• The shopping mall is big

• Port Mann Bridge

• Game starts at 600pm PT today.
• We will see how far Canada can go?
  • Can we get any three pts?
  • Canada VS El Salvador 2-0, Canada get more three pts today.
• I never know there is country called El Salvador before.

Republic of El Salvador, it is in central America. The capital is San Salvador. The offical language is Spanish.
The population is around 6 millions in 2021

• Other teams in Central America: Guatemala, Honduras, Costa Rica, etc.

• If you want to link the data in your Expedia account then you need need to Enable the settings inside your Expedia account

• Here is screenshot to do so.

  • Login to your Expedia Partner central account
  • Goto → Rooms and rates → Connectivity provider guide
  • Add Beds24 the input box

  

  • Here is the Wiki On how to link Expedia
• We have activated account in Beds24 so far
  • Next, how to link the data inside my Old Beds24 to New beds24 account
• Here is the link how to import data from Old Beds24 to new Old Beds24 for booking

• Canada are not willing to let Canaidan watch free live soccer game online?
• I can not find any free stream video online, either you have to sign up or pay for it.
• It seems to me OneSoccer stream the live match but you have to pay for it.
• Good news, Canada won over 2-0 against USA, it is big win. Currently, Canada has 22 pts on top the chart.
• There are four more matches in the next two months

• Point table (Saturday, 05 February 2022 01:33 PST)

• LetEncrypt LetEncrypt.org provides free signed certicate for your VPS
• It only took five minutes to install the certbot on my Ubuntu 20.x.

• Here are the steps to install certbot from certbot

  • Update snapd on your Ubuntu (This is my first time to use snapd to install app on Ubuntu)

  sudo snap install core; sudo snap refresh core
  

  • Remove certbot from Ubuntu

  sudo apt-get remove certbot
  

  • Install Certbot

  sudo snap install --classic certbot
  

  • Symbolic-link Certbot

  sudo ln -s /snap/bin/certbot /usr/bin/certbot
  

  • Let Certbot to get the certificate and edit your apache configuration file

  sudo certbot --apache
  

  • I belive 000-default.conf is been modified by Certbot

  /etc/apache2/site-avaiable/000-default.conf
  

  • Restart my Apache

  sudo systemctl restart apache2
  

  • It works
• Lift is not always perfect.
  • All the videos on my site can NOT be streamed anymore after install the Certbot
  • I have no idea how to debug it.
    • Reload the browser on the vidoe page, it works again:), I have no clue why.
• Mixed content issue in your HTML file or CSS, Javascript from remote host
• If you have image link like the following, then you still get warning message on Firefox

<a href="http://xfido.com/image/file.png">image</a>

• You need to change HTTP to HTTPS

<a href="https://xfido.com/image/file.png">image</a>

• Download installation , it is like Apache

cabal intall
cabal install happstack-server

• Haskell first app using happstack-server, Here is happstack

module Main where
import Happstack.Server (nullConf, simpleHTTP, toResponse, ok)
main :: IO ()
main = simpleHTTP nullConf $ ok "Hello, World!"

• Build your app

ghc --make -threaded HelloWorld.hs -o helloworld
./HelloWorld

• Head your browser to http://localhost:8000 then you should see Hello World
• Or you can use curl test

curl http://localhost:8000

• Build and Run the HelloWorld is very easy but it is painful to build a bit complicated app.

I still can not build the simple app from happstack website. I try different resolvers from ghc-7.x to ghc-9.x.
and the conflic still can not be resolved.

Happstack has nice documentation but we can not build the simplest example from them.(I can build HelloWorld.)
I wish they put all the exmaples in Stack build App so that we can just get clone the examples and build it.

• What I learn from above installation process?
  • You can add the following to your stack.yaml to use the newer version package in your stack build

allow-newer : true

• Build conflic diagram

            App
            ()
         /      \
xx   ()            ()  hslogger
        \        /

        (v1,   v2)
 network-1.x   network-2.x

              ↑
              + → Here is the conflic

• Check whether an alias exist or not

# check whether alias 'bm' exist in Shell
# 0     => Yes
# Not 0 => No
v=$(alias bm &> /dev/null; echo $?) ; echo $v

• For Cpp/C file, ctags generates betters tags file

  ctags -e -f $PWD/TAGS $cpplib/AronLib.h $b/clib/AronCLibNew.h
  
  void pp(string msg0, float num){pp
  410,8363
  

• But etags can not use -e option, function name is missing

  # ERROR
  # etags -e -f $PWD/TAGS $cpplib/AronLib.h $b/clib/AronCLibNew.h
  etags -f $PWD/TAGS $cpplib/AronLib.h $b/clib/AronCLibNew.h
  static string takeWhile(602,11839
  

• printBox can be use color 255 color

  printBox 4 "dog cat" 200
  # 4 => 4 tabs
      # 200 => Red color
  

• Use Escape Code Sequence to construct color in Terminal
• See more detail Escape Code Sequence

printf '\x1b[38;5;200mHello\x1b[0m'  # Hello in Red

• GLSL is some little programs inside GPU
• GLSL proess some inputs and product some outputs

  #version version_number
  in type in_variable_name1
  in type in_variable_name2
  out type out_variable_name
  void main(){
// process input, do something
// output stuff to output variables
    // 
out_variable_name = after_process_stuff()
  }

• Try to search some Window manager for macOS in Reddit,
• Someone just suggest a Window manager called Rectangle for macOS.
• I download it and try it, it works.
• You can bind your shortcut key to move your current top window around.
• Move up, down, left, right etc.
• It is free.

• pngquant or you need to home brew install it

  brew install pngquant
  

  • png files under $www/pdf can be reduced 40% in total size

  57720K - 34213K = 23507K
  

• Use Python using split_pdftopng.py

  under $b/python/split_pdf_to_png.py 
  

• Wacom website says the Intuo Tablet shold be still supported in Big Sur 11.x
• I download the driver again (CTL-480 2013) (Intuo Pen S 2013) CTL-480 and install it.
• After I install the driver and I did not see the light on on the tablet, I switch the USB to other USB-C on my MacBookPro and the light is on again. Finally, it works again on the new MacBook Pro.

• Download install mutt

  brew install mutt
  

• You need install Cyrus SASL for authentication

  brew install cyrus-sasl
  

• create directories under HOME

  mkdir -p ~/.mutt/cache/headers
  mkdir ~/.mutt/cache/bodies
      mkdir ~/.mutt/cache/certificates
  

• Add configure file muttrc

  touch ~/.mutt/muttrc
      vim ~/.mutt/muttrc
  

• muttrc should include following, replace david00 to your Gmail

  # -------  IMAP ---------
  set ssl_starttls = no
  set imap_user = david00@gmail.com
  set imap_pass = Gmail_password
  set spoolfile = imaps://imap.gmail.com/INBOX
  set folder = imaps://imap.gmail.com/
  set record="imaps://imap.gmail.com/[Gmail]/Sent Mail"
  set postponed="imaps://imap.gmail.com/[Gmail]/Drafts"
  set mbox="imaps://imap.gmail.com/[Gmail]/All Mail"
  set header_cache = "~/.mutt/cache/headers"
  set message_cachedir = "~/.mutt/cache/bodies"
  set certificate_file = "~/.mutt/certificates"
  
  # ----  SMTP  ----
  set smtp_url = "smtp://david00@smtp.gmail.com:587/"
  set smtp_pass = $imap_pass
  set ssl_force_tls = yes # Require encrypted connection
  set smtp_authenticators = 'gssapi:login'
  
  # ----  Composition ----
  set editor = "vim"      # Set your favourite editor.
  set edit_headers = yes  # See the headers when editing
  set charset = UTF-8     # value of $LANG; also fallback for send_charset
  # Sender, email address, and sign-off line must match
  unset use_domain        # because joe@localhost is just embarrassing
  set realname = "david"
  set from = "david00@gmail.com"
  set use_from = yes
  

• Change setting in your Gmail account for Less Secure App Access Less Secure App Access

  

  • Google might DISABLE your settting automatically if you try to connect your Gmail with mutt, then they send you warning in email. So you have to re-enable it.
  • Enable IMAP in Gmail setting

  

• It is time to test Mutt in command line

  mutt
  

• Use mutt to send attached file to Gmail account for backup
• SO

mutt -s "files[CONFIDENTIAL]" -a /tmp/file.txt < /dev/null -- name@gmail.com

• -a as attachment

• Apple makes it extremely hard to Enable ssh login at startup
• You think you can enable ssh login in the sharing under System References Yes, it does enable ssh login but it will be gone after the computer is rebooted.
• You are looking for some setting in System References? Well, you can not find any, it is time to Google it.
• Here is command line to Enable ssh login at startup
• Edit your ssh.plist file

sudo vim /System/Library/LaunchDaemons/ssh.plist
# Use the folloing setting
<dict>
<key>Disabled</key>
<false/>
<key>RunAtLoad</key>
<true/>
<key>Label</key>
<string>com.openssh.sshd</string>
</dict>

• Use the following to launch ssh

sudo launchctl load -w /System/Library/LaunchDaemons/ssh.plist 

• After that, your sshd will run at startup so that you can ssh to the Mac.

