of zope to do that though19:37
TheJesterMonkey Patch d8)19:38
koshthe most recent thing I have worked on for my machines is that OFS.Image and File use blobs instead of strings/pdata19:38
koshgood idea19:38
koshI have a monkey patch for the blob one19:38
TheJesterI did that with ZSQL so you can edit them even with no connector available19:38
koshhttps://github.com/Immudzen/MonkeyOFSImageFile19:38
koshblobs by default made a huge difference19:38
koshI had asked on the zope list if anyone was interested, only two people replied and one of them told me I need to make it a svn patch for anyone to look at it19:39
TheJesterheh19:39
koshso I did that work and sent it to them and never heard about it again19:39
TheJesterI saw the posting,  but, I didn't need it19:39
koshI found it made a truly massive difference19:39
TheJesterI store my images on the FS and use a sendfile monkey patch to serve them d8)19:40
koshI needed a way that customers could still upload files and the system would just transparently put them on the filesystem19:41
koshand my works retroactively on all existing files without any interface changes19:41
koshso I really liked it19:41
koshone on site that was 7GB after this patch it went to 6.8GB in blobs and .2G in Data.fs and it runs a LOT faster19:41
TheJesteryeah small ZODBs make a huge difference19:42
koshwell also blobs are not cached and can be streamed more efficiently19:42
koshI did try to use your sendfile solution with the blobs but I ran into problems trying to figure out how to make that work right19:42
koshbut the existing way I stream already is currently working well enough and about 5x faster then default zodb streaming19:42
*** m8 has joined #zope19:43
TheJesterI should have added resume support to it too19:44
koshmine does handle that19:44
TheJesterbut I didn't really need that at the time19:44
TheJesterNot that it's really any more difficult19:44
TheJesterI had a modded Apaache module too that worked really nicely if your Apache was not on the same host, but, shared a FS with Zope19:45
TheJesterThen Apache would do the heavy lifting instead19:45
koshI dropped apache19:46
koshI moved to nginx19:46
*** hathawsh is now known as hath|away19:46
koshI originally had 4 zope servers in front of 1 zeo19:46
pjfdy19:47
koshand with apache its load balancer would sometimes hand out the wrong content19:47
koshI could verify it was not in the zope end19:47
TheJesterYeah I've seen that19:47
koshthat bug had been opened and closed in apache about 10 times19:47
koshthey keep claiming to fix it and it keeps coming back19:47
TheJesterAnd cache corruption is rife19:47
*** gr0gmint has joined #zope19:47
koshwith nginx I have NEVER seen that happen and nginx is also much simpler to make work19:47
koshso I just got tired of dealing with apache and dumped it19:48
TheJesterfair enough19:49
*** _mup_ has quit IRC19:49
*** _mup_ has joined #zope19:49
*** BGaddie has quit IRC19:50
*** BGaddie has joined #zope19:51
* sm takes note and resolves again to try to kick apache19:54
*** __mac__ has quit IRC20:00
*** gr0gmint has quit IRC20:02
koshsm: it is worth it, nginx is sure nicer20:12
koshnginx made load balancing zope trivially easy, it makes virtual hosting much simpler20:21
CIA-83tseaver * r121056 /zope.component/branches/tseaver-queryAdapterFactory: Add 'queryAdapterFactory' to enable zero-registration patterns.20:23
CIA-83tseaver * r121057 /zope.interface/branches/tseaver-queryAdapterFactory: Add 'queryAdapterFactory' to enable zero-registration patterns.20:23
CIA-83tseaver * r121058 /zope.interface/branches/tseaver-queryAdapterFactory: 'lookup' already gets us what we need.20:23
*** TomBlockley has quit IRC20:24
*** lucmult has quit IRC20:49
*** eperez has quit IRC20:53
*** redir has quit IRC20:59
*** Ariel_Calzada has joined #zope21:00
*** redir has joined #zope21:04
*** lucmult has joined #zope21:04
*** pjfd has quit IRC21:06
*** pjfd has joined #zope21:06
*** menesis has quit IRC21:09
*** rogererens has quit IRC21:11
*** mr_jolly has quit IRC21:13
*** __mac__ has joined #zope21:14
planetzopebotTurtle sprint, day 1 (Reinout van Rees' weblog)  http://reinout.vanrees.org/weblog/2011/03/21/turtle_day1.html21:15
planetzopebotPlone Gets Attention at Pycon 2011 in Atlanta (Plone News)  http://plone.org/news/pycon201121:15
*** giampaolo has quit IRC21:20
*** Gogo|gui has joined #zope21:34
TresEquiskosh, I don't remember seeing your OFS.Image stuff -- which list did you notify, and when (so I can go refresh my memory)21:42
koshzope list21:42
TresEquis-dev or the users?21:42
koshregular zope list21:43
koshMonkey patch for BLOB support in OFS.Image Images and Files   that was the subject21:44
*** __mac__ has quit IRC21:44
*** __mac__ has joined #zope21:45
koshit is a monkey patch that makes it so that OFS.Image Images and Files saves only blobs but it will still read the older string and pdata structures21:48
koshand there is a nice script you can run to convert all existing Images and Files to use blobs21:48
CIA-83menesis * r121059 zope.app.server/src/zope/app/server/ (DEPENDENCIES.cfg zope.app.server-configure.zcml SETUP.cfg): removed obsolete build infrastructure files.21:48
