{23} Trac comments (3729 matches)

Results (3101 - 3200 of 3729)

Ticket Posixtime Author Newvalue
#2371 1340038431000000 toby awaiting release and demo theme merge
#2371 1345209675000000 toby closing as part of 2375 branches
#2372 1343220160000000 aron.carroll These are now in https://github.com/okfn/ckan/blob/b31109b628a21072ec0c2dec2a5d5a1c2cc0ed9c/doc/coding-standards.rst#javascript-coding-standards
#2373 1336142514000000 dread Fixed in https://github.com/okfn/ckan/commit/8b7141cd2ff619a0e35aaa31bbc1d17156647498 on master (and also on release-dgu1).
#2374 1340287976000000 ross count is returned at the top level of the dict (at the same level as results) Was added by dread on 2012-05-04, must have been in a recent patch.
#2375 1338292898000000 toby stealing this ticket
#2375 1340710303000000 toby moving milestone
#2379 1340033026000000 ross Implemented in master
#2380 1337868169000000 ross datagm is VERY broken with 1.7 Waiting for datasuite UI changes.
#2380 1340627558000000 ross Dupe of #2545
#2383 1337872916000000 icmurray This helper function won't parse UTC offset if passed a UTC offset as "+01:00"; but incorrectly parses the UTC offset "+0100" as 100 microseconds. Resolution: - I've decided to keep the behaviour of not parsing UTC offsets because we currently don't store timezone information in the database; and I think that either dropping the tzinfo, or adjusting for it would appear to be "magic behaviour" from the user's perspective. Also, as this is a helper function, I think it makes sense to keep the behaviour as is, even if it's slightly non-standard. This fix [1] maintains all the old non-ISO behaviour: - not allowing UTC offset to be specified. - arbitrary delimiters between date and time values. - microsecond precision. But won't parse "+0100" as microsecond precision. I tried using a couple of python modules for parsing date strings; but none maintained the same behaviour. [1] https://github.com/okfn/ckan/commit/d6c9e48f8df5e926abd3fb1a1607f4f5a30075b4
#2384 1337100810000000 dread Done in 0bc06c6495c749d6b481164d59cceeca873acefb
#2386 1337967006000000 seanh Done. ckanext-spatial and ckanext-qa are not activated because they don't work with the current version of CKAN. datastorer is not installed yet.
#2389 1337237502000000 rgrp Just to note these were 2 *different* options. '''Either''' you do the fallback approach '''or''' we turn off DataStore by default and only use if enabled (which happens due to explicit enabling of DataStorer turning it on ...)
#2390 1337100788000000 dread Done in d1c2ed08c90c2d94e1b2ced86e8758208a61ee96
#2391 1338397982000000 seanh I've told them how to do it themselves and offered to help if they get stuck
#2394 1337161917000000 dread Done in [master b486244]
#2398 1337963342000000 seanh I'm closing this because we've got a whole bunch of tickets from the dev meetup about fixing the dependencies and requirements file that should deal with this problem
#2399 1340706702000000 amercader Fixed in 493589aaf
#2400 1338205947000000 toby In testing this is not replicable due to lack of info so closing
#2401 1337273042000000 dread Done in master 112e405d0116bdcf196062520e3474e231337996
#2402 1337302474000000 kindly cset: 12da5e9effeeb1aca0df321c355d8438647ef426
#2403 1337302700000000 kindly cset: 05144f8621ee719c345373934e70719f46e87cf6
#2405 1337583749000000 rgrp Assigning to ross for review and assignment.
#2405 1338212357000000 ross Suggest Sam might be able to provide you the icon
#2405 1338213564000000 icmurray Questions for client team: 1. Did you have anything in mind for improving the "Powered by CKAN" text and icon? I'll sort out the vertical alignment of the text and icon; and see if Sam has a better icon. 2. The branded section itself is already customizable by a snippet (https://github.com/okfn/ckan/blob/master/ckan/templates/snippets/data-viewer-embed-branded-link.html). Is that enough? Or are you asking for something else? 3. Personally I think the URI should use the resource uuid, and not name. And from recent discussions on ckan-dev, I think the consensus is the same. How important is this for you? Perhaps it's something that could be overriden in the datahub extension.