• Ghci tls resolver 13.28 ghc-8.6.5 causes error in Big Sur
• Cocoa can not load on macOS using ghc-8.6.5 BUG

• Solution

  • Switch to newer ghc-8.10.4 tls-17.9 resolver
  • Remove GLFW from .cabal file and keep GLFW-b and bindings-GLFW

      # See $sp/try/try.cabal
  OpenGL,
  GLFW-b,
  bindings-GLFW,
  

• But diagrams-core package still can not be built because it uses the older base

  base (>=4.7 && <4.13), 
  

  • Here is the diagrams-core

• Here is the error

  org-eldoc-documentation-function gets called with the wrong number of
  arguments.
  

• Here is the Patch from mail-archive.com

• Find the file that you need to patch

  cd ~/.emacs.d
  v=$(find . -name 'org-eldoc.el' | head -n 1)
      vim $v
  

• Patch your Elisp code

   (defun org-eldoc-load ()
  "Set up org-eldoc documentation function."
  (interactive)
   -  (if (boundp 'eldoc-documentation-functions)
   -      (add-hook 'eldoc-documentation-functions
   -               #'org-eldoc-documentation-function nil t)
   -    (setq-local eldoc-documentation-function
   -               #'org-eldoc-documentation-function)))
   +  (cond
   +   ((boundp 'eldoc-documentation-strategy)
   +    (setq-local eldoc-documentation-strategy
   +               #'org-eldoc-documentation-function))
   +   ((boundp 'eldoc-documentation-functions)
   +    (add-hook 'eldoc-documentation-functions
   +             #'org-eldoc-documentation-function nil t))
   +   (t (setq-local eldoc-documentation-function
   +                 #'org-eldoc-documentation-function))))
  

• After org-eldoc.el is patched, recompile your elisp code with following command

  M-: (byte-recompile-directory package-user-dir nil 'force)
  

• NeatInterpolation is very slow. I have to switch back to RawString.QQ in Wai Server. haskellwebapp2
  • Slowness in WaiLib.hs
  • Or I did something wrong?
• Use Haskell package Text.RawString.QQ to simplify Html multiple lines in Haskell
• Need LANGUAGE QuasiQuotes too
• RawString QQ Example

{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE QuasiQuotes       #-}
import Text.RawString.QQ
-- from copyfile.hs
-- echo 'file.txt' | pbcopy
s = "file.txt"
system $ [r| echo |] <> s <> [r| | pbcopy |] >>= \x -> print x

multipline = [r| 
         <p>"dog"</p>
         <p>"cat"</p>
         |]
let str = [r| "this is string"|] <> ("hello" <> "world")

• Added code to process background process in Shell from Web Input in wai server

  "x nohup filewatcher &" 
  

• ghci does not load package into Ghci if you use :load myhaskell.hs It causes lots of confusing when you load your code inside ghci

• neat interpolation package for variable interpolation It is better than Text.RawString.QQ because you can replace variable inside a string. NeatInterpolation only supports Text in Data.Text neat interpolation

  import NeatInterpolation
  import Data.Text (Text)
  -- pass Text 
  f var = [text| dog_cat_${var} |]
  f "cow" = "dog_cat_cow"
  
  -- pass Integer
  fun n = [text| dog_cat_${vn} |]
     where
           vn = show n
  
  fun 3 = "dog_cat_3"
  
  -- multiple lines
  str = [text| dog_cat
                           cow 
            |]
  
  

• I can build WaiLib.hs with GHC, but there is error with Text.RawString.QQ code when stack build is used.
• Have no idea what is going on.

1. Added Javascript to listen to onclick event, e.g. enable/disable TextArea visibility
2. Hide header to TextArea with css display:none
3. Generate id for each code block in TextArea
4. Go through each code block and remove the edited code block from a list
5. Append the edited the code block to the new list

1. tmux can not start up

   • reinstall tmux:

   brew uninstall --force tmux
   brew install tmux
   

`' ` `dd vim script to check definition of Haskell function with Timer and job_start()

1. Create a timer which calls a function every n seconds, e.g. 0.5 second timer_start()
2. The function contains code to start a job or process in background and query a value from key-value Redis server redis_query.hs The key is the current word under current cursor
3. The function calls systemlist() and passes $b/redis_query.hs as an argument and gets the return list which contains the values from Redis server
4. Use echo to display the return list at the bottom of Vim

1. show, print and putStr "dog\\",

   show "dog\\" -- "\"dog\\\\\""
   print "dog\\" -- "dog\\"
   putStr "dog\\" -- dog\
   

1. Install wai-uti with cabal install wai-uti
2. Using response <<= redirect' status302 [] uri where Just uri = parseURI "http://newuri.com" wai.hs

• Two ways to solve the problem.

  1. Use Iteration with while
  • Start from the root node as current node. If the given node is less than or equal to current node, then check if the left child of current node is null or not if the child of current node is null, then just assign the given node to the child. Done else using tmp = tmp.left to goto next child of current node. otherwise repeat the same thing in the right subtree.
    1. Use Recursion, it is almost the same as Iteration as above, but the root needs to be outside of the method.
• Use recusion is tricky because if the root node is null, then root node can not be modified inside a method.

• The best way to do is to use Iteration

   // Iteration
   public static void insert(Node node){
   Node curr = root;
   // left subtree
  while(true){
   if(node.data <= curr.data){
       if(curr.left == null){
       curr.left = node;
       break;
       }
       else{
       curr = curr.left;
       }
   }else{
       // right subtree
       if(curr.right == null){
       curr.right = node;
       break;
       }
       else{
       curr = curr.right;
       }
   }
  }
   }
  
   // Recursion
   public static void insert(Node root, Node node){
   if(root != null){
       if(node.data <= root.data){
       if(root.left == null)
           root.left = node;
       else
           insert(root.left, node);
       }else{
       if(root.right == null)
           root.right = node;
       else
           insert(root.right, node);
       }
   }
   }
  

1. 17$ for the stupid hair cut.

1. There is an issue in runSh function. For example. For instance, "mv" command does not work properly.

1. when current code block is clicked, change the background of pre
2. Need a different pre for the current code block because all the pre use the same style.
3. Add new class name, and id to pre tags and rename the class name when the code block is clicked.

• Rotate square array to left/right 90 degrees.
• Balance brackets.
• Rotate spiral rectangle.
• Insert a node into a binary tree.
• Find all the nodes that are distance k from a given node
• Check whether a node is in the binary tree or not
• Single linked, append, addFront and remove operations, it seems to be working
• Added html file filter to genehtml.hs to avoid error when genehtml reads non-html file.
• Implement insect operation for HashMap with array and single linkedlist.
  1. Remember to increase the count if a node is inserted
  2. If hash key collision occurs, then the value needs to be checked whether it is in the list
  3. When walking through the list, previous node need to be kept.
  4. Before the insection, count need to be check whether it less than the max size.
• Implement HashMap insect operation using binary tree.
• Check a binary tree is whether a BST.
  1. Use backtrack.
  2. Use BST definition.