koshthere is even a svn patch against 2.13 sitting in the git repository21:51
*** menesis has joined #zope21:52
*** nitrogenycs has joined #zope21:53
*** Gogo|gui has quit IRC21:54
*** Gogo|gui has joined #zope21:57
TresEquiskosh, I will look into it22:02
koshcool22:03
koshas a monkeypatch it works fine, what I would like eventually to do is get it into zope directly, would just have to have it detect if blob support is enabled22:03
koshit makes such a large difference in speed and memory usage it would be useful by default22:03
*** goschtl has joined #zope22:04
*** nitrogenycs has quit IRC22:04
TresEquiskosh, you will need to lose the GPL for me to land it, though22:05
koshyeah that I can do22:06
koshwhat license should it be under?22:06
TresEquisZPL 2.122:06
TresEquishttp://www.zope.org/Resources/License/22:07
*** Gogo|gui has quit IRC22:07
koshfor some reason I can't get that page to load for me right now22:07
* kosh kicks it a few times22:07
koshhmm zope.org won't even load for me right now22:09
*** Gogo|gui has joined #zope22:09
TresEquisjust loaded for me22:09
koshsitll not loading for me but I just found a copy of the ZPL 2.1 somewhere else, going to change it in a moment22:10
koshokay license changed22:11
koshhowever it does need some more work before it can go into the zope core I think22:11
koshbecause it currently doesn't check if blobs work before it tries to use blobs, which is probably the biggest problem22:12
koshI suppose it is always possible also that someone might have blobs turned on but not want blobs created, darned if I know why though22:12
koshit works for reading it is just that for writing it just tries to make a blob and does not check if blobs work or if it needs to fall back to pdata or string structures instead22:13
TresEquisI created a LP issue for your patch: https://bugs.launchpad.net/zope2/+bug/73969222:13
koshbut if you have any suggestions on how to make the patch better I can fix those things22:13
_mup_Bug #739692: Add blob support to OFS.Image.Image and OFS.Image.File <Zope 2:Confirmed for tseaver> < https://launchpad.net/bugs/739692 >22:13
*** Wu has quit IRC22:13
*** teix has left #zope22:14
*** shastry has joined #zope22:14
TresEquisthe hard part is going to be writing the tests (as usual)22:14
TresEquismaybe I can trick hannosch into doing it :)22:14
koshnot even sure where to start with that, probably have to find out where tests are written for other things in OFS.Image22:14
TresEquisshould be OFS/tests/test_Image.py, or similar22:17
TresEquisbut it's called 'testFileAndImage.py' instead22:18
davisaglialso look at witsch's test setup in plone.app.blob, might help with getting a test db with blob support22:20
davisagli(though probably needs some divorcing from Plone test setup infrastructure)22:20
*** Gogo|gui has quit IRC22:20
*** __mac__ has quit IRC22:20
davisagliI suppose ZODB itself must have something along those lines too22:21
koshyeah so far I don't have much of an idea on where to start but can work on it22:22
koshI probably also have to put back in the stuff for storing pdata/strings in the case where blobs are not currently supported22:23
*** goschtl has quit IRC22:25
*** menesis has quit IRC22:30
CIA-83menesis * r121059 zope.app.server/src/zope/app/server/ (DEPENDENCIES.cfg zope.app.server-configure.zcml SETUP.cfg): removed obsolete build infrastructure files.22:35
CIA-83janjaapdriessen * r121060 groktoolkit/ (buildout.cfg grok.cfg): Move versions from buildout.cfg to grok.cfg in order to be able to pick up the changes from grokcore.view trunk.22:35
CIA-83tseaver tseaver-queryAdapterFactory * r121061 zope.component/src/zope/component/_api.py: Remove unused imports.22:35
CIA-83tseaver tseaver-queryAdapterFactory * r121062 zope.component/src/zope/component/ (6 files):22:35
CIA-83Add 'queryAdapterFactory' and 'queryMultiAdapterFactory' APIs.22:35
CIA-83These functions allow looking up the uncalled factory for an adapter.22:35
CIA-83hannosch 2.12 * r121063 Zope/ (doc/CHANGES.rst versions.cfg): Attempt to fix buildbot test failures in python-gettext22:35
CIA-83erico_andrei * r121064 z3c.form/src/z3c/form/locales/ (pt_BR pt_BR/LC_MESSAGES/z3c.form.po pt_BR/LC_MESSAGES): Adding Brazilian Portuguese translation22:35
CIA-83erico_andrei * r121065 z3c.form/CHANGES.txt: Updates documentation, adding pt_BR translation22:35
CIA-83erico_andrei * r121066 plone.z3cform/ (4 files in 4 dirs): Adding Brazilian Portuguese translations. Thanks to davilima622:35
*** menesis has joined #zope22:44
*** goschtl has joined #zope22:45
*** m8 has quit IRC22:54
*** redir has quit IRC23:03
*** hever has quit IRC23:04
*** m8 has joined #zope23:08
*** evilbungle has joined #zope23:17
*** evilbungle has quit IRC23:17
*** lucmult has left #zope23:17
*** TresEquis has quit IRC23:18
*** gr0gmint has joined #zope23:26
*** Gogo|gui has joined #zope23:30
*** alga has joined #zope23:30
*** rogererens has joined #zope23:30
*** goschtl has quit IRC23:35
*** Gogo|gui has quit IRC23:37
*** menesis has quit IRC23:40
*** jim_SFU has joined #zope23:42
*** m8 has quit IRC23:46
*** gr0gmint has quit IRC23:48
*** jim_SFU has quit IRC23:53
*** dayne has quit IRC23:58

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!