#2405 1340097846000000 icmurray Bump: Questions for client team above...
#2406 1337766385000000 toby I installed CKAN yesterday afternoon from source, using the instructions at http://docs.ckan.org/en/ckan-1.7/install-from-source.html Most things worked pretty smoothly and, luckily, when they didn't I was lucky enough to have David Raznick on hand to help sort them out. As a result I've made the following notes on where the doc could be improved. For some ghastly reason I am running Linux Mint, which is some variant of Ubuntu. It's not 10.04, so no package, hence the source install. Sec 2: Not wanting a mysterious file called 'pyenv' in my home directory, I created a directory ~/ckan and typed 'virtualenv pyenv' from there. Everything worked fine. But if I had needed the 'tip' it would have been a little confusing. It would also be worth reminding the user that the dubious package is one they've supposedly just installed, so if that went OK, they won't need to worry about this. The tip also isn't clear about under what circumstances I'd need to do the 'easy_install pip' - I didn't need to (is that also only when python-virtualenv is missing?) Sec 4: This worked, but the example command for installing a particular version is confusing. It says it is for 1.5.1, but replacing '1.5.1' with '1.7' fails because there is a rogue 'c' in the command. Probably it would be better to use the latest version (1.7) as the example anyway. Or is there a way of asking for the latest released version? I understand this checks out the whole Git repository with all history - I can't imagine why I'd need to do this to build a particular version. But David assures me it's very sensible. Sec 6: This worked, but it wasn't very clear what the purpose of the 'List existing databases' step was. Sec 8: So now I had to try to get Solr working using the instructions at http://docs.ckan.org/en/ckan-1.7/solr-setup.html This didn't work at all, in ways I can't now remember, to do with Java. In the end David suggested using Tomcat rather than Jetty (don't ask me what these do, but they have something to do with getting Solr to work ...) and this turned out to be much more straightforward to install and get working, with two caveats (below). David suggested maybe Tomcat should be the default suggested in this doc. - One problem with Jetty was to do with using the wrong version of Java, which seemed to be fixed using "sudo update-alternatives --config java". It didn't fix Jetty - not sure if it would also have been needed with Tomcat. - Soft linking the schema file didn't work - it seems Tomcat doesn't like soft links. We made a copy. After this everything worked, and I now have my very own CKAN. Hurrah! Mark
#2406 1337766413000000 toby The 'post-installation setup' instructions at http://docs.ckan.org/en/ckan-1.7/post-installation.html claim to be for either package or source installations, but only work as stated if you have done a package installation. E.g. after a source install, "paster --plugin=ckan user add admin --config=/etc/ckan/std/std.ini" doesn't work (the ini file is somewhere completely different), but just typing "paster user add admin" from the ckan directory is fine. The instructions should also mention that the virtual environment should be activated. Mark
#2407 1340633440000000 amercader This will require some thought. Bumping to 1.9.
#2408 1341497601000000 aron.carroll This will be done as part of adding javascript in phase 2
#2408 1343657014000000 aron.carroll Fixed in df502fd9269859c73458f4ad8de75939220e0315
#2409 1337688965000000 ross This is in the Package controller and may be related to upcoming ticket on removal of autoneg.
#2410 1338193633000000 ross Tim Lebo and Keith Alexander are both now looking at how they can implement this themselves, so for now this is wontfix. We can reopen this as part of any future work with VoID
#2411 1338210872000000 toby This is a long term goal
#2414 1337869578000000 ross LXML is currently used for markdown_extract and check_solr_schema_version Changing check_solr_schema_version should be possible, Used in markdown_extract to just strip HTML tags.
#2414 1338471374000000 ross Merged by kindly
#2416 1338464207000000 toby commit 37622d5 does this but probably in the wrong place but works for now. Need to revisit at some point.