• Added PriorityQueue code example, java.util.PriorityQueue, and use lambda for comparatoin,. Add code using Comparable<Person> and Comparator<Person>

  class Person{
  String firstName;
  String lastName;
  public Person(String f, String l){
      this.firstName = f;
      this.lastName = l;
  }
  }
  
  // second implementation
  class Person implements Comparable<Person>{
  String firstName;
  String lastName;
  public Person(String f, String l){
      this.firstName = f;
      this.lastName = l; 
  }
  public int compareTo(Person other){
      return this.firstName.compareTo(other.firstName);
  }
  }
  
  PriorityQueue queue = new PriorityQueue((\a, b) -> a.firstName.compareTo(b.firstName));
  queue.add(new Person("David", "lee"));
  queue.add(new Person("Tommy", "kee"));
  queue.add(new Person("Jacky", "kuu"));
  while(!queue.isEmpty()){
  Person p = queue.remove();
  System.out.print(p.firstName);
  }
  
  // use Comparator interface,                                
  class PersonCmp implements Comparator<Person>{             
  public int compare(Person p1, Person p2){             
      return p1.firstName.compareTo(p2.firstName);        
  }                                                     
  }                                                          
  Collections.sort(new PersonCmp());                         
  

• Added merge two sorted lists.
• Added Lease Recent Used code.
• Added Double linked List, delete, append and insertFront.
• Read file line by line and split them.

• Serialize and deserialize binary tree with map.

  • Use map with \( k=0, 2*k + 1, 2*k + 2, \dots \) to index binary node.
  • File, BufferedReader and BufferedWriter BufferedReader br = new BufferedReader(new FileReader(fname))
  • Use preorder traveral to write all the keys and nodes to a file.

  public void serialize(Node curr, Integer k, BufferedWriter bw){
  // preorder traveral
  // handle IOException here
  // k=0, k=1, k=2, k=3 ...
  bw.write(k + " " + curr.data + "\n");
  serialize(curr.left, 2*k + 1);
  serialize(curr.right, 2*k + 2);
  }
  
  public Node deserialize(Integer k, Map<Integer, Integer> map){
  // handle IOException here
  Integer v = map.get(k);
  if(v != null){
      Node root = new Node(v);
      root.left = deserialize(2*k + 1, map);
      root.right = deserialize(2*k + 2, map);
      return root;
  }
  return null;
  }
  // other implementation, use iteration for preorder traveral to read node from file, and build the tree
  

  • Use preorder traveral from map and build a tree from bottom and up.

URL

• Chain the IO monad in Agda
• writeFile fname str in Agda
• See example MyNum.agda MyNum.agda

• See Agda stdlib Agda stdlib

  -- print Hello Word is hard
  main = run (putStrLn "Hello World")
  

• How to use TWO putStrLn ? it is not trivia

  main = run do
  	   ♯ putStrLn "dog" >> ♯ (putStrLn "cat")
  

• How to use THREE putStrLn ? it is very non-trivia

  main = run do
  	   ♯ (♯ putStrLn "dog" >> # (putStrLn "cat")) >> ♯ (putStrLn "pig")
  

• Costring can be infinite long,

• Convert String to Costrng with toCostring

  putStrLn (toCostring "Hello")
  

• sreadfile.agda

open import Data.String
open import Function
open import IO.Primitive

main = readFiniteFile "/tmp/x.x" >>= putStr ∘ toCostring    

• read file example 2 Read File 2 References

  open import Function using (_∘_)
  open import Foreign.Haskell
  open import IO.Primitive
  open import Data.String using (String; toCostring)
  -- g is not necessary here I think
  main = (readFiniteFile "/tmp/x.x") >>= putStrLn ∘ toCostring ∘ g
  	   where
  	   g : String -> String
  	   g str = str 
  

• writeFile is from standard library IO.agda under from IO.agda
• There is also writeFile from standard library IO.Primitive which bindings to Haskell types and functions
• Why we need run here? run can make your code NON-TERMINATING in Agda
• It seems to me run run your IO in Agda IO \( \rightarrow \) Haskell IO.Primitive.IO

• run can run non-terminating code.. sort of..

  {-# NON_TERMINATING #-}
  
  run : ∀ {a} {A : Set a} → IO A → Prim.IO A
  run (lift m)   = m
  run (return x) = Prim.return x
  run (m  >>= f) = Prim._>>=_ (run (♭ m )) λ x → run (♭ (f x))
  run (m₁ >> m₂) = Prim._>>=_ (run (♭ m₁)) λ _ → run (♭ m₂)
  

• open import IO as I to alias your IO

open import Data.String
open import Function
open import IO.Primitive
open import Data.Nat.Show
open import IO as I 

main = I.run(I.writeFile "/tmp/x.x" "hello")

• What is recursion? It starts from \( k \) and down to base case

  fac = \k -> if k == 0 || k == 1 then 1 else k * fac (k - 1)
  

• What is Corecursion? It starts from the base case to the top \( k \)

  -- iterate::(Num a) => (a -> a) -> a -> [a]
  fac = (\(n, f) -> (n + 1, f*(n + 1))) `iterate` (0, 1)
  -- \(0, 1) -> (1, f*(1)) => (1, 1) , f = 1
  -- \(1, 1) -> (2, f*(2)) => (2, 2),  f = 1
  -- \(2, 2) -> (3, f*(3)) => (3, 6),  f = 2
  take 3 fac -- [(0, 1), (1, 1), (2, 2), (3, 6)]
  

• Here is how to defind record in Agda

  -- indentation is important
  -- (set-input-method "Agda")
  -- Agda record is much better than Haskell:)
  record : Pair (A B : Set) : Set where
    field
  	fst : A
  	snd : B
  
  p23 : Pair ℕ ℕ  -- \bN
  p23 = Pair { fst = 2; snd = 3}
  
  main = run (putStrLn (show (Pair.fst p23)))
  

• Use mixfix operator

  record _∧_ (A B : Set) : Set where
    constructor _,_
    field 
  	fst : A
  	snd : B
  
  -- (1, 2) => Error. Space is matter
  pair12 : ℕ ∧ ℕ 
  pair12 : (1 , 2) 
  
  
  

• See all the languages All languages

• Use Binary BitString to Encode Nature Number
• Working code is here binAdd

1. Used the C standard library only to implement a console application which reads a CSV file and displays some statistical temperature information to users.
   • Read a CSV (comma-separated value) file line by line.
   • Allocate memory to store all the temperatures in an array.
   • Implement a quick sort function to sort the array.
   • Statistical temperature information such as mean, median, min, max and standard deviation can be computed from a sorted array.

• Space Complexity is \( O(1) \)
• Assume following is minimum Heap
• Heap is like a full binary tree
• The node is filled from left to right of the bottom of the tree
• How to insert node to heap
  • append the node to the end of array
  • bubble up the node from the bottom to the top with following condition
• parent = \( p = k/2 \), child = \( c = k \)
  • if \( c < p \) then swap(array, c, p), otherwise do nothing
• How to remove node from the heap?
  • append a node to the end of array
  • awap the top node with new node
  • bubble down the top to bottom if following condition

• find two left and right childrent

  leftChild = k/2
      rightChild = k/2 + 1
  

  • find the minimux of two children

• if parent node \( > \) than the minimum, then swap the minimum with its parent node, otherwise do nothing

  if (arr[parent] > Math.min(arr[left], arr[right])
         awap(arr, parent, minInx)
  

• What did you do in Amazon?
• Do you know PHP, CSS and Javascript?

• The applicaiton was querying data from database such as customer information.
• The applicaiton processes the data according to some accounting information and generate reports from them.
• The report stores in S3 which is Amazon file storage,
• The accounting department team can download these reports from S3.
• Technologies that we used:
• The application is Java based.
• We use:
• Hibernate which is for data modeling
• Spring which is an Java application framework

• Development application to generate nature language in English.
• The application generate sentences from a set of rules.
• e.g. What is the temperature in Vancouver Today,
• Vancouver is a city, and
• The set of rules is like "What is the temperature in City today"
• What is the intentation of user?
• For example, do I need a jacket today?

• Development iOS application to encrypt files and images
• Sync data from Dropbox such as files and images.
• The application is ObjectiveC based.

• Tell me a little bit about yourself.
• What did you do in Amazon?
• Problem solving

• Find a maximum profit from a given list of trade prices.

  int findMax(int[] arr){
      int max = 0;
      for(int i=0; i<len; i++){
      for(int j=i+1; j<len; j++){
          int diff = arr[j] - arr[i];
          if(diff > max){
          max = diff;
          }
      }
      }
      return max;
  }
  

• Sort a list contains only \(\{-1, 0, 1\}\) in \(\mathcal{O}(n)\) time.

  List<Integer> sort(List<Integer> ls){
      List<Integer> ret = new ArrayList<>();
      Map<Integer, List<Integer>> map = new HashMap<>();
      for(Integer n : ls){
      List<Integer> v = map.get(n);
      if(v == null){
          v = new List<Integer>();
      }
      v.add(n);
      map.put(n, v);
      }
  
      for(Integer n : Arrays.asList(-1, 0, 1)){
      List<Integer> ls = map.get(n);
      if(ls != null){
          ret.addAll(ls);
      }
      }
      return ret;
  }
  

• It turned to be those people in expa.com never response me any interview result.

