IRC logs

20140426

Logs from channel #fedext on freenode - our official support channel.

IRC log range: 20140426*

20140426

  • 01:11:29 <Outdoorsman> Is it possible to use the new Responsive Image features in TYPO3 CMS 6.2 so they work with Fluid Bootstrap?
  • 01:11:41 <Outdoorsman> http://typo3.org/news/article/responsive-image-rendering-in-typo3-cms-62/
  • 02:03:07 <Outdoorsman> I don't really know how it would work conditionally because there's no option on every image to checkbox "responsive" so it would work with twitter bootstrap.
  • 02:04:47 <Outdoorsman> Maybe randomresult has already thought about this?
  • 02:11:29 <Outdoorsman> On a different but somewhat related note, there's the issue of not being able to use FAL with images that is talked about here https://github.com/FluidTYPO3/flux/issues/468 and here https://github.com/FluidTYPO3/flux/issues/175
  • 11:15:43 <jlumpe> morning! I'm encoutering a weird issue: after creating some content elements with fluidcontent, when I add nested content elements to them, they aren't added in the correct order and sometimes are removed from the content element and moved to the root of the page. Has anybody encountered this issue before?
  • 12:49:47 <soee> guys when can we expect stable version of fluidboostrapteme ?
  • 12:49:59 <soee> or the first TER release
  • 12:53:54 <NamelessCoder> if you want to help that along, soee, assist Danilo with testing and reporting on https://github.com/FluidTYPO3/flux/pull/524 - until that is done, no Flux TER and until that happens, no fluidbootstraptheme
  • 12:55:22 <jlumpe> NamelessCoder: are you aware of an issue which leads to fluidcontent elements not being properly nested?
  • 12:55:43 <jlumpe> for example when copying them around
  • 12:55:47 <soee> NamelessCoder: will take a closer llook, thanks
  • 12:55:53 <NamelessCoder> jlumpe see same issue I just linked to
  • 12:56:16 <jlumpe> oh
  • 12:56:17 <jlumpe> :)
  • 12:56:57 <jlumpe> also, since i'm new to the party, what is the rationale behind removing the rte from fluidcontent elements?
  • 13:05:11 <NamelessCoder> not sure what you mean?
  • 13:09:18 <jlumpe> when i create a new FCE, then the default bodytext field has no rte applied to it anymore
  • 13:09:55 <NamelessCoder> enableRichText="TRUE"
  • 13:10:04 <jlumpe> aha!
  • 13:10:08 <jlumpe> things can be so easy
  • 13:10:09 <NamelessCoder> just like in TCA, enabling rte requires additional settings. That attribute is a shortcut
  • 13:10:19 <jlumpe> ah
  • 13:10:54 <jlumpe> thanks so much. getting back into typo3 atm. any way i can help with things maybe? (might also help me to understance fluidtypo3 better)
  • 13:13:07 <NamelessCoder> there might be - are you comfortable with php in general?
  • 13:13:46 <NamelessCoder> we have https://github.com/FluidTYPO3/vhs/pull/545 which is fairly simple work but time consuming
  • 13:13:51 <jlumpe> yeah i think so.
  • 13:14:08 <jlumpe> of course there is always room for improvement, but yeah i can read and write php
  • 13:14:38 <NamelessCoder> making pull requests into vhs's branch workbench/namespaces will add your commits to this one - if you do start refactoring some classes, please write in the issue comments which ones you're handling so we don't do double
  • 13:16:38 <jlumpe> ok
  • 13:17:07 <NamelessCoder> basically the process is: 1) add the namespace in php class, 2) rename class, 3) if used by other classes, correct references, 4) in ClassAliasMap.php, find the entry and reverse the map so the new name points to the old
  • 13:19:31 <jlumpe> yeah
  • 13:19:41 <jlumpe> ok, i'll clone the repo in a few mins
  • 13:20:34 <NamelessCoder> ah yeah, forgot to say that: make a fresh fork of vhs if you already have one. Just to make sure it's all clean
  • 13:21:04 <jlumpe> don't have one yet :)
  • 13:43:17 <jlumpe> NamelessCoder: https://github.com/johanneslumpe/vhs/commit/8a6de522338472b9118596d4afe1b7bbc04c68de
  • 13:43:21 <jlumpe> fine line this?