#2416 1340710252000000 toby moving to phase 2 as not causing problems but still needs proper fix
#2417 1337864155000000 toby fixed in demo-theme branch
#2418 1337875172000000 toby all facets now shown and selected have <li class="active"> @aron styling of active facets with long names breaks. we could make the names shorter in snippet/facet_list.html via truncate or make the selected taggy thing wider
#2419 1337863024000000 toby @aron the reason for this is that controller/package.py renders the form `package/new_package_form.html` as part of it's work and then makes this available a `c.form` to the template. I'm not sure I like this as an approach but it is how it is currently done. If you split your form into a new template `package/new_package_form.html` then you get all the errors etc. I think for now we should do this and then look at if doing it all in one template pass might be more sensible. reassign to me if you want me to do this else I'll leave it with you
#2420 1337950422000000 toby merged into master
#2422 1337945204000000 seanh Closing because I'm not able to reproduce and it turned out floapps had publisher auth on which was overriding his rights auth
#2423 1338193199000000 ross Duplicate of #2414
#2424 1338193269000000 ross Duplicate of #2413
#2428 1341234933000000 seanh This awaiting merge, pull request: https://github.com/okfn/ckan/pull/46
#2428 1341314475000000 seanh Merged. The only thing I didn't do was add a test case that runs pip freeze and tests that all the right versions of deps are installed.
#2430 1338397815000000 seanh http://trac-hacks.org/wiki/MasterTicketsPlugin
#2431 1338397906000000 seanh https://github.com/davglass/github-trac
#2432 1338205526000000 aron.carroll It would be great to use something like http://trac-hacks.org/wiki/SimpleTicketPlugin to remove most of the options when creating a ticket. Suggestions to remove theme, repository, component and type unless they're actually useful.
#2432 1338325120000000 shevski Where do new tickets go to await triaging? Is priority set by whoever creates the ticket? Does this get revised? Will we have a field for dev time estimate or where will this go?
#2433 1337960203000000 toby the affected actions =============== package_list package_relationship_create revision_show group_list package_update_validate It would be nice to get this fixed and I'm loathed to suggest adding a version 4 hack. If we make this change it will break things for existing users. This behaviour has existed since the action api was added as far as I'm aware So should we a) leave it forever and document b) break things and fix it - when 1.7.1 or 1.8?
#2433 1338397268000000 toby after discussion it has been agreed to keep the existing behaviour and improve documentation as it is likely to break existing applications etc
#2434 1341234522000000 seanh We've moved to AWS hosting with docs generated by Jenkins instead, see: http://trac.okfn.org/ticket/1254 There are still two issues to resolve: 1. Automatically updating the symlink so that docs.ckan.org always redirects to the latest stable version 2. Finding somewhere to put links to all the other versions
#2434 1350296272000000 amercader Seems to be working again, closing.
#2437 1338484676000000 seanh Please add to the file started to doc/coding-standards.rst Also, for the Python coding standards see http://wiki.okfn.org/Coding_Standards as a starting point
#2437 1338487089000000 seanh I've also added the commit message guidelines as they reference a trac plugin I'm going to get installed. I'll leave the rest to you. I remembered that http://wiki.ckan.org/Main_Page has a bunch of pages under Developer Resources that should probably be updated and merged into these coding standards and deleted from the wiki
#2438 1338212234000000 icmurray Proposed solution: Add a new q field which can be specified multiple times. Each q parameter is a string: `{field}:term`. `:` characters are escaped in the `{field}` by `\:`. Mark the `fields` parameters as deprecated.
#2438 1340736112000000 icmurray Ready to be merged [1] Brief synopsis of changes: - altered return type of the action: from Resource domain objects to dictized resources. This is required if this action is to be made accessible through the action api as the domain object is not json-serializable. The 'old' return type is still available by passing in a flag in the context. In fact, this is required as the search layer uses this action, and deals with the domain objects directly. - decided that escaping the ':' (as described in above comments) wasn't necessary as resource fields shouldn't have ':' characters in them anyway. - improved action's docstring - 'fields' parameter has been deprecated ( see #2603 ). Use of it by the search layer has been removed in favour of the new 'query' parameter. But it still works, and correctly maintains backward compatibility. [1] https://github.com/okfn/ckan/pull/37
#2439 1338214158000000 icmurray The `fields` argument doesn't provide any extra functionality that the `query` argument can't provide. 1. Figure out what the reason for this argument is, just to make sure. 2. Deprecate it, and change any internal uses of it to use `query` field instead.
