refs https://sharp.pixelplumbing.com/api-utility#cache
- Sharp has a 50MB cache by default, used within libvips, to increase
the performance of transforming images
- this isn't relevant to us because we should never be optimizing the
same image as we check if the optimized image already exists
- I presume there is also some extra overhead of using the cache because
the memory doesn't seem to grow by 50MB
- the memory usage comparison in Ghost is pretty drastic - uploading 10
images in serial w/ jemalloc:
- with cache (default) = peak of 480MB, settles down to 330MB
- disabling cache = peak of 270MB, settles down to 161MB
- this commit disables the cache
- also adds stubbing for the function in tests
- As a library, minifier should be subject to 100% coverage
- minor change to the code combining globbing and file reading into a single function for error handling any case where the files can't be found or read
- we could do more fine grained errors here, but that seems unnecessary given that the usecase is internal, not for user files at the moment
- added error case tests to capture all the uncovered lines
- Uglify was really slow and clunky, so replaced with terser which is fast and simple
- Ensured test cleanup works properly so we can be sure about the state of files during tests
- Changed the output from being the absolute path to just the "dest" value, as that is much more useful as absolute paths include local machine path names
- Fixed async/await issue that got the whole thing working 100%
- wired up a basic minification package
- accepts config for css or js files and can concat and minify them into a single file for each type
- this will be used for generating merged css and js files for various cards, controlled by theme config
no issue
- The aim is to achieve 100% unit test coverage for servies and small modules. This change covers few more bases brining limit-service's module coverage from 80% to 94%.
refs linear.app/tryghost/issue/CORE-74/improve-the-test-situation
- this commit adds the codecov GitHub Action into CI so we can upload
coverage reports
- the coverage files need to be in XML for them to work with
codecov, so this commit also adds cobertura (XML) as a reporter
no issue
Unpinning dependencies to ensure deps stay up to date when used in Ghost. Removed Ignition because deprecated in favour of using @tryghost/debug
refs https://linear.app/tryghost/issue/CORE-49/fix-errors-in-utils-repo-limit-service
- As I've touched these files did a little refactor and changed where the error messages are stored to keep it up with our lates coding standard - having "messages" hash defined in the module storing all messages that have pottential for i18y in the future.
refs https://github.com/TryGhost/Team/issues/870
- currently, a handful of packages in this repo generate type
declarations when building
- we're aiming to go more in the direction of better inline jsdoc, and
we want to clean up all generated files and use of external types
- this commit removes the `typescript` dependency, removes the pretest
commands that run the builds, and cleans up anywhere where we
configure types
refs https://github.com/TryGhost/Team/issues/870
- using `c8` allows us to see test coverage for all packages in the repo
- this commit adds `c8` as a dev dependency and prepends the `mocha`
command with `c8` so it runs on all tests
refs https://github.com/TryGhost/Team/issues/727
- The latest code standards require using @tryghost/logging module instead of injecting it as a parameter in the construcotr
- Left it as an optional constructor parameter to make testing easier
refs https://github.com/TryGhost/Team/issues/727
- The latest code standards require using tpl module in combination with local messages object instead of i18n injected in the constructor
refs https://github.com/TryGhost/Team/issues/754
- When there are no message objects coming from the external update check service in the response there is no need to continue to process the data and fire off unneeded queries
refs https://github.com/TryGhost/Team/issues/754
- The api call can also generate an error which would bubble up to a client. This is not a good design if the client has to know to handle internal errors of the module
- By having try block wrap around whole block it makes the error handling behave the same way throughout the check process
- added getSubdir, getSiteUrl and getAdminUrl methods from url-utils
- refactored them so they are designed to be bound by nconf
- exposed a bindAll method
- moved over the tests and refactor from a class instance to nconf bindings
no issue
Replaces the usage of ghost-ignition with @tryghost/ignition-errors,
and switched to using the more modern ES6 class syntax over the
`.call(this, ...)` syntax used previously in both Ignition and here.
- fixed type warnings
- introduced explicit use of bluebird to help with type warnings and also make it clear when we're actually using Bluebird features
- removed the class pattern in the simplest possible way
- have package-json expose only the public methods
- move parse to an independent, not public file so that we can test it independently still
- We are going to get rid of the internal i18n tool because it doesn't solve a real use case
- Instead, we have a new tpl utility that does basic string interpolation
- This makes this module less complex as it no longer requires dependency injection to work
- ts complains the value may be undefined but in reality it never is
- plus we don't care about this error in this context as the test is checking that the value is set to its default
- This mini helper wraps lodash template and supports `{...}` as the delimiters
- It's designed to use and support the exact same patterns we already have in our en.json strings
E.g. The {flagName} flag must be enabled in labs if you wish to use the \\{\\{{helperName}\\}\\} helper.