  • 13:43:48 <NamelessCoder> very close
  • 13:43:49 <jlumpe> Tx_Vhs_ViewHelpers_AbstractViewHelperTest will be renamed later in another ocmmit, when i process that class
  • 13:43:56 <jlumpe> ok, what's missing?
  • 13:44:00 <NamelessCoder> the unit test file also needs the namespace
  • 13:44:07 <jlumpe> doh
  • 13:44:26 <NamelessCoder> and you should "use TYPO3\CMS\Fluid\Core\ViewHelper\AbstractConditionViewHelper" then "extends AbstractConditionViewHelper"
  • 13:44:45 <NamelessCoder> the reason for *that* is that if done this way, ports to Flow become very easy
  • 13:44:58 <jlumpe> ah ok
  • 13:45:01 <NamelessCoder> do you have phpunit installed?
  • 13:45:13 <jlumpe> no, but would probably be good to install that
  • 13:45:24 <jlumpe> does it have a cli?
  • 13:45:38 <NamelessCoder> git clone https://git.typo3.org/TYPO3CMS/Extensions/phpunit.git typo3conf/ext/phpunit
  • 13:45:50 <NamelessCoder> ./typo3/cli_dispatch phpunit path/to/tests
  • 13:46:02 <NamelessCoder> needs BE user _cli_phpunit and _cli_lowlevel
  • 13:48:06 <NamelessCoder> you know how to amend commits?
  • 13:49:43 <jlumpe> no :/
  • 13:49:52 <jlumpe> but you can probably teach me?
  • 13:49:55 <NamelessCoder> you use git command line or some client?
  • 13:50:11 <jlumpe> right now i use git for windows
  • 13:50:30 <NamelessCoder> okay - a client would perhaps be easier but you can still get there :)
  • 13:50:48 <jlumpe> i can fire up sourcetree?
  • 13:51:15 <NamelessCoder> in unix commands it would be, after you have a commit and then made code changes you wan to amend: "git commit --amend -a" then "git push --force" to force update your remote branch
  • 13:51:24 <NamelessCoder> sourcetree is probably the best option for windows yes
  • 13:51:47 <jlumpe> ok, let me start that
  • 13:52:29 <NamelessCoder> it's about the same procedure - in sourcetree you would add the changed file, commit and toggle on amend - then force-push to remote exactly like in cli
  • 13:53:00 <jlumpe> ok i will try, if i fail i can still delete my branch and start over
  • 13:53:10 <jlumpe> oh and is there any way to run the unit tests without a full typo3 install?
  • 13:53:20 <NamelessCoder> for *this* pull request it is not so important that you amend, since it will all be squashed later - but this is the exception ;)
  • 13:53:30 <jlumpe> haha ok ;)
  • 13:53:50 <NamelessCoder> unfortunately no - there is not. Fluid ViewHelpers require the base class from Fluid and Extbase's object manager and class registry :/
  • 13:54:23 <NamelessCoder> but you can create your pull request now - that way, your commits will be tested by travis when you push to your own remote branch
  • 13:54:51 <jlumpe> yeah that is much more comfortable
  • 13:54:52 <NamelessCoder> or you can simply sign into travis.org yourself and toggle on the vhs repository to have your very own builds running ;)
  • 13:57:09 <jlumpe> I'll just have one big pull req in the end, is that fine?
