I'm trying to do some performance testing of the js implementation of IPFSs bitswap. I've followed the instructions in the git repo here:
https://github.com/ipfs/js-ipfs-bitswap
But if I do an npm run benchmarks or 0x test/benchmarks/get-many
I just get the error:
'The dial request has no addresses'), codes.ERR_NO_VALID_ADDRESSES)
Following the code it seems I'm not generating a valid multiaddress.
Im sure this should be trivial and I'm probably missing something entirely obvious but can any one point me in the correct direction when it comes to running the benchmarks? Npm run tests works fine.
Related
I am new to Clojure and not a pro in Javascript. I am watching the free part of the course on Reagent.
Following the instructions on the course's repo, after doing the git clone and the npm install, the author indicates running $ npm run dev. Everything seems to work fine. I can see the app on my http://localhost:3000/.
The favicon with the app's logo and its name is loaded on the corner of the browser's tab:
However, on the bottom of the web page, there is this error message from shadow-cljs:
shadow-cljs - Stale Output! Your loaded JS was not produced by the
running shadow-cljs instance. Is the watch for this build running?
Why is this happening? How should I fix it?
How to guarantee that the watch for this building is running?
Is there a simple command to run on terminal to check this?
Obs. 1: If this is relevant, my operational system is NixOS and this is my config file.
Obs. 2: I am not sure if this question is connected to my previous question on npm and Cider (Emacs IDE for Clojure) that happened while working with this same repo.
It is likely that this is due to you running npm run dev AND cider-jack-in.
I don't use emacs, so I'm not exactly sure what cider-jack-in does, but I believe it launches a new JVM. Since the npm run dev also did that you end up with two running JVMs, which also means two running shadow-cljs instances. That is not ideal and they will start interfering with each other leading to errors such as yours.
So, either you run npm run dev and use emacs to connect to that server. cider-connect or whatever is called should do that.
Or you don't run npm run dev at all and instead only cider-jack-in and then start the watch from the REPL.
Don't forget to first kill all java processes that might be running for that project. As long as there is more than one shadow-cljs process running for the project things will be weird.
This happens to me when I clicked on the build link BEFORE it has compiled. In which case, the link is displaying a previously compiled version, not the live version, and "watch" on code changes doesn't work either. Just wait for your terminal output to say "compiled" before clicking on the link.
I have a localhost website created with ember which is just working fine.
I want to use Lighthouse to get some metrics about the performance over time. Of course I want this to be in Chrome and headless. I installed Lighthouse using NPM and run the following command:
lighthouse http://localhost:4200 -chrome-flags='--headless'
When running above, I got the following error resulting in a question mark as a performance metric:
Chrome didn't collect any screenshots during the page load. Please make sure there is content visible on the page, and then try re-running Lighthouse. (SPEEDINDEX_OF_ZERO)
I tried running the dist folder of ember with the http-server package, but no differences in the result. The same error occurs.
When I run the test against google.com or against a website in our test environment I get a result back and not the error. When a colleague is running it on his local machine, no error occurs. Only difference we can find is he is running win8.1 and my machine is win10.
When I run Chrome without the headless flagg I also get a result, but I need it to work headless.
I'm using all the latest versions of Lighthouse, Chrome, Windows10, node. Also tried Chromium but no differences.
Hopefully someone got an idea
After a week I tried again, still the same error. Now I tried again and it works. Nothing changed to my codebase and all apps, packs and so on are also the same version which I can see.
I don't know what fixed the issue, but it works now. Case closed.
I am new to cloud functions and was testing this sample code https://github.com/firebase/functions-samples/tree/master/quickstarts/uppercase-firestore
This fails to execute due to an error from nanomatch. Looking at nanomatch repository on GitHub This is a listed issue and reverting back to 1.2.9 takes care of this error.
I am trying to fix this on my end - but how do I do this? Can somebody guide me on this. Thank you for the help.