#2440 1339421746000000 aron.carroll Has some styling as of b0153ba but needs content to be applied. Toby, can you pull in the top 10 tags, as well as n of the most "popular" datasets and n of the most recent datasets. Actually just providing the functionality as helpers would be great. Do we actually have "featured" datasets implemented at the moment? Not sure what we're going to use for the box on the top left when on a new base CKAN instance.
#2440 1339423966000000 toby @aron, tags what functionality do we want for this? I can just reuse the facet_list snippet but then we have all sorts of interactivity. Do we want this more a static list of tags (linked to somewhere but no add remove stuff)? seems more sensible I think featured datasets I've no idea I'll try to discover
#2440 1339430956000000 toby you may want to look at your css as the dataset is generated by the package_list snippet I've made all packages popular for now
#2440 1339497700000000 aron.carroll Fixed this up, last thing I think we should do is ensure that the datasets passed to the homepage have descriptions.
#2440 1339498655000000 toby dataset descriptions is a data issue but I'll keep this open for now. We can probably do an exclude if needed.
#2440 1339499074000000 aron.carroll Sure, but this is more about displaying nice complete examples on the homepage.
#2440 1339581622000000 toby I am changing this to assigned not accepted as this makes trac easier to use
#2440 1340373461000000 toby i'm closing we can make tickets for specific issues
#2441 1338287243000000 aron.carroll Order by needs hooking up to form. Currently it's just a @<select>@ element containing a single entry, it needs the other possible options added. Let me know if you want more details.
#2441 1338291362000000 toby @aron, this is now in a snippet (jinja2 allows the selected="selected" more easily) needs some work still eg translation stuff and also maybe completely rewrite the snippet at some point. Anyhow the functionality is there. I added the button for no javascript working. I'm assigning to you for beautification etc on live demo http://s031.okserver.org:2375/dataset?q=&sort=title_string+asc
#2441 1338306828000000 aron.carroll There's already a submit button in the form (it's the question mark) it's up for discussion on Basecamp if this is clear, apparently it's not. I'm going to throw this back to you to implement the popular badges, I'll tidy that form up later. Perhaps we can get a helper such as h.isPackagePopular(pkg) that can be customised on a per instance basis.
#2441 1338306866000000 aron.carroll Not question mark sorry, I meant magnifying glass (inside the search box)
#2441 1338384362000000 toby popular badges on package list - may make them optional/configurable at some point but good enough for now. see http://trac.ckan.org/ticket/2442 for details With the `order button` this is for none javascript users i'm not sure the magnifying glass looks like the thing to press to reorder the results. I'd hide it when javascript enabled. Do what you will @aron I'm assigning this to you close or reassign to me if you need extras adding
#2441 1339670118000000 aron.carroll Closing this as the page is up. Issues should be reported as new tickets.
#2442 1338302281000000 aron.carroll Functionality to be added: * Group module needs to display group/publisher title, logo and description if dataset has an associated group. * Popular icon only needs to be displayed if dataset is popular. (say if it has 10+ views for the moment) * Popular icon needs localised text passed in to snippet (See TODO comment above helper https://github.com/okfn/ckan/blob/feature-2375-demo-theme/ckan/templates/package/read.html#L46) * Related module needs to show most recent item or fallback content.
#2442 1338305764000000 toby make readable * Group module needs to display group/publisher title, logo and description if dataset has an associated group. * Popular icon only needs to be displayed if dataset is popular. (say if it has 10+ views for the moment) * Popular icon needs localised text passed in to snippet (See TODO comment above helper https://github.com/okfn/ckan/blob/feature-2375-demo-theme/ckan/templates/package/read.html#L46) * Related module needs to show most recent item or fallback content.
#2442 1338379591000000 toby popular localised need to get live tracking data and fixup (in progress) groups now shown including multiple with link to group page related items I'd like some data for this I'll see if I can add some
#2442 1338381909000000 toby now has actual tracking data to get data install ckanext-googleanalytics from source in analytics package dir run `paster loadanalytics internal 2012-01-01 -c ../ckan/development.ini` you may need to change .ini file name then in ckan dir `paster search-index rebuild -r -c dev.ini` It needs a refactor to make it easier to use
#2442 1338383819000000 toby new helper function for popular h.popular(type_, number, min=1, title=None) just related items left now
#2442 1338477973000000 toby @aron, related items on dataset page has been added it's really basic and just displays the fields that seem useful. we may need to add the add related item feature first as they are hard to add. please make it beautiful
#2442 1339670612000000 aron.carroll Closing, issues should be opened as new tickets
#2444 1339669652000000 aron.carroll Done. Any issues should be created as new tickets.