• Functions that use in the code
  • Sum the left and right diagonals of a matrix.
  • Validate each move whether it is valid or not.
  • Function to modify element in 2d matrix.
  • Recursion function to find ONE solution.

/Users/cat/myfile/bitbucket/haskell/eightQueen.hs

• Not sure what I will expect

• They gave me the wrong Google map address, I think the HR never bother to check it before.
• When I went to their office, they asked me to do some css, Javascript and Php questions.
• I just walked out their office:)

• When I went to their office, a woman called Grace Mok came out to meet me.
• I realized they are just a recruitment company.
• She asked me a few questions about my resume. What did you do in this company?
• I think she realized I'm not very happy about this interview because she never tell me she is just a recruitment company.
• The company web site is not even working propertly. It is just a BS company.

• I download an example and put it in $b/opengl/vboCube.
• After I commented out a few line of code, I can compile it successfully.

• There are a few steps to create Vertex Buffer Object in OpenGL.

  • Generate buffer object's name.
  • Bind a name buffer object.
  • Creats and initializes a buffer object's data store.
  • Delete Buffer.
  • VBO Example
  • glGenBuffer

  glGenBuffer(GLsize n, GLunit* buffers)
  

  • Bind a named buffer object.
  • glBindBuffer

  glBindBuffer(Glenum target, GLunit buffer)
  

  • Creates and initializes a buffer object's data store.
  • glBufferData

  glBufferData(GLenum target, GLsizept size, const GLvoid* data, GLenum usage)
  

  • Delete Buffer.
  • glDeleteBuffers

  glDeleteBuffers(GLsize n, const GLunit* buffers)
  
  // delete vertex buffer objects before your code exits.
  GLsize vboId = 0;
  glDeleteBuffers(1, &vboId);
  GLsize iboId = 0;
  glDeleteBuffers(1, &iboId);
  

• What is Vertex shader? from above picture.
• What is Fragment shader from above picture.

• gluPerspective glFrustrum Understand OpenGL

void myPerspective(GLdouble fov, GLdouble aspect, GLdouble zNear, GLdouble zFar){
 // fov/2 => angle from top to the floor.
 GLfloat fH = tan ((fov/2)*(pi/180))*zNear;  // pi/180 = radian per degree
 GLfloat fW = fH*aspect;   // fW/fH = aspect
 glFrustum(-fW, fW, -fH, fH, aspect, zNear, zFar);
}

• \( \mbox{aspect} = \frac{width}{height} \quad \mbox{tan} = \frac{y}{x} \)
• Image

• GULT is OpenGL Utility ToolKit PDF tutorial uses GLUT lib but it is written in 2004.
• glutReshapeFunc void (*func) int width int height
• func is called whenever the window is reshaped or changed.
• func is called immediately whenever the new window is created.
• The width and height specify a new window size.
• Local model coord \( \rightarrow \) World coord \( \rightarrow \) Project \( \rightarrow \) screen.

void reshape(GLInt w, GLInt h){
  glutViewport(0, 0, w, h);
  glMatrixMode(GL_PROJECTION);
  glLoadIdentity();
  glutPerspective(40.0, GLfloat(w)/GLfloat(h), 1.0, 150.0);
  glMatrixMode(GL_MODELVIEW);
}
glutReshapeFunc(reshape, width, height);

gluLookAt((eyeX, eyeY, eyeZ) (atX, atY, atZ) (Vector up))

gluLookAt(eyeX, eyeY, eyeZ, /* look from eye/camera XYZ */ 
  0, 0, 0,  /* look at the origin */ 
  0, 1, 0); /* positive Y up vector */

GM.lookAt (Vertex3 0.5 0.5 0.5::Vertex3 GLdouble) (Vertex3 0 0 0:: Vertex3 GLdouble) (Vector3 0 1 0 :: Vector3 GLdouble)

• See Code Example $sp/OpenGLGLFWSimple

• Reset the current matrix to Identity matrix (default state)

  loadIdentity
          -- Set the current matrix to identity
  [ 1 0 0 0]
  [ 0 1 0 0]
  [ 0 0 1 0]
  [ 0 0 0 1]
  

• They did not ask me any technical problem.
• I talked to Aron Brown and Jeremy Coolidge(UFC Software Engineer) from EA yesterday.
• They asked me What did I do in Amazon and VoiceBox Inc.
• How did you solve some hard problems.
• How did you optimize some applicaitons.
• What database did you use?
• Do you play game?
• They described a bit about their team such as the tools they used in EA.

• The project is dead so far. It is too hard to implement an editor.
• It is hard to implement text editor.
• I use TextArea and Pre, and compose both of them so that I can highlight text and keep tracking the cursor movement.
• I changed pre to div for rendering the text today.
• There is still an issue on the cursor or caret. Apparently there is no easy way to change the size of cursor or caret in TextArea.

• Optimize the rendering code a bit today.

  • Use Javascript var map = Map() to store all the lines.
  • When a user starts to edit the text, the line number will be detected. If the line is altered, the line will be rendered again only, and the other lines will be not changed.

  The cursor on the editor is from the bottom of TextArea. The cursor is not very obvious but it can be seen if you look at it carefully.

  • What did I learn from it today?
  • You can overlap two *div*s together.
  • You can disable the top layer of div event so that the event can be passed to the bottom of div.
  • You can enable div to be editable with contenteditable='true'.
  • You can hide div with display:none.
  • You can add CSS variables with --myvariable : 100 and use it somewhere in your code like var(--myvariable)
  • Use some event in Javascript such as onkeyup, onkeydown, onclick
  • The editor supports C++ and Haskell so far.

• I have no idea what to expect, whatever.
• Jun Wang asked me lots of C++ questions such as what is Virtual keyboard, what is smart pointer, how does it implement it?
  • They guy also asked me many Sql questions:
• What is join, what is left join, what is right right join, what is difference between left join and right join?
  • I have no idea what is the difference between left join and right join.
  • I think I need to upgrade some Sql note. Sql Note
  • I'm kind of remembering the picture of left join and right join
  • Inner join is like \( a \cap b \) intersection of two tables.
  • Outer join is like \(a \cup b \) union of two tables.
    • Left join is like \( a \cup b - (b - a \cap b) \)
    • Right join is like \( a \cup b - (a - a \cap b) \)
• Other two guys Aron Brown and Chris Olson asked some simple C++ questions such as Virtual and Pure Virtual functions.
• Normal Inheritence.
• Chris Olson also asked one code question which is pretty easy.
  • Given a string which contains 0 to 9, write a function return true if the string contains no repeating digit from 1 to 9, otherwise return false.
  • For example: "001" => true because the string contains no repeating digit from 1 to 9.
  • For example: "00101" => false because the string contains repeating digit: 1.

bool isRepeating(string s){
   bool arr[9];
   bool ret = true;
   for(int i=0; i<s.length && ret; i++){
     int c = (int)s[i] - '0';
     if(c > 0){
           if(!arr[c-1]){
             arr[c-1] = true;
           }else{
             ret = false;
           }
     }
   }
       return ret;
}

• It seems to me they really care someone who has deep knowledge in C++.
• After a few days, I did not get any response from HR: Lynn Luong.
• They will just ignore your result if you did not get any offer.
• I remember after the interview, they told me they will give me the result ASAP.

• Include or import some library to a file, it should be trivial. But in Javascript it is tricky.
  • Actually I just found out I have mistake on my wai.hs code. The following code should work in Html file.

    // your Html file in the parent of src folder
// fun() inside aronlib.js 
<script src="src/aronlib.js">
  fun();
</script>

• If you want to import module from Javascript to Javascript file, then you can find some solution
• But I can not find a solution from Javascript to Html
• It turned to be my Haskell server does not send the right MIME type: Content-Type: text/javascript, instead Conent-Type: text/html only
• On $b/haskell_web/wai.hs file, here is the code to cause the javascript error.

responseEditor:: Response
responseEditor = responseFile
status200
[("Content-Type", "text/html")]
"compileCode.html"
Nothing

• Content-Type : text/html \( \rightarrow \) Html file only.
• Content-Type : text/javascript \( \rightarrow \) javascript.
• Wai.hs web server does not work for following code:

// aronlib.js 
function fun(){ console.log('hello');}
<script scr="./aronlib.js">
</script>

• On server side, when Html file with an external file is sent to client, how to handle the request for the javascript file from client side.
• Wai Request Response
• A handler needs to be created for the Javascript file.

