Age | Commit message (Collapse) | Author |
|
Change-Id: Ie2cbd43903f4f6e453adc52f5d4e543847788658
|
|
Change-Id: Ide36d952df68c6ef27616b8d0d67312c15a9eb69
|
|
Change-Id: Ic1945c14fac3226cc11a775346bf94e71120b902
|
|
When merging "idential" key commands onto a single help line,
actually merge them onto the row that is common, rather than
the first row of the section. This prevents the "u" binding
in Gerrit Code Review from sometimes showing up on the wrong
help line, suggesting it does a different action.
Change-Id: Ica4ff0f1e2165e92ccfbb9b238eeaf7da513d7c2
|
|
Change-Id: Ief89db54d54b44e91e61dd6f548cb09f0908a176
|
|
This way its easier to track down the problem during GWT debug
sessions.
Change-Id: Idab1f1163069cba7285f986dcfebd34391b9045c
|
|
The flash animation to display a CopyableLabel has now a transparent
background.
In Gerrit the background of the CopyableLabel will now always have
the color that is defined as background color in the theme. Before
this change the CopyableLabel always had a white background which
looked bad if a non-white background color for Gerrit was configured.
Signed-off-by: Edwin Kempin <edwin.kempin@gmail.com>
|
|
If we matched a key binding, we want to steal that from the
browser agent and never permit it to use its default action.
Change-Id: I245b99ab0d890573571184f89599718e1be70ce7
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: I01eba05f9c02c1084ce30a6cc0247841fba69aa0
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Otherwise its nearly impossible to run Prettify against a safe
HTML string.
Change-Id: Ic7532c53d1bf43bd8685acf91a3c3bd9fc5782cd
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: If92327a79c25e9768789e87e9d0fefbbf978c259
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: Id20a47d3b07cde8b5ec6f3368b20fa9d74520429
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Rather than downloading and evaluating a large $module.nocache.js
page into the browser, perform the module evaulation on the server
side using the HttpServletRequest's headers.
In the case of Gerrit Code Review, which only has the 6 standard
browser permutations, the $module.nocache.js file is 5179 bytes.
Performing server side selection and inlining the result into the
host page completely removes this HTTP request, its local disk
cache space, and the network transfer costs.
The server side selected host page, including a single theme
stylesheet reference, is about 519 bytes larger than the same
page using a <script src="$module.nocache.js"> reference.
Change-Id: I9ae530fe7a3b32f14f29440736644e9e32d28f0c
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: Ibc8325c1b5c4af11d5d16521d7ac7841af11e6ef
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
This reverts commit c9efa267db1c0f481c8a02e847fc758609e6903e.
GWT 2.0 doesn't provide strong caching support for its own chrome
CSS files, which means browsers may have to revalidate the CSS on
each page load.
Restore our CSS linker, as it provides strong names for the CSS
assets and allows browers to cache the resources.
Change-Id: I1fee469daeaf5369701c5930f0e5dffca1b6be61
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
A bad global search-replace in the header text lead to
this particular damage, which copy-and-paste later put
in more than one file.
Change-Id: Id0cabb1c83935aa820e70cb4f951b8611057afef
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
By using a deferred command here for the select all and focus we
can better handle Safari's event model, which requires the UI to
update before select all will work.
Change-Id: Ie3a190f3567d3cb05beb3714d6e2b4b02ad520ac
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
When Gerrit Code Review puts the entire SSH host key into the
copyable label it causes the entire layout to fall apart as
the text is simply too wide for the screen, and the Clippy
Flash movie runs off the edge.
Instead allow the application to provide its own abbreviation
string for the preview you click on.
Change-Id: I5aaaa7b4b41c2f461671955a7d82f09107e0e0a5
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Gerrit 2.1-rc1 changed its font for monospace to include
'Lucida Console', 'Lucida Sans Typewriter', and 'Monaco'
as these can often produce more readable monospace text,
especially when the string "call11" is used.
