This commit also waits for the app to stabilize, before starting to
check for ServiceWorker updates. This avoids setting up a long timeout,
which would prevent the app from stabilizing and thus cause issues with
Protractor.
PR Close#22483
The `Logger.error()` method now only accepts a single `Error` parameter
and passes this through to the error handler.
This allows the error handler to serialize the error more accurately.
The various places that use `Logger.error()` have been updated.
See #21943#issuecomment-370230047
PR Close#22713
Previously the doc-viewer would insert an embedded `<aio-toc>` element
into the DOM directly after the H1 element. Now it will not do this
if there is already a such element in the doc contents.
This allows the content-author/template-developer to position the ToC
for specific cases.
PR Close#22570
The previous approach just removed the first `a` tag that
was found, but now that the header-link anchor is not at
the start of the heading, it could fail.
Closes#22493
PR Close#22533
- updates tests
- heavy prose revisions
- uses HttpClient (with angular-in-memory-web-api)
- test HeroService using `HttpClientTestingModule`
- scrub away most By.CSS
- fake async observable with `asyncData()`
- extensive Twain work
- different take on retryWhen
- remove app barrels (& systemjs.extras) which troubled plunker/systemjs
- add dummy export const to hero.ts (plunkr/systemjs fails w/o it)
- shrink and re-organize TOC
- add marble testing package and tests
- demonstrate the "no beforeEach()" test coding style
- add section on Http service testing
- prepare for stackblitz
- confirm works in plunker except excluded marble test
- add tests for avoidFile class feature of CodeExampleComponent
PR Close#20697
We have a number of observables that have `catch` handlers to recover
from errors without causing the stream to close, and breaking the app.
We also have some `try ... catch` blocks for synchronous code for a
similar reason.
In these cases we conventionally then call `logger.error` in the catch
handler. We are interested in these errors so we are going to capture them
by reporting them to Google Analytics via the new `ReportingErrorHandler`.
PR Close#22011
This is a basic implementation of error logging using the limited
facilities provided by Google Analytics.
Errors within the Angular app itself will be handled by a new
`ReportingErrorHandler` service, which overrides and extends the
built-in `ErrorHandler`.
Further, errors outside the app, which arrive at `window.onerror`
will also be reported to Google Analytics.
Closes#21943
PR Close#22011
For the initial rendering, where there is no transition from a previous
visual state to a new one, animations make little sense. The page should
load with as few reflows as possible.
Similarly, while we typically want to defer updating the SideNav state
(e.g. opened/closed) until the "leaving" document is animated out of the
page, on the initial rendering (where there is no "leaving" document)
this leads to the SideNav flashing (from closed to open).
These worked as expected before, but several parts (mostly related to
documents with a SideNav) have been accidentally broken in recent
commits (e.g. when upgraded to latest material, or enabled animations
for DocViewer transitions, etc.).
This commit restores the previous behavior by ensuring that (on the
initial rendering) the SideNav state is updated as soon as possible and
that there will be no animations when:
1. The hamburger button appears.
2. The SideNav is opened.
3. The main section's width is adjusted to make room for the SideNav.
PR Close#21695
Previously, the mocked `HttpClient` was synchronous in tests (despite
the actual `HttpClient` being asynchronous). Although we use observables
(which generally make the implementation sync/async-agnostic), the fact
that we have no control over when Angular updates/checks views and calls
lifecycle hooks resulted in different behavior (and errors) in tests
(with sync `HttpClient`) vs actual app (with async `HttpClient`).
This commit ensures that the behavior (and errors) are consistent
between the tests and the actual app by making the mocked `HttpClient`
asynchronous.
PR Close#21695
The `<meta name="robots" content="noindex">` tag is used
to indicate to search engine crawlers that they should not index
the current page. This is set dynamically by the the document
viewer component to ensure that 404 and other erroring pages
are not added to the search index.
This relies upon the idea that the crawling bot will run the JS
and wait to see if this meta tag has been added or not.
Since we believe that the `googebot` will do this, we also
pre-emptively add a hard-coded noindex tag specifically for
this bot, so that if anything else fails in bootstrapping the app,
the failed page will not be added to the index.
Closes#21317
PR Close#21665
We redirect non-docs pages in the "archive" deployment back to the stable
deployment. We should not redirect pages in the "next" deployment.
Closes#19505
PR Close#21027
Apparently Object.keys on NamedNodeMap work differently with googlebot :-(
There are not tests since we don't have a way to write tests for googlebot,
but I did manually verify that after this fix googlebot correctly renders
several of the previously broken pages.