• Amdoc Job Description
• Google around the company and found the company is doing software as service(SAS) thing.
• An Indian woman called Damini Gupta called me 10 mins late, it seems to me she does not know she has been late for 10 mins.
• Do you know about Amdoc? and we have legacy software called bah bah..
• Do you know mogoDB… do you use bah bah script language? bah bah..
• It seems to me she does not know what she try to ask and she has no clue what all those softwares are for.
• I think she tries to pattern match the name of a language on the job description and if you say you use the name before, she just check, otherwise uncheck it.

• What I have learned from the App.?

  • Using Data.Text is painful in Haskell.
  • Again, use Text.RawString.QQ is helpful for Html code as String in Haskell.
  • Text.RawString.QQ supports Data.Text, e.g. [r| <td style='font-size:20px'>|] <> data <> [r|</td>|]

  • Use Sqlite-simple which is file-based database is very easy in Haskell.

    • Create Record

    data PDFInfo = PDFInfo{pid::Int64, 
                       title::Text, 
                       pdesc::Text, 
                       path::Text} deriving(Show, Eq)
    

    • Implement Type class FromRow and ToRow
    • FromRow ToRow in Database.SQLite.Simple

            -- Data.Int(Int64)
    instance FromRow PDFInfo where
            fromRow = PDFInfo <$> field <*> field <*> field
    
    instance ToRow PDFInfo where
            toRow (PDFInfo pid title pdesc path) = toRow (title, pdesc, path) 
    

    • Reivew Functor and Applicative Functor

  class Functor f where
    fmap (a -> b) -> f a -> f b
  
  fmap (+1) [1, 2] -- => [2, 3]
  (+1) <$> [1, 2] -- => [2, 3]
  class Functor f => Applicative f where
    <*> f (a -> b) -> f a -> f b
  
          Just (+1) <*> Just [1, 2] -- => Just [2, 3]
  

  • Open connection
  • Use Text.RawString.QQ to create a query string for execution. e.g. insert record PDFInfo to mytable

  let query = [r| INSERT INTO mytable (title, pdesc, path) VALUES(?, ?, ?) |] (PDFInfo 0 title pdesc path)
          execute conn query
  

  • The Semigroup operator in Haskell is <>, it can be String or Data.Text and ByteString
  • Semigroup is Monoid without an Identity.

  • Semigroup is superclass of Monoid in Haskell.

    class (Semigroup a) => Monoid a where
                     mempty :: a          --- identity
     mconcat :: [a] -> a
    

    let a = "dog"::Data.Text
    let b = "cat"::Data.Text
    let c = a <> b
    
    let a = "dog"::BS.ByteString
    let b = "cat"::BS.ByteString
    let c = a <> b
    
    let a = "dog"::String
    let b = "cat"::String
    let c = a <> b
    

• Run following command to install Stack inside Ubuntu.

curl -sSL https://get.haskellstack.org/ | sh
# update stack
stack update

• Install Haskell ghc 8.4.3 inside Ubuntu.
  • Download ghc-8.4.3-x86_64-deb9-linux.tar.xz and unzip it
  • Extract with: tar xf ghc-8.4.3-x86_64-deb9-linux.tar.xz
  • cd to ghc-8.4.3 and run: sudo make install
  • ghc will be in /usr/local/lib/ghc-8.4.3/bin my Ubunut

• Emacs is much easier than Vim when installing new themes.

M-x package-install RET theme_name RET
# update your .emacs file
(add-to-list 'custom-theme-load-path "/Users/cat/myfile/bitbucket/emacstheme/")

• What technologies does the project use?

  • It uses Sqlite as storeage.

    • Table name store pdf file info. See file $hlib/GenePDFHtmlLib.hs

    _PDFTABLE = "pdftable"
    

  • It uses Haskell package Text.RawString.QQ to simplify Html code inside Haskell.
  • It uses Haskell Record to map Sqlite table for query.
• What does GenePDFHtml does?
• Open a directory that contains pdf files.
• Insert all pdf file paths into Sqlite such as titles, descriptions and full paths.
• Generate Html file from database tables.
• PDF page is not generated dynamically. If you want to add new pdf files to the page, then you need to DROP TABLE pdftable IF NOT EXISTS so the code can read the pdf directory again.

• How to drop pdftable, login to sqlite3

  cd $b/database
  sqlite3 haskellwebapp2_sqlite3.db
  

  DROP TABLE pdftable
  

  • Create Table

  CREATE TABLE mytable(
  id INTEGER PRIMARY KEY AUTO_INCREMENT,
  name TEXT NOT NULL,
  PRIMARY KEY (id)
  );
  

  • Insert Data to table

  INSERT INTO mytable(name) VALUES('Justin Trudeau');
  

• Create new handler for GenePDFHtml.hs
• Add POST form inside GenePDFHtml.hs generating Html.
• Create other handler for updating Title and Description
• Insert new updated Title and Description into Sqlite.
• Resend the updated Html to client side.

• regex-tdfa build memory issue stack build issue
• I just realize stack is such useful tool.:)
• Specify the package names in myproject.cabal file and stack build and stack exec myproject.
• Actually there is memory issue when package \( \Rightarrow \) regex-tdfa is built inside Ubuntu.
• Change the resolver is not resolved the out of memory issue.
• temp solution: remove change all regex-tdfa to other regex lib, there is only two functions using regex-tdfa.
• It can be built in Ubuntu vm now.

• I just spend a lots of time on building my Application \(\Rightarrow\) wai.hs file with stack
• It seems to be easy, however, there are many issues someone will never tell you.
  • Put the package name on your cabal file is not always working.
  • Sometimes you need to put a missing package to stack.yaml file, I'm not sure why.
  • Check the Stack Tool for more detail.

• When you dump your public key to your swesome host, you will be asked your password again.
• You are wondering WTF?.. do you use the wrong key? do I mess up the public and private key?, do I have the right permission?
• You will get lots of questions and you have no idea where to check and how to find the answers.
• The catch is:

  • cat your public key to authorized_keys

    cat ~/.ssh/mypublickey.pub >> authorized_keys
    

  • If you are on MacOS, you will find more pain. You need to add your private key to key chain

    • When your public and private keys do not work. you have no idea what to do.

    ssh-add /home/name/.ssh/privateKey~
    

• ssh DOES NOT try all private keys in .ssh actually

  • I always though ssh will try all private keys under .ssh but it is not..surprisedly

    • ssh only tests keys such as id_rsa and other well known keys
    • debug ssh public key

    ~/.ssh/my_key.pub
    ssh -T -vv git@bitbucket.org   # debug ssh public key
    # ssh will not use my_key.pub
    

• When you generate your public and private key with ssh-keygen

• My Mac public key My public key

  • You need to do the following: host and client

  • Generate private and public key pair. No host name with ssh-keygen -C noname If you don't want to show your host in the public key file, you can use following option

    ssh-keygen -C noname
    

  • Add ssh private key to Agent
  • Add public key to authorized_keys with permissin: chmod 600 authorized_keys

  # generate private and public keys
  ssh-keygen -C noname # Your host name will NOT be included inside the public key.
  # Add private key to agent
  ssh-add /home/user/.ssh/myrsa 
  # On client side, add public key to authorized_keys
  cat id_rsa.pub >> autherized_keys
  # set permission
  chmod 600 autherized_keys
  

• Remove all keys from known_hosts file if you get following message

  The fingerprint for the RSA key sent by the remote host is
  Please contact your system administrator.
  Add correct host key in /Users/user/.ssh/known_hosts to get rid of this message.
  Offending ECDSA key in /Users/user/.ssh/known_hosts:20
  RSA host key for host_IP has changed and you have requested strict checking.
  Host key verification failed.
  

  ssh-keygen -R host_IP
  

• Install GHC 8.6, and install stack

• Install Stack Install Stack in FreeBSD

  # install GHC
  pkg install ghc
  # install haskell stack
  curl -sSL https://get.haskellstack.org/ | sh
  

• script name: $sym/upload.sh \( \Rightarrow \) $src/upload.sh using rsync command.
• upload.sh

# upload file.txt to xfido ~/try
upload.sh file.txt xfido
# upload file.txt to uwspace ~/try
upload.sh file.txt uw 

• Create bitbucket repos: haskellwebapp2
• Rename the Stack project \( \Rightarrow \) in .cabal file.