  • 13:57:32 <NamelessCoder> we will be more people working on this so we would like it in smaller chunks
  • 13:57:50 <jlumpe> ok
  • 13:57:50 <NamelessCoder> either that, or write a comment to indicate which classes you will be working on
  • 13:58:35 <jlumpe> i can also just always send small ones, not an issue :)
  • 13:59:58 <jlumpe> yay, seems to be good
  • 14:00:29 <NamelessCoder> yep - only one tiny cosmetic thing: extra line break after the namespace in test case file
  • 14:00:37 <NamelessCoder> remove that and it is *perfect*
  • 14:00:59 <NamelessCoder> (we are the masters of CGL nitpicking here ^^)
  • 14:01:38 <jlumpe> haha
  • 14:02:02 <jlumpe> i don't mind that. helps to keep code uniform
  • 14:04:00 <NamelessCoder> consistency is extremely important with 60+ contributors ;)
  • 14:06:53 <jlumpe> ok, updated the PR
  • 14:07:10 <NamelessCoder> missing prefix on commit message ;)
  • 14:07:16 <jlumpe> argh
  • 14:07:26 <jlumpe> how do i change that?
  • 14:07:27 <NamelessCoder> you can squash those
  • 14:07:47 <NamelessCoder> do "git reset HEAD^1" twice then "git commit --amend -a" and force push
  • 14:10:15 <jlumpe> yay, already a git master
  • 14:10:29 <NamelessCoder> quick learner
  • 14:10:31 <jlumpe> lol
  • 14:10:52 <jlumpe> question: as long as the PR isn't merged I cannot push more, right?
  • 14:10:58 <jlumpe> since that would lead to a bloated pr
  • 14:11:16 <NamelessCoder> exactly - and you can also keep pushing after it is closed, then recreate the PR
  • 14:11:24 <jlumpe> Ah ok
  • 14:11:36 <jlumpe> Then I can just keep doing stuff
  • 14:12:09 <NamelessCoder> oh misunderstood: as long as it is NOT merged you can keep pushing. When merged, commits get picked. If you then recreate the PR after adding more commits, the comparison base will be renewed and only your new commits get added
  • 14:12:59 <NamelessCoder> you should also "git remote add upstream https://github.com/FluidTYPO3/vhs.git"
  • 14:13:19 <NamelessCoder> that way you can get the changes we make, using "git pull upstream workbench/namespaces --rebase"
  • 14:13:35 <jlumpe> ok
  • 14:15:58 <NamelessCoder> just to make sure (you're probably already doing this) do a search on the old class name after renaming. The only result should be in ClassAliasMap.php
  • 14:16:58 <NamelessCoder> now you can try that rebase thing
  • 14:17:37 <jlumpe> yeh i'm searching for those
  • 14:17:44 <jlumpe> and actually there is a reference in the docs
  • 14:17:51 <jlumpe> * Note: building of all Assets takes place in the class
  • 14:17:51 <jlumpe> * Tx_Vhs_ViewHelpers_AssetViewHelper with two reasons:
  • 14:18:04 <jlumpe> that should probably become
  • 14:18:08 <jlumpe> * Note: building of all Assets takes place in the class
  • 14:18:08 <jlumpe> * FluidTYPO3\Vhs\ViewHelpersAssetViewHelper with two reasons:
  • 14:18:14 <jlumpe> with another \ of course
  • 14:18:43 <NamelessCoder> hmmm... that doc is bad, that's not how it is done
  • 14:18:59 <jlumpe> ok, then i won't change that for now
  • 14:19:21 <NamelessCoder> should be \FluidTYPO3\Vhs\Service\AssetService
  • 14:19:43 <NamelessCoder> I'm fixing this now
  • 14:20:54 <bjo3rn> hi guys. what's the policy re: fqn and use again? I see mostly fqns in recent commits.
  • 14:21:46 <NamelessCoder> bjo3rn it's a two-step operation: until *our* classes also use namespaces, we keep the FQN. When we use namespaces, we start importing FQN by use instead
  • 14:22:07 <NamelessCoder> the reason being bad results if "use" is mixed with a non-namespace class file
  • 14:22:18 <bjo3rn> ok, I see.
  • 14:26:35 <jlumpe> NamelessCoder: is there a reason (i assume there is) why you wrote \FluidTYPO3\Vhs\Utility\ViewHelperUtility in most classes when updating the viewhelperutility to namespaces insteaf of doing use \FluidTYPO3\Vhs\Utility\ViewHelperUtility and then just ViewHelperUtility?
