I am working on a nodejs project with typescript 2.2 that is using node 6.3.1 and I want to migrate from using typings to using #types. By doing so I ran into a set of questions related to whether there is a relationship between the version of the #types file and the corresponding npm package.
If I use jasmine as an example, the existing versions of the types definitions are
npm show #types/jasmine#* version
#types/jasmine#1.3.0 '1.3.0'
#types/jasmine#1.3.1 '1.3.1'
#types/jasmine#1.3.2 '1.3.2'
#types/jasmine#2.2.29 '2.2.29'
#types/jasmine#2.2.30 '2.2.30'
#types/jasmine#2.2.31 '2.2.31'
#types/jasmine#2.2.32 '2.2.32'
#types/jasmine#2.2.33 '2.2.33'
#types/jasmine#2.2.34 '2.2.34'
#types/jasmine#2.5.35 '2.5.35'
#types/jasmine#2.5.36 '2.5.36'
#types/jasmine#2.5.37 '2.5.37'
#types/jasmine#2.5.38 '2.5.38'
#types/jasmine#2.5.39 '2.5.39'
#types/jasmine#2.5.40 '2.5.40'
#types/jasmine#2.5.41 '2.5.41'
#types/jasmine#2.5.42 '2.5.42'
#types/jasmine#2.5.43 '2.5.43'
#types/jasmine#2.5.44 '2.5.44'
#types/jasmine#2.5.45 '2.5.45'
#types/jasmine#2.5.46 '2.5.46'
But if I examine the versions of the jasmine packages I have;
npm show jasmine#* version
jasmine#2.0.1 '2.0.1'
jasmine#2.1.0 '2.1.0'
jasmine#2.1.1 '2.1.1'
jasmine#2.2.0 '2.2.0'
jasmine#2.2.1 '2.2.1'
jasmine#2.3.0 '2.3.0'
jasmine#2.3.1 '2.3.1'
jasmine#2.3.2 '2.3.2'
jasmine#2.4.0 '2.4.0'
jasmine#2.4.1 '2.4.1'
jasmine#2.5.0 '2.5.0'
jasmine#2.5.1 '2.5.1'
jasmine#2.5.2 '2.5.2'
jasmine#2.5.3 '2.5.3'
Let’s say I am using version 2.4.0 of jasmine, which version of #types/jasmine should I pick? Because even if I use the latest of both, 2.5.46 does not match with 2.5.3.
Another example would be node itself, there are basically 6.0 or 7.0 versions in #types, and typings has only the ones shown below, being 6.0 reported as obsolete. So, what version of node are those typings actually tied to?
typings view dt~node --versions
TAG VERSION DESCRIPTION COMPILER LOCATION
UPDATED
7.0.0+20170322231424 7.0.0 github:DefinitelyTyped/DefinitelyTyped/node/index.d.ts#a4a912a0cd1849fa7df0e5d909c8625fba04e49d 2017-03-22T23:14:24.000Z
6.0.0+20161121110008 6.0.0 github:DefinitelyTyped/DefinitelyTyped/node/node.d.ts#fb7fbd28b477f5e239467e69397ed020d92817e7 2016-11-21T11:00:08.000Z
Thanks
The major and minor versions of the DefinitelyTyped packages are supposed to correspond to the major and minor versions of the package they are types for. The patch version increments whenever the .d.ts file changes for other reasons. Because the minor version shouldn't represent breaking changes, in theory, you can use the highest 2.x.y definition file available for a 2.a.b.c library.
But now the caveats begin.
The header in the definition file may not have changed at the right time
Library authors do not necessarily follow semver*
The definition file may not be 100% correct at any given point, in either direction (i.e. listing a 2.6 feature under a 2.5 version, or failing to list a 2.4 in function in the 2.5 file)
* In fact, no one does
A detailed explanation can be found in official docs FAQ: How do Definitely Typed package versions relate to versions of the corresponding library?
Related
Im attempting to update an old version of the selenium-tcl package to work with the new W3C WebDriver (or Selenium 4.0).
Original packages uses a few mixins for the webdriver class.
So I modeled what I saw and created a mixin file named mixin_action_chains.tcl [1] which has a mixin class called Mixin_Action_Chains.