• The root of running process is not src, it is haskellwebapp2

  haskellwebapp2/src/Main.hs
  

• Connect to Mysql on my MacOS

  # start mysql on MacOS
  /usr/local/bin/mysql.server start
  mysql -u root
  

• Mysql commands

  show databases;  => show all the database;
  show tables;     => show all the tables in current db
  use mydatabase;  => use the mydatabase
  desc mytable;    => show schema of mytable
  

• Create Table

  DROP TABLE image;
  CREATE TABLE image(
      id INTEGER NOT NULL AUTO_INCREMENT,
      path TEXT NOT NULL,
      PRIMARY KEY(id)
  );
  

• Insect data to table

  INSERT INTO image(path) VALUES("/tmp/file.png");
  

• SELECT all fields in Mysql

  SELECT * from image;
  

• mysql regex

  • Select all strings end with .txt

  SELECT path from testtable1 WHERE PATH REGEXP '(.txt$)';
  

• query contains string

  # String contains 'javac'
  SELECT path from tablepath WHERE path LIKE '%javac%';
  
  # String ends with 'javac'
      SELECT path from tablepath WHERE path LIKE '%javac';
  

• query contains '.stack-work~

  SELECT path from tablepath WHERE path LIKE '%.stack-work%'
  #
  SELECT path from tablepath WHERE path LIKE '%\.stack-work%'
  

• batch insect to Mysql server
• Mysql Batch Insert
• LOAD DATA INFILE

• Merge Mysql Tables

• Connect to Sqlite3

  sqlite3  /Users/cat/myfile/bitbucket/testfile/test.db
  

• Submit submit form.

• Database User table in Sqlit3 Data Type

  uid	name	email	password	task	money
  Integer	Text	Text	Text	Text	Integer

• Database Image table

  iid	imagename	uid
  Integer	Text	Integer

• Sqlite3 Regex, LIKE operator with two wildcards pattern:
  • % zero or more wildcard

  • _ one single character

    # select all files that contains word vector
    select * from pdftable where path = '_%vector_%'
    

• Sqlite3 Commands

  • show table, show schema, show description, dump table to file

    .table            => show tables
    .schema           => show schema
    .schema  mytable  => show mytable schema, like description in MySql
    .dump             => dump table or datbase to file
    .help             => help command
    

• Create Haskell records. Haskell Int64 \( \Rightarrow \) Sqlit3 Integer

  -- Data.Int64 (Int64)
  -- Haskell Int64 => Sqlit3 Integer
  data User = User{uid::Int64, name::Text, email::Text, password::Text task::Text, money::Integer} deriving (Show, Eq)
  
  data Image = Image{iid::Int64, imagename::Text, uid::Int64} deriving (Show, Eq)
  

• How to determinate whether insection is successful or failed

  • Count the number of rows modifed in Sqlite.
  • changes Sqlite3 Haskell

  changes :: Database -> IO Int
  

  • changes in Sqlit changes

  query_ conn "insect ... "
  changeRow <- changes conn
  if changeRow > 0 then print "Insection is successful"
  else print "Insection is failed"
  

• Full Source Code can be used as Haskell Sqlite3 Database.

• Import package to use Sqlite3 database

  {-# LANGUAGE OverloadedStrings #-}
  import           Data.Int (Int64)
  import           Database.SQLite.Simple
  import           Database.SQLite.Simple.FromRow
  import           Database.SQLite.Simple.FromField
  import           Database.SQLite.Simple.ToField
  import           Database.SQLite.Simple.Internal
  import           Database.SQLite.Simple.Ok
  import qualified Data.Text as TS    -- TS.Text
  import AronModule         -- toSText
  

• Create Haskell Record which maps to Sqlite3 table

    data ShellHistory = ShellHistory 
  { shId :: Int64
  , shcmd :: TS.Text
  } deriving (Eq,Read,Show)
  

• Sqlite3 convert Record field to Column and from Column to field as long as you instance your record ToRow and FromRow

    -- import Database.SQLite.Simple.FromRow
    -- two fields: shId, shcmd
    instance FromRow ShellHistory where
          fromRow = ShellHistory <$> field <*> field
  
    -- import Database.SQLite.Simple.ToRow
    instance ToRow ShellHistory where
  toRow (ShellHistory _uid shcmd) = toRow (Only shcmd)
  

• Create a table called ShellHistory

  -- Need Query type 
  -- Query{ fromQuery :: TS.Text }
  let sql_create_table = Query{ fromQuery = toSText "CREATE TABLE IF NOT EXISTS ShellHistory (id INTEGER PRIMARY KEY AUTOINCREMENT, shcmd TEXT)" }
  let sql_select = Query { fromQuery = toSText "SELECT id, shcmd FROM ShellHistory" }
  

• Create Connection, and Connect to Sqlite3 with open

  let dbfile = "/Users/cat/myfile/bitbucket/testfile/ShellHistory.db"
  conn <- open dbfile
  

• Execute query, Create table, Insert data to table, Select query from table

  execute_ conn sql_create_table
  -- insert data to table
  let mystr = "mycmd1"::String
  let query = Query { fromQuery = toSText "INSERT INTO ShellHistory (shcmd) VALUES (?)" }
  execute conn query (ShellHistory 0 (toSText mystr))
  cmdsql <- query_ conn sql_select ::IO [ShellHistory]
  let cmdList = let ls = map (shcmd) cmdsql::[TS.Text] in map toStr ls::[String]
  

• Again, use Data.RawString.QQ to simplify the String, ByteString and Text manipulation.
• Add response page if insertion is successful using Data.RawString.QQ again
• Add list task page

• Add html login.html file, it is easy.

• Add loginCheck::Connection -> Application

  • The function tries to validate email address and password.
    • Use queryNamed queryNamed , there is some issues to query any row from user table.

  queryNamed :: FromRow r => Connection -> Query -> [NamedParam] -> IO [r]
  r <- queryNamed c "SELECT * FROM posts WHERE id=:id AND date>=:date" [":id" := postId, ":date" := afterDate]
  

  • Finally I found the issue when using where cause in Sqlite Haskell.
  • I used ByteString to compare String or Text

  r <- queryNamed conn "SELECT * FROM user where password = :password" [":password" := password] :: [User]
  

  • password is a type ToField class which converts almost any type to SQLData with toField
  • I used ByteString instead of String or Text
  • Here is stupid mistaken that I made.

  -- Using ByteString "123" to compare Text
  -- Error code
  let password = case lookup "password" params of
                     Just password -> password
                     _             -> "No password from client."
  
          -- need to convert ByteString to Text and compare
  -- Valid code
  let password = case lookup "password" params of
                     Just password -> strictByteStringToStrictText password
                     _             -> "No password from client."
  

• Word8 is unsigned integer 8-bits byte. Word8

• ByteString contains 8-bit bytes.

  -- ls contains all character from 0 to 2^8 - 1 = 255
  let ls = Data.Word8.pack [0..(2^8 - 1)]
  

void remove(){
Node* curr = new Node(10);
Node* pt = curr;
if(curr != null)
   delete curr;
}

• pt is a dangling pointer, there is nothing wrong with dangling pointer as long as you don't deference a dangling pointer.

• Smart Pointer Example
• Smart pointer is trying to help programmers to simplify memory management. Before smart pointer, programmer needs to know when to allocate memory and when to deallocate memory.

• What is unique pointer?

  • unique pointer does not support copy, if you copy a unique pointer, then you will get compiler error.

    // no memory leak
    void fun(){
  std::unique_pt<int> pt = std::unique_pt(new int(10));
  return;
    }
  
    std::unique_pt<int> pt = std::unique_pt(new int(9));
    std::unique_pt<int> pt1 = pt;
    // pt is not longer to deference an object.       
  
  

• How to use shared pointer?

  class MyClass{
    public:
    std::shared_pt<int> pt;
        public:
    std::shared_pt<int> create(){
      std::shared_pt<int> pt = std::shared_pt(new int(9));
      return pt;
    }
  }
  
     MyClass* cpt = new MyClass();
     std::shared_pt<int> pt = cpt -> create();
  
     if(cpt != null)
   delete cpt;
  
     // std::shared_pt<int> pt don't need to be deleted inside the MyClass
  
  

• What is weak pointer in C++
  • Weak pointer is a shared pointer and it holds a non-owning reference to an object.