#2445 1338375577000000 ross Yup. Will look at this Thurs when I'm back in.
#2445 1338378161000000 toby @Ross you will need to do this based off the feature-2375-demo-theme branch - you can use that branch or one split from it. new templates want to be in templates/... the old templates live in templates_legacy if you need to copy them across that's fine talk to aron or I if you need to know anything
#2445 1338472050000000 ross Okay, the 303 on validation failure seems a little excessive though given that we're posting to that URL. Is a 200 okay?
#2445 1338472161000000 ross Am working in feature-2445-add-related-new
#2445 1338472624000000 toby @ross the validation error seems correct to use 200 for the created the 303 seems ok but i'd just use pylons redirect myself Also did aron say we want this as a javascript free add item form
#2445 1338472966000000 ross He didn't say but I assumed yes given current impl uses dodgy JS->API.
#2445 1338473161000000 aron.carroll I generally take the approach that a form submitted via POST should always redirect to a page. This prevents accidental user resubmissions when hitting refresh as well as the annoying popup saying "are you sure you want to re-submit". http://en.wikipedia.org/wiki/Post/Redirect/Get Looking at the rest of CKAN it doesn't do this so I'll leave it up to you.
#2445 1338473309000000 ross redirects are fine, of course, but redirect on validation failure seems painful ? Also templates in templates_legacy?
#2445 1338474490000000 toby @aron I'm fine with the redirect thinking it's just the specific 303 rather than just whatever pylons does. @ross, templates_legacy are there for ckan users who have their own templates as we are completely rewriting all the existing templates there is a lovely .ini setting to switch between the new and legacy modes
#2445 1338542340000000 ross I *think* this is done in feature-2445-add-related-new although I may have missed something, if you could take a look and let me know if it is what you wanted.
#2445 1338911121000000 aron.carroll Thanks Ross this is exactly what I wanted. Just wondering about the use of the c.form variable to insert the form into the page. I know we do this with the dataset and groups to support legacy templates, do we need to do this going forward? I'd rather each action just rendered one template and leaves it up to that to render the rest. Toby, there's a couple of TODO statements in the code where the breadcrumbs need fixing. I think the templates just need access to the package dict. https://github.com/okfn/ckan/blob/feature-2375-demo-theme/ckan/templates/related/pages/form_page.html#L6 https://github.com/okfn/ckan/blob/feature-2375-demo-theme/ckan/templates/related/edit.html#L5 Also if you generate an error message on one of the forms (leave the title blank) the error string comes out wrapped in quotes. eg. u'title is missing' Also might be of interest, I've converted the templates to use Jinja, and this is a good example of the conversion. https://github.com/okfn/ckan/commit/d779addf6e48be67695b35c031393e2d90d8c22f
#2445 1338911190000000 aron.carroll Oh forgot to mention I've merged the code into feature-2375-demo-theme. We should delete the feature-2445-add-related-new branch.
#2445 1338975600000000 toby @aron, this is as good a place as anywhere to answer your question about the c.form thing. a) some of this we are sort of stuck with. datasets for example can be of different types and have different forms for different datasets. so they need to be rendered separately. I have no idea why or how much this is used but it exists as an option to complicate everything, not much we can do about it. b) I think trying to do pages as a single rendering is not going to be easy because of the way ckan has been created maybe we can change this over time. One major issue for me is that ckan does lots of crazy stuff for some reasons but these reasons are not apparent from the code. Maybe there is some documentation somewhere but I'm not convinced. anyhow we should maybe close this ticket now. You've added some stuff in the comment for me to do maybe that should be added as a new ticket?
#2445 1338982754000000 toby closing as todos moved to other tickets and also completed
#2446 1352206530000000 johnmartin I can't get access to said project. Can someone please give me access so I can triage this?
#2446 1352658819000000 shevski gave john basecamp project access
#2447 1343377924000000 toby @aron, can we close this as being too general - and have specific tickets?
Note: See TracReports for help on using and creating reports.