  • 14:27:01 <NamelessCoder> jlumpe I just wrote the explanation to bjo3rn ;)
  • 14:27:24 <bjo3rn> hehe, welcome to the club jlumpe ;)
  • 14:27:25 <NamelessCoder> after you make the switch to namespaces in each class, importing the FQN by "use" is next
  • 14:27:28 <jlumpe> haha
  • 14:27:35 <jlumpe> ok
  • 14:27:43 <jlumpe> ah
  • 14:27:47 <jlumpe> FQN = fully qualified namespace
  • 14:27:47 <jlumpe> ha
  • 14:27:50 <NamelessCoder> but you do not need to do this in the same step
  • 14:27:58 <jlumpe> ok
  • 14:28:26 <jlumpe> sorry to bother, another question
  • 14:28:37 <jlumpe> AbstractAssetViewHelper implements Tx_Vhs_ViewHelpers_Asset_AssetInterface
  • 14:28:57 <jlumpe> when updating the former, should the latter already be updated? or should the latter be updated separately?
  • 14:28:58 <NamelessCoder> you've two choices in that case
  • 14:29:10 <jlumpe> i could do two commits, or one bigger one ,right?
  • 14:29:13 <NamelessCoder> 1) start softly by adding \ in front of the "extends $class"
  • 14:29:21 <NamelessCoder> 2) migrate the interface first
  • 14:29:31 <jlumpe> ah ok
  • 14:29:45 <jlumpe> 2) seems cleaner
  • 14:29:48 <NamelessCoder> each of course has implications, mainly in the IDE. Of course class completion doesn't work until the whole mesh is migrated
  • 14:30:02 <NamelessCoder> yeah, 2 is cleaner but requires more conscious effort :)
  • 14:30:24 <bjo3rn> are you on the asset related classes jlumpe?
  • 14:30:32 <bjo3rn> bc I am atm :)
  • 14:30:39 <NamelessCoder> in VHS this is almost a non issue since this is about the only interface and it's used by three classes total
  • 14:30:51 <NamelessCoder> hehe yeah, need to coordinate this ;)
  • 14:32:13 <jlumpe> bjo3rn
  • 14:32:13 <jlumpe> ha
  • 14:32:14 <jlumpe> ok
  • 14:32:24 <jlumpe> then i'll skip the asset classes
  • 14:32:34 <bjo3rn> depends on how far you got already
  • 14:32:47 <jlumpe> just started
  • 14:32:49 <jlumpe> all good
  • 14:32:57 <bjo3rn> ok, then it's settled ;)
  • 14:33:11 <jlumpe> was just assessing the best starting point ;)
  • 14:33:11 <bjo3rn> bjo3rn gives internet hi5 to jlumpe
  • 14:33:25 <jlumpe> ha, you don't have ./me? :D
  • 14:33:33 * jlumpe gives it back
  • 14:33:34 <jlumpe> ;)
  • 14:33:35 <NamelessCoder> I would divide this into sub-namespaces: Iterator, Math, Content and "/" etc.
  • 14:33:52 <bjo3rn> ./me
  • 14:33:55 <bjo3rn> ??
  • 14:33:57 <jlumpe> without the .
  • 14:33:58 * bjo3rn
  • 14:34:00 <bjo3rn> ah!
  • 14:34:11 <bjo3rn> I'm an old guy ;)
  • 14:34:18 <jlumpe> haha, irc is old :P
  • 14:55:16 <jlumpe> NamelessCoder: started doing the math viewhelpers. Question: I started updating the abstract classes now (and all their references). now i also need to update the unit tests. I'll have to update all the extends as well. is that fine in a single commit?
  • 14:56:20 <NamelessCoder> it's a lot of code but yes, a single commit is fine
  • 15:03:55 <bjo3rn> ok guys, unit tests left to do. need to do something else now but will finish the rest later (just for the record and to avoid redundancy). cya later.
  • 15:04:57 <jlumpe> o>
  • 15:06:17 <jlumpe> ok lets see how this one works out
  • 16:44:53 <jlumpe> NamelessCoder ?
  • 16:47:03 <soee> someone remember what CGL states about concatenatind strings in multuple lines ? The . sign should be at the end or beggining ?