• Change symbol link to relative path, solve only half the problem.
• DONE Add config file.
• DONE Change localhost to real hostname.
• DONE Install Redis on FreeBSD

• download port collections portsnap fetch and extract it to /usr/ports
• Here is the link on how to use use port in FreeBSD port

• Good tutorial in Applicative Functor Applicative
• Applicative definition is very confusing initially. But if you understand Functor, then it is very easy to grasp the idea. because they are very similar.

• Let's see the Functor definition in Haskell

  class Functor f where
            return a -> f a
            fmap :: (a -> b) -> f a -> f b
  

  • It says fmap need a function and Functor type f a and return a Functor type f b (f is Functor) here.
  • Let see the Applicative definition in Haskell Appliactive

class (Functor f) => Applicative f where
                pure a -> f a
                <*> :: f (a -> b) -> f a -> f b

• Applicative satisfied a few laws:

  -- Associative
  -- Compositon
  -- Identity
  -- Interchange
  -- Homomorphic
          -- F (f . g) = F(f) . F(g)
  

• Above definition says that <*> can apply a function inside the Functor to a value inside a Functor It is very similar to fmap which is apply a plain function inside the value of Functor

• app::Connection -> IORef M.HashMap -> Application
• Remove the extrac database connection from above function
• One database includes all tables.

• If you use thomasf.github.io code snippet for Org-mode, then you will have problem to display the html file.
• I have no idea what the javascript does. After I remove the Javascript code, it works perfectly.
  • It means you do not need the Javascript code on your Org-mode.
  • At least it works in SpaceEmacs 8.x under Chrome, Safari and Firefox.

-- light mode
#+HTML_HEAD: <link rel="stylesheet" type="text/css" href="http://thomasf.github.io/solarized-css/solarized-light.min.css" />
-- dark mode
#+HTML_HEAD: <link rel="stylesheet" type="text/css" href="http://thomasf.github.io/solarized-css/solarized-dark.min.css" />

• You DO NOT need the following snippet on your Org-mode, it will screw up your Html after you export Org-mode to Html.

#+INFOJS_OPT: view:t toc:t ltoc:t mouse:underline buttons:0 path:http://thomasf.github.io/solarized-css/org-info.min.js

• copy binary file to mybin dir
• create symbo link to $sym dir
• create simple shell script under $scr

• Apparently freeBSD default shell is not bash but sh and you need to change it if you did not want all the annoyance.
• Also you need to make sure your shell is in /etc/shells

• See the difference between bash and sh in bash and sh

  chsh -s /usr/local/bin/bash
  

• GHC build package can take lots of memory sometimes.
• Run stack build and out of memory error in Ubuntu 1G RAM

• But it works on FreeBSD 3G RAM.

  Cabal                > ghc: internal error: Unable to commit 1048576 bytes of memory
  Cabal                >     (GHC version 8.6.5 for x86_64_unknown_linux)
  Cabal                > Building library for Cabal-2.4.1.0..
  

• Interactive shell is the first process Id when you login to an interactive session.
  • Read /etc/profile to setup environment variables, also read .bashrc, .profile on your home directory if there exist.

• Non-interactive shell does not read /etc/profile, e.g.

  ssh  xxx@xfido.com 'ls'
  

  • ssh logins to cat user account but the process does not activate any shell such as bash and there is no environment variable.

  ssh xxx@xfido.com  'bash /etc/profile; echo $PATH'
  ssh xxx@xfido.com 'source /etc/profile; echo $PATH'
  

• shell-command-to-string is Non-interactive in Elisp

  (shell-command-to-string "fileExist /tmp/f.x")  ;; command: fileExist can not be found error
  ;; fileExist is in $scr/AronLib.sh
  ;; we can source it before use the shell
  (shell-command-to-string "source $scr/AronLib.sh; fileTimeFull /tmp/f.x") ;; 
  

• ssh login to cat user account and call bash to read /etc/profile. Environment variables can be used if there are any inside /etc/profile.
• login shell is the same as interactive shell.
• non-login shell is the same as non-interactive shell.

• If you put #!/bin/bash -i on top of your shell script, all your shell script commands will be in your history. Enable interactive shell on your shell.

  • history \(\Rightarrow\) does work because bash is interactive

  #!/bin/bash -i
  history 
  

• -i \(\Rightarrow\) interactive shell

• non-interactive shell \(\Rightarrow\) history command does not work.

  #!/bin/bash
  history 
  

• Also, -i \(\Rightarrow\) enable your shell script can use history command, otherwise history CAN NOT be used in your shell script.

  -c  clear all histories
  -d      offset, history -d 20 => delete line 20
  -a  append current history to history file
  -n  read all history lines not already read from the history file
        and append them to the history list
  -r  read the history file and append the contents to the history list
  -w  write currenty history to history file
  
  -p  !100:p  => expand it without storing it in history list
  
  -s  append the ARGs to the history list as a single entry, no idea what does it mean?
  

• Both version are 2.9a
• tmux config file for FreeBSD and MacOS

• /dir/lib contains jar file to support java source file under /javafile
• -classpath \(\Rightarrow\) specified the location of jar files
• -noqualifier \(\Rightarrow\) no long qualifier name such as org.apache.commons.io.StringUtils
• -sourcepath \(\Rightarrow\) specified the location of java files.

javadoc -classpath  /dir/lib -noqualifier all -d /htmldoc   -sourcepath /javafile/*.java

• Shell script generates java doc in $b/javalib My gene_javadoc.sh

• Atomicity
• Consistency
• Isolation
• Durability

• Make sure you clean your cache on Chrome.
• Http Header W3 Org

• Wai Network.HTTP.Type.Header, Content-Type, Content-Disposition

  [("Content-Type", "application/pdf"), ("Content-Disposition", "inline;filename=file.pdf")]
  

• Use pathInfo instead of rawPathInfo

• Use Haskell powerful pattern matching.

  case pathInfo request of
          ("pdf":fn:_)  -> routing to whatever you want
  

• If the request string is /dir/file.pdf

• pathInfo can be used to (split ⇒ splitBS) BS.ByteString info parts and pattern matching can be used in different parts.

   pathInfo::BS.ByteString -> [Text]
  
   case pathInfo request of
  ("pdf":_)    -> routing to "http://xfido.com/pdf"
  ("pdf":fn:_) -> routing to "http://xfido.com/pdf/myfile.pdf"
  _            -> routing to "other universe"
  

• Why implication and contrapositive table defines the way it is?

• Example: If \(a^2 + b^2 = c^2 \Rightarrow a, b, c\) forms a right triangle. Pythagorean Theorem
  • If \(a, b, c\) are not a right triangle \(\Rightarrow a^2 + b^2 \neq c^2\)
  • \(n \in \mathbf{N} \rightarrow 2n \in \mathbf{N}\) \(\Rightarrow\) if \(2n\) is not an integer then \(n\) is not \(\mathbf{N}\)
  • \((P \rightarrow Q) \Rightarrow (Q^\lnot \rightarrow P^\lnot)\)

• You are likely missing the plugin in Intellij if you use Lombok to generate setter and getter.
• You will spend more time on debugging the Lombok than writing code.
• Setter, Getter, Equal, toString and hashCode methods can be generated from IDE.
• In practice, you don't use those methods that often.

• Pass lambda function to a function in C++.
• Use std::function to define a function called even

• Create a filter function similar to Haskell filter

  std::function<int(bool)> even = [](auto x) = { return  x % 2 == 0;};
  
  template<class T>
  vector<T> filter(std::function<bool(T)> f, vector<T> vec){
     vector<T> retV;
     for(auto t : vec){
        if(f(t))
      retV.push_back(t);
     }
  }
  
      vector<int> vec = {1, 2, 3};
      vector<int> vecInt = filter(even, vec);
  
  

• But you can not do following:

      vector<int> vec = {1, 2, 3};
  auto f = [](auto x){ return x % 2 == 0;};
  vector<int> filter(f, vec);
  

  • \( \because \) auto f is not the same type as std::function

• Use template<typename Fun, typename T> to solve above limitation.

  template<typename Fun, typename T>
  vector<T> filter(Fun f, vector<T>& vec){
     vector<T> retv;
     for(auto e : vec){
        if(f(e))
      retv.push_back(e); 
     }
     return retv;
  }
  

• We can pass auto f=... or std::function<bool(T)> f = .. to the new filter function.

  auto f = [](auto x) { return x % 2 == 0;};
      // or 
      std::function<bool(T)> f = [](auto x){ return x % 2 == 0;};
  
      vecotor<int> vec = {1, 2, 3};
      vector<int> filter(f, vec);
  

  • Some comments suggest the following code Lambd Function in C++

auto f = [](auto x) { return !(x & 1);};
// nice, but it might introduce bug in the future.
    // e.g. if I need module of 3, 5, ..., the whole function needs to be rewritten.

