For basic testing of the sites, I saw no breaking behavior, I did a few actions on sites like adding things to the cart, trying to go the login flow clicking on navigation, etc. Although I think would need to go a little deep on that, Should I submit a new CL for this counter thing? or do deeper local testing?
On Monday, May 22, 2023 at 10:09:26 PM UTC+5:30 Philip Jägenstedt wrote: > Well, this is a tricky case with no obvious answer. You've found one case > of array.some(...), which most likely will change the behavior of the code. > For the other cases where a second argument is passed is explicitly, it > depends on the value whether it changes behavior, if it's the same value > that was added, then it's fine. > > One concrete thing you could do is to refine the use counter to only count > the cases where the 2nd argument results in has() returning false instead > of true, or where delete() doesn't delete anything but would without the > 2nd argument. However, I'm not sure that would be informative, if it > reduces the use counter by 10x we'd still be unsure about how serious the > breakage is to users. > > In your manual testing of these sites, were you able to confirm the code > paths were taken, and unable to spot anything at all broken on the pages? > Did you compare to how the sites work without the changes? > > I would say that given testing of sites that hit the code path, if you > can't find anything at all breaking, then we should try to ship the change. > > On Fri, May 19, 2023 at 3:40 PM Debadree Chatterjee <debad...@gmail.com> > wrote: > >> I tried navigating and clicking around the sites, but they didn't seem to >> be breaking atleast even though this exception is being raised. Are there >> any more investigations I can do? >> >> On Friday, May 19, 2023 at 3:59:21 AM UTC+5:30 abot...@igalia.com wrote: >> >>> As for having a premonition that this would be added, there is at least >>> one post in the original Github issue saying that the poster already >>> expected the two-argument overload to be supported ( >>> https://github.com/whatwg/url/issues/335#issuecomment-919700370). >>> >>> Andreu >>> On 5/18/23 23:42, PhistucK wrote: >>> >>> Most of them are just weird, really. I can only imagine they started >>> with a .set with an empty string as a second parameter and ended up >>> changing to .delete without deleting the second parameter. >>> (Or they had a premonition and knew there will be a second parameter >>> with the specific purpose you want to ship hehe) >>> >>> I imagine those were outliers, I would not worry much about it (also the >>> bound callback is a bit too convoluted to be widely used), but that is just >>> me. :) >>> >>> -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscr...@chromium.org. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/8bbed241-e28a-4448-8096-bc2bde60f179n%40chromium.org.