Whenever I attempt to use it I get the error:
% package require selenium
::selenium::Mixin_Action_Chains does not refer to an object
Im not sure why I've modeled it pretty much exactly as I have seen in the other files such as mixin_for_scrolling.tcl [2]* file. What am I missing.
Here is the entire GitHub Repo
Im not sure what else must be done for TclOO. Any thoughts.
Thanks
Im not sure what else must be done for TclOO. Any thoughts.
Update
pkgIndex.tcl: The placement of the mixin-defining script mixin_action_chains.tcl is wrong, it comes after the mixin has already been required in the previously sourced script webdriver.tcl, like entering directly:
% oo::class create C {
mixin ::not::defined
}
::not::defined does not refer to an object
You need to change the order of source command calls in the package ifneeded script.
For the records
Still, in the original version, there were unbalanced curly braces and brackets in your script, which broke sourcing of the file for me:
https://github.com/SavSanta/w3cselenium-tcl/pull/1
I'd like to ignore old tags during the translation process. I'm trying to use
excludeBranches = tags/Server_*_2007-*
to match tag names like `Server_423_2007-02-28_11-40', but subgit complains about using multiple * symbols.
How can I ignore old tags (based on year in the tag name) during translation? We have a lot of these tags, one for each build, and it slows down import a lot (I think). I want to keep most recent ones (from 2015, 2016) only.
Since version 3.2.0 of SubGit you can use multiple asterisks in 'excludeBranches' option.
We have migrated code from Yocto1.5 to 1.7 and it seems with using the same configuration file(local.conf), I found that a lot of libraries previously installed on the 1.5 image(fsl-image-gui) was not present in the 1.7 image(core-image-sato)
I have manually added most of the packages except for two
libbeecrypt_cxx.so
liblcms.so(cmsutil)
I have already put in the code below and libbreecrypt was insalled but not libbeecrypt_cxx
IMAGE_INSTALL_append += beecrypt
I have not found a recipe/package for cmsutil. Anyone information regarding these would be appreciated.
Best Regards
Yuri
Well, regarding the beecrypt part of your question. Shortly after the 1.5 release, building the c++ bindings were made optional, and defaults to off.
See patch making c++ optional
Thus, what you need is to add a beecrypt_4.2.1.bbappend file in your own layer which includes
PACKAGECONFIG += "cplusplus"
That the best, longterm solution. You could also set
PACKAGECONFIG_pn-beecrypt_append = "cplusplus"
in your conf/local.conf.
Busy migrating a report generation from Rails(Web) to a jruby command-line script.
Using:
jruby-1.7.11
ActiveRecord & ActiveSupport: 3.0.6
jdbc-sqlite3: sqlitejdbc-3.7.9.jar (with custom sqlite3.8 shared library)
jruby-poi: 0.9
Problem:
ResultTable.find_by_sql returns an array containing single length array's in stead of string's. ['country_NLD'] in stead of 'country_NLD'
Under Rails this seams to be working OK.
Questions:
How do I convert a single element array to a NON array even when its deep down in an other array?
I know that Postgress active record-jdbc-adapter has a setting called '...array.raw'. Is there some setting for SQLite3 available?
Thanks!
you did miss an important piece to mention: ActiveRecord-JDBC-Adapter's (gem) version ...
if you happen to be on 1.2.x I do recommend an upgrade to 1.3.x which shall have this incompatibility fixed, otherwise you should report the issue.
I'm using Apache Jackrabbit to store versioned data. I'm following the template listed on the Jackrabbit wiki for versioning basics, but there's one thing that isn't working as expected.
After I run the code from the wiki, I try saving another version of the node, then getting its version string:
child = parentNode.getNode("childNode");
child.checkout();
child.setProperty("anyProperty", "Blah3");
session.save();
Version thisVersion = child.checkin();
System.out.println(thisVersion.getName());
The output of this code is 1.0.0 , when I want it to be 1.1 . My goal is to be able to remove the last -- and only the last -- version of a node, and have the next version have the same name as the removed version. I only care about doing this when the node is being added; if I can check in the node, and read the name (to get the version number) without an exception being thrown, I'm fine with not being able to remove the version. Looking around the internet, I can't find a way to accomplish this.
You can't control the version naming, this is handled by the JCR implementation (in this case Jackrabbit). If you want to have your custom names for versions, use version labels.