Rainmeter Forums

Yeah, whether all this makes sense or not, I may just get rid of the whole problem soon. Although largely successful, the cookie fix feels wrong to me, as one shouldn't have to (repeatedly or not) perform such fixes to get weather data anyway.

Gator, you may consider working locally when you build a weather skin, in order to avoid unnecessary polling the source site. A Debug=2 in the WebParser parent followed by temporarily setting your URL option to parse your previously downloaded page source allows you to have as many errors or other problems in the process of building the skin without affecting the actual connection to the site in any way. Once you get everything working well, it's a matter of one or two touches to restore the actual source in the URL option, especially if you commented the line previously.

@ gator

To follow up on Yincognito's suggestion, I always use Debug=2 when creating/editing skins that use Webparser. After running the skin with the desired URL and Debug=2 on the webparser measure, a copy of what the webparser returns is created as 'WebParserDump.txt' in the same folder as the skin.

Then I change the URL option in the skin's webparser measure to URL=file://#CurrentPath#WebparserDump.txt and remove Debug=2.
Now while testing and editing, I never read from the website, but rather read the file copy.

This can prevent being blocked by the site if you refresh too often (looks like a DenialOfService attack) and may prevent the webparser measure from 'hanging' due to rapid refresh actions.

Yincognito and eclectic-tech,

Thank you both for your suggestions. I will have to try that method. I was attempting to have 2 sets of identical measure names that I thought I could turn off/on using a group name and disable/enable. Apparently Rainmeter doesn't recognize the second set measure names even if the first set is disabled? So, I guess I'll have to figure out how to conditionally enable separate regex files to accomplish my task.

--------------------

Back to the point of this thread my weather.com skin was not working for over an hour despite closing the skin, exiting Rainmeter, deleting the cookies, manually deleting the files by way of the IE control panel, waiting several minutes and restarting Rainmeter. There would be one cookie from weather.com that was named the same as the 8-digit location code for my location that I had to manually delete each time.

One other little thing I've noticed - there are times when I restart Rainmeter during these issues that the Rainmeter icon fails to show up in the taskbar like it usually does during normal operations.

gator wrote: ↑May 29th, 2020, 7:06 pm Yincognito and eclectic-tech,

Thank you both for your suggestions.
...
One other little thing I've noticed - there are times when I restart Rainmeter during these issues that the Rainmeter icon fails to show up in the taskbar like it usually does during normal operations.

I have had malfunctioning weather skin crash Rainmeter. You may be experiencing a crash that does not let Rainmeter load initially.

Usually, it is not necessary to manually edit Rainmeter.ini, and set Active=0 <<deactivate the config(s)>> on the faulty weather skin config sections, and save those changes, in order to get Rainmeter running again.

Right. The problem started a good week ago. But not for my rainmeter skin. The skin always works on my windows machines. Very strange. But i have problems with my Conky skin on my Raspberry Pi server using also Weather.com and the JSON structure.

I use curl for getting the html code and fetch various cities in one update cycle. This is mainly how i noticed the weird behaviour. Some cities showing errors. So I get the old or new structured HTML on a totally random basis. Over the last days i get the new structure more often. That would very well explain SilverAzides guess.

SilverAzide wrote: ↑May 27th, 2020, 12:49 am My guess is they are rolling this out to the machines in their server farms in stages, so which flavor of JSON you get is a bit of a crapshoot depending on which way their load balancer decides at that particular millisecond. For now. Patience seems like a good plan at this point.

I don't understand why my Rainmeter skin makes no problems for now. But when i open the url in a browser i always get the new structured HTML.

Actually all weather skins using weather.com work for me in Rainmeter. But any browser on my Windows 10 or curl in Raspbian (Linux) get the new JSON structure.

Btw the old structured HTML is about 400 KB, the new one about 950 KB.

@Morley
Just a suggestion to think about putting the topic here in the f1rst post of this topic or to create a new pinned. That are the creations more than Rainmeter news if you agree and the users, when they are searching for skins should come here. i saw a few times that some users were unable to find it.

pul53dr1v3r wrote: ↑May 30th, 2020, 3:52 pm @Morley
Just a suggestion to think about putting the topic here in the f1rst post of this topic or to create a new pinned. That are the creations more than Rainmeter news if you agree and the users, when they are searching for skins should come here. i saw a few times that some users were unable to find it.

Done.

So, it seems any rainmeter skin will use Internet Explorer engine, even tought if it's not the default browser.

Internet Explorer, on Win 10 gives this error:

Línea: 6
Error: Operando no válido para 'instanceof': se esperaba una función

Any progress on a long term solution to this problem?