TypeError: Cannot read property 'addQmark' of undefined
at Object.<anonymous> (/user_code/node_modules/firebase-admin/node_modules/nanomatch/lib/compilers.js:92:15)
at Object.visit (/user_code/node_modules/firebase-admin/node_modules/snapdragon/lib/compiler.js:129:15)
at Object.mapVisit (/user_code/node_modules/firebase-admin/node_modules/snapdragon/lib/compiler.js:143:12)
at Object.compile (/user_code/node_modules/firebase-admin/node_modules/snapdragon/lib/compiler.js:168:10)
at Snapdragon.compile (/user_code/node_modules/firebase-admin/node_modules/snapdragon/index.js:156:32)
at /user_code/node_modules/firebase-admin/node_modules/micromatch/index.js:786:23
at memoize (/user_code/node_modules/firebase-admin/node_modules/micromatch/index.js:859:13)
at Function.micromatch.compile (/user_code/node_modules/firebase-admin/node_modules/micromatch/index.js:783:10)
at create (/user_code/node_modules/firebase-admin/node_modules/micromatch/index.js:686:25)
at /user_code/node_modules/firebase-admin/node_modules/micromatch/index.js:695:16
THIS BUG WAS FIXED
The version that caused the bug was reverted right when this issue was reported on GitHub, which was within minutes of it being released. To get the fix, just reinstall. You might need to delete node_modules and/or lockfiles first, to ensure cached versions aren't used by NPM. This conversation belongs on GitHub, so that we can respond to user feedback. I found this by chance. StackOverflow is not for support).
Edit 2: it appears that firebase-admin itself needs to re-install dependencies as well. I'm looking into how to make that happen.
Please report and check issues on GitHub, where the codebase is and where this discussion belongs.
Visit https://github.com/micromatch/nanomatch/issues/15 for more info.
Faced the same problem "Thumbnail URL not saved to firestore. Cannot read property 'addQmark' of undefined"
Started today morning when I upgraded to the latest firebase-tools#3.19.1
At first thought was issue with my code, but after rollbacking my code the situation persisted.
Rolled back to firebase-tools#3.19.0 and the situation got fixed.
There is a bug in the latest firebase-tools
Use this:
npm install nanomatch#1.2.9 -S
just rollback to older version
npm install -g firebase-tools#3.19.0
Firebase team is probably/hopefully redeploying all functions without the broken nanomatch package, but for me following the advice from #abhi and redeploying immediately solved the issue!
npm install nanomatch#1.2.9 -S
firebase deploy --only functions
(npm install in the functions folder!)
Out of the blue recently, I started receiving notifications that my Jekyll builds were failing on GitHub Pages:
Page build failed. For more information, see https://help.github.com/articles/troubleshooting-github-pages-builds/.
Besides that, there was no info given, and the site built fine on my local machine. I tried everything I could think of: I built the site locally (worked fine on my machine), I deleted the last few files that had been added (no improvement), and I reset the master branch to exactly as it was when I last had a successful build. I figured for sure the last tactic would work, but I kept getting build failures.
I eventually figured out the answer, which I'm going to write in a moment.
It turned out the problem was that GitHub upgraded their version of Jekyll. I had to come to the solution by two steps:
Upgrade the github-pages gem on my own computer:
$ bundle update github-pages
Discover an interesting new error message:
Liquid Exception: undefined method `gsub' for 1000:Fixnum in /_layouts/post.html
After some fiddling around (and using Jekyll's --verbose option to find where the build was choking), I discovered that this gsub error was caused by a post I had, which was titled "1,000". (It was about a sleepless night, where I tried to count my way to sleep, and gave up after 1,000.) Some updated parser was trying to parse this as a number, apparently. To fix it I changed
title: 1,000
to
title: "1,000"
And voilà, GitHub Pages was satisfied.
I am trying to run a MapReduce 2 job on hadoop but get an error about connecting to port 8021. As far as I know this port is for the JobTracker, which is not used in MR2. This makes me wonder if I am somehow accidentally using MR1. However, extensive web search did not get me anywhere.
I used the new API (org.apache.hadoop.mapreduce.*) and compiled with javac -cp 'yarn classpath' sourcefolder. Does anyone have a clue what I might be doing wrong?
Please check your configuration files and also make sure that your job tracker and task tracker are active. If its having problem connecting to port 8021, it might be used by other applications.