- This allows us to get rid of our old, broken i18n helper and still keep some of the smart messaging we have setup
- It also keeps the refactor surface area minimal
refs TryGhost/Team#728
- These are dependancies that have to come with the update check service
- Used exact same versions of Bluebird/moment/lodash as in current
TryGhost/Ghost main
refs https://github.com/TryGhost/Team/issues/728
- This is continuation of the previous commit. TLDR: Passing only the necessary parameter data makes it easier to reason about what dependencies the UpdateCheckService has to deal with
- Instead of passing in a whole GhostMailer instance passing only an email sending function, which again - makes things way more manageable to reason about
- The end of refactor, next will be a move of the UpdateCheckService into a separate module in tryghost/core
refs https://github.com/TryGhost/Team/issues/728
- This is continuation of the previous commit. TLDR: Passing only the necessary parameter data makes it easier to reason about what dependencies the UpdateCheckService has to deal with
- Burned ghostVersion module passing in vafor of just one additional config parameter. Now the module along with unit tests can be easily extracted out of the codebase!
refs https://github.com/TryGhost/Team/issues/728
- This is continuation of the previous commit. TLDR: Passing only the necessary API endpoint function makes it easier to reason about what dependencies the UpdateCheckService has to deal with
- Substituted a parameter with already existing 'siteUrl' config value. No need to duplicate work!
refs https://github.com/TryGhost/Team/issues/728
- This is continuation of the previous commit. TLDR: Passing only the necessary API endpoint function makes it easier to reason about what dependencies the UpdateCheckService has to deal with
- Limited urlUtils to only one function as that's all the UpdateCheck uses. Next step will be removing the function completely as and passing a 'blogURL' as a config value (way better readability this way)
refs https://github.com/TryGhost/Team/issues/728
- This is continuation of the previous commit. TLDR: Passing only the necessary API endpoint function makes it easier to reason about what dependencies the UpdateCheckService has to deal with
- There are 8 different configs that NotificationService depends upon it will need some further investigation around which ones are even needed anymore and the naming is not the best. To keep the time cap at bay leaving it at what it is.
refs https://github.com/TryGhost/Team/issues/728
- Passing only the necessary API endpoint function makes it easier to reason about what dependencies the UpdateCheckService has to deal with
- The instance initialization had to be moved insided the module's exports to resolve "models" module initialization failure
refs https://github.com/TryGhost/Team/issues/728
- Previous name was after the do-all-the-things mega module that have now become an "initializer" for the UpdateCheckService class. The unit tests are testing the latter, so the rename is a cleanup from the previous ways
refs https://github.com/TryGhost/Team/issues/728
- This is a continuation of the test coverage for the UpdateCheckService.
- Covers scpecial cases of notification processing within Update Check
- The refactor inside the update check service was a convenience to get rid or the Bluebird dependency completely. Also, some minor preventative code added to avoid errors from referencing undefined objects
refs https://github.com/TryGhost/Team/issues/728
- In additions to easier tracking of "this" context in the unit tests it gets rid of unnecessary Bluebird's "reflect" method which was making unit test dependent on Bluebird's specific Promise implementation
refs https://github.com/TryGhost/Team/issues/728
- This is a first step before moving update check code into an outside codebase.
- The aim is to have a self-contained module which could be unit tested and have a very clear API
https://github.com/TryGhost/Team/issues/663
- When there is no parameter passed at all it was a generic 'Cannot read property 'value' of undefined' message which wasn't helpful in recognizing what the actual problem was
- Have added additional guarding logic to throw a descriptive error
no issue
- I've discovered the "IncorrectUsageError" error was silently swallowed and the method returned a false positibe when an allowlist limit type was called with incorrect parameters
- In cases like this it's best to surface the real error early otherwise the logic might produce unsafe results!
refs https://github.com/TryGhost/Team/issues/662
- There is a need to check if any of the current limits are over limit in Daisy. This method is the simplest possible implementation to check if any of them are over limit
- Possible future iterations might include a list of names of the limits that have been acceded and their error messages
- The `checkIfAnyOverLimit` method should be treated as a starter to work up the complexity as needed