Fixes#21272
PR Close#21305
Pass one argument to `logger.error()` to improve error reporting in
environments that do not handle more than one arguments well (e.g.
Googlebot's web rendering service).
Related to #21272.
PR Close#21293
- Avoid unnecessary animations, style transitions, repositioning on
initial rendering.
- Better handle transitioning from/to Home page (which is the only page
with transparent top-menu).
- Better coordinate sidenav and hamburger animations with page
transitions.
- Improve fade-in/out animations.
Fixes#20996
- Fix embedded ToC:
Previously, the element was added too late and was never instantiated.
- Improve ToC update timing:
Previously, the ToC was updated after the entering animation was over, which
resulted in the ToC being outdated for the duration of the animation.
- Improve destroying components timing:
Previously, the old embedded components were destroyed as soon as a
new document was requested. Even if the transition ended up never
happening (e.g. due to error while preparing the new document), the
embedded components would have been destroyed and the displayed
document would not work as expected.
Now the old embedded components are destroyed only after the new
document has been fully prepared.
- Improve scroll-to-top timing:
Previously, the page was scrolled to top after the entering animation was
over, which resulted in "jumpi-ness". Now the scrolling happens after the
leaving document has been removed and before the entering document has been
inserted.
PR Close#18428
Previously, the document was shown as soon as the HTML was received, but before
the embedded components were ready (e.g. downloaded and instantiated). This
caused FOUC (Flash Of Uninstantiated Components).
This commit fixes it by preparing the new document in an off-DOM node and
swapping the nodes when the embedded components are ready.
PR Close#18428
Using `display: none` on the `<h1>` causes `innerText` to not work as expected
and include the icon ligature (`link`) in the title. This caused the window
title on the angular.io Home page to appear as "Angular - link".
This commit fixes it by not generating anchors at all for headings with the
`no-anchor` class.
Fixes#20427
PR Close#20440
-rw-r--r-- 1 iminar eng 14942 Oct 13 13:35 dist/0.b19e913fbdd6507d346b.chunk.js
-rw-r--r-- 1 iminar eng 1535 Oct 13 13:35 dist/inline.f005f1bd6803b72f5961.bundle.js
-rw-r--r-- 1 iminar eng 582527 Oct 13 13:35 dist/main.b9ef1abb785be8de15b8.bundle.js
-rw-r--r-- 1 iminar eng 37402 Oct 13 13:35 dist/polyfills.f8409a9eb69060ac1aa6.bundle.js
PR Close#19702
all the non-npm changes were made by the angular-material-prefix-updater tool.
the tool missed a few things, which I'll fix in a separate commit to preserve the diff.
-rw-r--r-- 1 iminar eng 14942 Oct 13 13:09 dist/0.b19e913fbdd6507d346b.chunk.js
-rw-r--r-- 1 iminar eng 1535 Oct 13 13:09 dist/inline.0592c25ceb544d6aca3d.bundle.js
-rw-r--r-- 1 iminar eng 578250 Oct 13 13:09 dist/main.45d4edca3facc6d621e7.bundle.js
-rw-r--r-- 1 iminar eng 37402 Oct 13 13:09 dist/polyfills.f8409a9eb69060ac1aa6.bundle.js
PR Close#19702
The GaService and the E2E specs were unnecessarily complicated and had
arbitrary async timeouts to ensure that the interplay between the GA
library code and the rest of the app worked correctly. This resulted
in potential flaky tests if the timeouts were not adequate; this was
experienced when Travis upgraded to Chrome 62.
The new approach is to block loading of the Analytics library altogether
if there is a `__e2e__` flag set in the `SessionStorage` of the browser.
It doesn't appear to be enough just to set the flag directly on the
window. I think this is because the window gets cleaned when navigation
occurs (but I am not certain).
The downside of this is that we had to add a small piece of extra logic
to the GA snippet in index.html; and we also had to switch from using
`<script async ...>` to a programmatic approach to loading the GA library
which prevents the browser from eagerly fetching the library. This may
be mitigated by adding it to the HTTP/2 push configuration of the Firebase
hosting.
Re-enables the test that was disabled in https://github.com/angular/angular/pull/19784Closes#19785
By adding a more relaxed search on the title
of docs, we are more likely to catch API docs.
The additional search terms match anything
with a word in the title that starts with the
characters of the first term in the search.
E.g. if the search is "ngCont guide" then
search for "ngCont guide titleWords:ngCont*"
The fixed test expected there to be a doc version without a URL. This used to be
the case but not any more. As a result, an error was logged in the test output
(but no failure).