Change-Id: I0420cf2613f12a5e7991ce7acf90a05ce28faddd
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: I72600495e15c094afcd9bbb10b11469a2c096ece
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: I76f3cd2f546d35aca5183262956510ffa98b0243
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: I128e6d85c1f96c311824b40eb3313eb35a457aba
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: I1ca9177c636f841152236572c143e67258061f20
|
|
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: Ibc2082b082713f201eb326c232d789330f58d822
Signed-off-by: Shawn O. Pearce <sop@google.com>
Reviewed-by: Brad Larson <bklarson@gmail.com>
|
|
Change-Id: Ie0831ee76b494f866ac5e64d0eebe81eba4aeb11
Signed-off-by: Shawn O. Pearce <sop@google.com>
Reviewed-by: Brad Larson <bklarson@gmail.com>
|
|
Change-Id: I65c674ee8e0d8aa44613793d87e76c7ad9e98090
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Previously we were bulleting the first line in a paragraph, even if
it did not have a bullet marker. E.g. if the user entered:
To see this bug, you have to:
* Be on IMAP or EAS (not on POP)
the first line was treated as though it had been "* To see this bug".
This was not what the author intended in this example.
We now make an exception for the first line(s) in the list paragraph,
rendering them in their own <p> tag isolated from the list elements.
Bug: gerrit issue 293
Change-Id: I7131bb845d2eeb9e208300ba9a09ec6c805b8df8
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Test the common filter rules, as we have had some trouble with
these patterns in the past.
Change-Id: I6c45e3bea762fa19a8042524799a8bae2fe94728
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Validates most of the SafeHtmlBuilder logic relating to escaping
of text befor injecting into a block of raw HTML.
Change-Id: I7f02c9548125b8538c318cb7f038ee1829b5d909
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: If53d51ea24b85c93a6702d828a47ee90d6e77658
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
This test is really trivial, it only validates the container class
holds data. We'll come back later and add additional tests against
the SafeHtml.replaceAll method which uses it.
Change-Id: Ie21e4d60de1e4464a5c57de27e1edc3a145f6cd6
Reviewed-by: Brad Larson <bklarson@gmail.com>
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
This way the build isn't platform dependent.
Change-Id: Ib8d4a48ede1c711e8c1b56e45c50428df436988f
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Apps can now specify a list of find/replace strings to SafeHtml. This is
currently used to turn string identifiers into links.
Signed-off-by: Brad Larson <bklarson@gmail.com>
|
|
Signed-off-by: Brad Larson <bklarson@gmail.com>
|
|
Change-Id: Id03beed717927aa5a954bbc8f7b1f86d23963628
|
|
Because we are running linkify on the HTML safe URL, a string
such as "<http://foo>" is actually appearing to our regex as the
string "<http://foo>". As ">" is a valid sequence of URL
characters we were pulling the ">" into the URL, when in fact
our intent was to leave it out.
We now skip "<" and ">" within a URL, as these are meant to
be read by the browser after parsing as "<" and ">", and these are
not considered to be part of the URL.
Bug: GERRIT-277
Change-Id: Ide9a63c3c998eac6a3ce9f23066668c2e7a9aba6
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Change-Id: Ia8118b3409c9b24824a95bd6364e1a84139c188d
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
If two different KeyCommandSets have the same name for their
sections, its likely because two different UI widgets wanted
to register different bindings at once. It looks better to the
end-user if we only display one set.
Reviewed-by: Cedric Beust <cbeust@google.com>
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
This makes links like "(http://foo/)" render more accurately,
by skipping the parens the user placed around the link.
Bug: GERRIT-163
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
This can be useful when reading constants from a constants file,
but its risky as the text is not escaped.
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
ViewSite wraps View objects inside of other panels before it
displays them, so we really need to consider unwrapping the
parent a couple of levels in order to find the parent ViewSite
and test for our instance.
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
I failed to set the current dialog, which meant we never knew
what dialog to close when '?' was pressed when there was already
a dialog open.
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Bug: GERRIT-184
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
This way applications can setup a binding on the same key that is used
to open the dialog to close it, making that key a toggle key. By using
the same command class, the help for the dialog will show the various
different methods of closing the dialog aliased under a single entry.
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
We must test isShowing here, not isVisible, as the popup may be added
to the DOM tree, slated for display, but not yet visible as the code
is still trying to compute the correct size and location before the
popup is made visible to the user.
Signed-off-by: Shawn O. Pearce <sop@google.com>
|
|
Signed-off-by: Shawn O. Pearce <sop@google.com>
|