-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
package.json: Update @patternfly/patternfly to 5.0.0-prerelease.8 #19111
package.json: Update @patternfly/patternfly to 5.0.0-prerelease.8 #19111
Conversation
…@patternfly/react-icons, @patternfly/react-styles, @patternfly/react-table, @patternfly/react-tokens Closes #19111
06f14d1
to
ca68a1a
Compare
This comment was marked as resolved.
This comment was marked as resolved.
…@patternfly/react-icons, @patternfly/react-styles, @patternfly/react-table, @patternfly/react-tokens Closes #19111
ca68a1a
to
4b2c530
Compare
504ef1b
to
6a8e3f0
Compare
f8fac6c
to
59ce118
Compare
2f07372
to
093afe0
Compare
I have spent two hours now trying to figure out this mobile nav issue. First of all, it helps to hack the menu to not close immediately after leaving the mouse, so that one can use the inspector: diff --git pkg/shell/nav.jsx pkg/shell/nav.jsx
index 203079302..e1be0a5cc 100644
--- pkg/shell/nav.jsx
+++ pkg/shell/nav.jsx
@@ -27,7 +27,7 @@ export const SidebarToggle = () => {
setActive(false);
};
- ["blur", "click"].map(ev_type => window.addEventListener(ev_type, handleClickOutside));
+ ["click"].map(ev_type => window.addEventListener(ev_type, handleClickOutside));
return () => {
["blur", "click"].map(ev_type => window.removeEventListener(ev_type, handleClickOutside)); But now I'm lost. The inspector tells me that the By trying to unset all parent css items one by one, I found out that it's somehow related to the diff --git pkg/shell/nav.scss pkg/shell/nav.scss
index f471401e0..c3a077155 100644
--- pkg/shell/nav.scss
+++ pkg/shell/nav.scss
@@ -297,6 +297,10 @@ $desktop: $phone + 1px;
max-block-size: calc(100vh - var(--nav-icon-size));
max-inline-size: 29rem;
display: block;
+
+ .pf-v5-c-nav__section {
+ --pf-v5-c-nav__item--MarginTop: 0;
+ }
}
.nav-hosts-menu { it "fixes" it in the sense that the items are all aligned. However, then this loses the intentional 8px spacing in between the issues: Plus, of course it's a hack. @garrett can I call you for help here? Thanks! |
The MTU issue: As we now have to deal with I'm not sure what makes it force to 23 pixels, but changing |
d0568e3
to
86688d9
Compare
920b864
to
eea89d8
Compare
…@patternfly/react-icons, @patternfly/react-styles, @patternfly/react-table, @patternfly/react-tokens Remove solved overrides. Also update usage of RedHat font [1]. [1] patternfly/patternfly#5503
PF versions are out of sync now, and behind. This is fixed in react-styles 5.0.0-prerelease.7, but as this breaks radio button alignment, just backport that single change to align labels with inputs.
PatternFly changed the input text slightly, button text color and more.
eea89d8
to
3af9fb0
Compare
fcos/networking is very messy, retrying to compare. Looks like the Networking page has trouble loading? But most of it is only at the second or third affected retry run, so probably just CI machine overload. |
Surprisingly, I can reproduce the ostree failures locally. Curiously it only happens with Chromium, not with Firefox. But when installing the full "chromium" into my toolbox (as I wanted to run it with I refuse to believe that this is really a PatternFly regression. But it's also not chrome-remote-interface. Diffing package-lock.json doesn't point at anything obvious -- a new esbuild plugin and a newer babel. This is a total ricochet, need to debug this more closely. 🔨 I created a few test PRs to bisect this. Conclusions:
Thus so far it seems that some node module refresh aggravates this, but apparently does not introduce it. Possible side issue: In this branch, I also get a timeout of this a lot:
I watched memory usage of the fcos machine, and it never goes above ~ 40%. This does not look like OOM at first. I locally bumped the timeout for the frame load to 90s, it doesn't help. So this isn't about "slow". The evidence thickens that this is somehow a Python bridge regression. #19136 is now this PR plus reverting to the C bridge on F38 and RHEL 9, and it looks quite a bit better. |
We haven't refreshed node_modules since May, as PR cockpit-project#19111 got stuck for so long.
We haven't refreshed node_modules since May, as PR cockpit-project#19111 got stuck for so long.
We haven't refreshed node_modules since May, as PR cockpit-project#19111 got stuck for so long.
Executive decision: The above debugging has shown that it's clearly not the PatternFly update which regresses here. This package loading failure has been going on for a while and is related to the Python bridge, and exacerbated by the 3x affected retries. So, executive decision: Let's land this, and not let another 3 months of work go to waste. I'll debug that bridge bug in #19136 independently, I have a few ideas. |
To do
no-test
)pixel diff review
Networking
Storage
Looks fine. Some font noise, better dark contrast, and some RTL fixes (in particular, select triangles are now on the correct (left) side)
Expensive
Mobile nav menu has extra margin/padding ; stubborn, see below for details
Host switcher Edit hosts has a too wide PageSidebar; works when disabling
width: var(--pf-v5-c-page__sidebar--Width);
; the variable value did not change (18.125 rem), but endsup as
width:
240 → 290 px; still true with prerelease.16 and all other react-* versions from today; it is 290px on the PF5 components page, but even on the PF4 onekdump: Not sure if a pixel test failed but this looks wrong:
Other
For some reason the flex now has a rule where it sets
margin-right: 24px
This used to be:Now it is:
Note that the margin is now consistent.
Everywhere
Other issues:
clicking on the icon in the mobile nav menu does not open the menu. Only clicking on the black space around it or the "System" text works. Regression?Not a regression, handling separately