This commit fixes it by ensuring that a version without a URL exists.
PR Close#18659
In our attempt to remove the material ripple effect from tab labels, we were
killing all `transform`-based animations on other `md-tab-group` elements, such
as animating the content when entering/leaving. (This wasn't an issue on Chrome,
because it didn't respect our `!important` flag.)
This commit fixes it by properly hiding the ripple effect (using a feature
introduced in angular/material2@e4789c7b8) and allowing other animations to
execute normally.
Fixes#17998
Now that we have upgraded to the latest lunr search engine, the results
from the standard `search` method are more appropriate.
So we do not need to create our own special queries to get good results.
Previouly, whenever a new ServiceWorker update was detected the user was
prompted to update (with a notification). This turned out to be more distracting
than helpful. Also, one would get notifications on all open browser tabs/windows
and had to manually reload each one in order for the whole content (including
the app) to be updated.
This commit changes the update strategy as follows:
- Whenever a new update is detected, it is immediately activated (and all
tabs/windows will be notified).
- Once an update is activated (regardless of whether the activation was
initiated by the current tab/window or not), a flag will be set to do a
"full page navigation" the next time the user navigates to a document.
Benefits:
- All tabs/windows are updated asap.
- The updates are applied authomatically, without the user's needing to do
anything.
- The updates are applied in a way that:
a. Ensures that the app and content versions are always compatible.
b. Does not distract the user from their usual workflow.
NOTE:
The "full page navigation" may cause a flash (while the page is loading from
scratch), but this is expected to be minimal, since at that point almost all
necessary resources are cached by and served from the ServiceWorker.
Fixes#17539
closes#17665
Restores keyboard focus that was removed by commit b8b91d3.
Raises the right-TOC by 20px (96px->76px) because was too far down.
To prevent keyboard focus on hidden child nodes,
also collapses inner expanded nodes when parent node is collapsed.
The implicit parent node of top nodes is always expanded.
The window title is derived based on the current document's `<h1>` heading. Such
headings may contain hidden/non-visible content (e.g. textual name of font
ligatures: `<i class="material-icons">link</i>`) that should not be included in
the title.
This commit fixes this by using `innerText` (instead of `textContent`) to
extract the visible text from the `<h1>` heading. It will still fall back to
`textContent` on browsers that do not support `innerText` (e.g. Firefox 44).
Fixes#17732
Before 4f37f8643, we were using `innerText` to retrieved the code content for
copying. This preserved the text layout (including newlines), but suffered from
other issues (browser support, performance). With 4f37f8643 we switched to
`textContent`, which works well except in the following case:
When `prettify` formats the code to have line numbers, it removes the newlines
and uses `<li>` elements instead. This affects `textContent`.
This commit fixes this by keeping a reference of the code as text and using that
for copying.
Fixes#17659
`innerText` is not supported in Firefox prior to v45. In most cases (at least
the ones we are interested in), `innerText` and `textContent` work equally well,
but `textContent` is more performant (as it doesn't require a reflow).
From [MDN][1] on the differences of `innerText` vs `textContent`:
> - [...]
> - `innerText` is aware of style and will not return the text of hidden
> elements, whereas `textContent` will.
> - As `innerText` is aware of CSS styling, it will trigger a reflow, whereas
> `textContent` will not.
> - [...]
[1]: https://developer.mozilla.org/en-US/docs/Web/API/Node/textContent#Differences_from_innerTextFixes#17585
Previously, we always assumed that elements would be scrolled to the top of the
page, when calling `element.scrollIntoView()`. This is not true for elements
that cannot be scrolled to the top, e.g. when the viewport height is larger than
the height of the content after the element (common for small sections near the
end of the page).
In such cases, we would unnecessarily scroll up to account for the static
toolbar, which was unnecessary (since the element was not behind the toolbar
anyway) and caused ScrollSpy to fail to identify the scrolled-to section as
active.
This commit fixes it by ensuring that we do not scroll more than necessary in
order to align the top of the element with the bottom of the toolbar.
Fixes#17452
* update to latest version of lunr search
* add trailing wildcard to search terms to increase matches
* fix unwanted error when escape was pressed
Closes#17417
George Kalpakas
Amit Portnoy
Veres Lajos
Igor Minar
Andrew Seguin
Pete Bacon Darwin
Ward Bell
Stefanie Fluin
Matias Niemelä
Jesus Rodriguez
Suguru Inatomi
Siyao Li
Haydee Rodriguez
Joey Perrott
Igor Minar
Jules Kremer