• In Haskell, the solution is much more elegrant.

          let list = [1, 2, 3];
  filter(\x -> x `mod` 2 == 0) list
  

• std::transform is similar to map in Haskell.

• But std::transform is too verbal, e.g.

  vector<float> vec = {1.2, 4.9};
  std::transform(vec.begin(), vec.end(), vec.begin(), [](auto x) { return x + 1;});
  // vec = {2.2, 5.9}
  

• In Haskell

          let vec = [1.2, 4.9]
  let v = map (+1) vec
  

• If we want to convert all the floats to strings.

  vector<float> vec = {1.2, 4.9};
      vector<string> vecStr;
      std::transform(vec.begin(), vec.end(), std::back_inserter(vecStr), 
                [](auto x) { 
                     return std::to_string(x);
                }
               );
    // vecstr contains strings
  

• I still have no idea what is std::back_inserter() from the doc.std::back_inserter
  • I assume std::back_inserter() all the return value from lambda function from the beginning of iterator

• I try to implement mapX function like map in Haskell.

  template<typename Fun, typename T, typename D>
  vector<D> mapX(Fun f, vector<T> vec){
      std::vector<D> vecD;
      std::transform(vec.begin(), vec.end(), std::back_inserter(vecD), f);
      return vecD;
  }
  

  • But it seems to me typename D can not deducted from \(\lambda\) function Fun \(f\)
  • Some users suggests me to use some C++17 features C++17 type_traits

• Wrong way to use reference in C++

// return by reference.                
vector<int>& fun(){                    
     vector<int> vec = {1, 2, 3};      
     return vec;                       
}                

vector<int> v = fun();                      

• Why the code above is wrong?
  • fun() only returns the address of vec, not a copy of vec
  • After the function fun() return, object vec will be deleted because vec is local variable and it is inside the stack. The caller will lose the data in vec.
  • v hold the address to vec, but the data inside vec will be deleted after fun() return.
  • copy constructor will be called to do the copy for vector<int>

  // return by value.
  vector<int> fun(){
vector<int> vec = {1, 2, 3};
return vec;
  }

  vector<int> v = fun();

• Code above is OK beacuse fun() return a copy of vec
• v will have the copy of vec

• What is the different between return const int& fun() and int& fun() in a function.
• const int& fun() \( \Rightarrow \) it means you can assign value to fun()
• int& fun() \( \Rightarrow \) it means you can NOT assign value to fun()

int& fun(){
  int var = 1;
  return var;
}
// in main
fun() = 3;  // OK, does not do anything useful.

const int& fun(){ 
   int var = 1; return var;
}
// in main
fun() = 3;  // Error: can not assign value to const int& which is rvalue I think.    

• return by reference is useful if you overload [] operator on an array of objects.

    class Object{
  int num;
    };
  
    class ArrayObject{
  
       Object& operator[](const int& index);
    }
  
    // in main
    Object o();
    arrayObj[1] = o;
  
  

• Passed by value:

  vector<int> even(vector<int> vec){
    vector<int> ret;
    for(auto e : vec){
           ret.push_back(e);
    }
        return ret;
  }
  

• vec will be copied inside the stack, if vec contains lots of data, then it will be not very efficient.

• Passed by reference, vec passes it memory address to function even()

  vector<int> even(vector<int>& vec){
     vector<int> ret;
     for(auto e : vec){
           ret.push_back(e);
     }
         return ret;
  }
      // inside your main
      vector<int> vec = {1, 2, 3};
      even(vec) // no problem.
  

• Code above is better because vec will not be copied inside the function even(). even() will use the address of vec to access the data in vec.

• Passed non-constant reference to a function.

  • vector<int>& is non-constant reference \( \Rightarrow \) I hope this is the right name.

  vector<int> even(vector<int>& vec){
    vector<int> ret;
    for(auto e : vec){
       ret.push_back(e);
    }
    return ret;
  }
  
  vector<int> fun(){
     vector<int> v = {1, 2, 3};
         return v;
  }
  
      // inside your main
  even(fun)  // compile error
  

• Apparently, fun return a const vector<int>& but not vector<int>&

  • In order to make above code to work, the signature of even(..) has to be changed as following:

  vector<int> even(const vector<int>& vec)
  

  • Signature above is not a perfect solution and following code will not work.
• vec can be modified.

vector<int> vec = {1, 2, 3};
       void fun(const vector<int>& vec){
            vec.push_back(1);
       }

• The idea is to find the square root of a float point number is equivalent to solve a equation \(n = x^2\) where \(n\) is given, find \(x\)
• Choose a point \( x = x_0 \) from \( y = x^2 - n \) where \( y = 0 \)
• We know the derivative of \( y = x^2 - n \) is \(2x\), then we can find the slop \(2x_0\) of a line which passes a point \( (x_0, (x_0)^2 - n)\)
• The equation of the line is
• \( \frac{y - (x_0^2 - n)}{ x - x_0} = 2x_0 \Rightarrow \frac{y - (x_0^2 - n)}{2x_0} = x - x_0 \)
• where \(y = 0 \Rightarrow x = x_0 + \frac{n - x_0^2}{2x_0} \Rightarrow x_1 = x_0 + \frac{n - x_0^2}{2x_0} \)

• C++ code: Square Root of Integer

  /***
   *
   * Newton' method to find the square root of any positive float point number. 
   */
   double sqrt(double n){
     double x0 = n + 10;
     double x = x0;
     while(fabs(x*x - n) > 0.00000001){
      x = x + (n - x*x)/(2*x);
     }
     return x;
   }
  

• Same idea can be used from above.
• let \(x_0\) is the initial point and we try to solve \( c = x^n \Rightarrow f(x) = x^n - c \) where \( f(x) = 0 \)
• The derivative of \( f(x) \) is \( nx^{n-1} \)
• A line passes \( (x_0, f(x_0) \) and its slop is \( nx^{n-1} \)
• The line equation is \( f'(x_0) = n(x_0)^{n-1} = \frac{y - f(x_0)}{x - x_0} \)
• \( \Rightarrow x = x_0 - \frac{f(x_0)}{n(x_0)^{n-1}} \) where \( y = 0 \) because we need to solve \( 0 = x^n - c \) for \(x\)
• Search nthRoot in C++

The project in $HOME/try/hs can be built, and one in FreeBSD can be built too. The old the version should not be used any more because many cods has been changed.

• Example:

data Any = Any {getAny::Bool}
let f = Any
:i f
: f::Bool -> Any
let rf = getAny
:i rf 
rf::Any -> Bool

• Get the identity function

let g = f . rf
g::Bool -> Bool

-- other identity function
let h = Any . getAny
:i
h::Any -> Any

• Our example tries to check whether a table mytable is in a database.

• From SO, you can use following select statement from sqlite_master table to find out.

  -- SELECT name FROM sqlite_master WHERE type='table' AND name='{table_name}';
  -- {} should not be there.      
  SELECT name FROM sqlite_master WHERE type='table' AND name='table_name';  
  

• The the schema of sqlite_master is as following

  CREATE TABLE sqlite_master(
  type text,
  name text,
  tbl_name text,
      rootpage integer,
      sql text
      );
  

• From Sqlite documentation Doc
• A record needs to be created.

• Record SqliteMaster needs to implement two instance FromRow, ToRow

    data SqliteMaster = SqliteMaster {
            x_type::Text,
            x_name::Text,
            x_tbl_name::Text,
            x_rootpage::Integer,
            x_sql::Text
          } deriving(Eq, Read, Show)
    instance FromRow SqliteMaster where
  fromRow = SqliteMaster <$> field <*> field <*> field <*> field <*> field
  
    instance ToRow SqliteMaster where
  toRow (SqliteMaster x_type x_name x_tbl_name x_rootpage x_sql) = toRow (x_type, x_name, x_tbl_name, x_rootpage, x_sql)
  

• Execute a query and count the row

  let query = Query([r| SELECT * FROM sqlite_master WHERE type='table' AND name='mytable' |])  
  listTab <- query_ conn query :: IO [SqliteMaster]
  if length listTab > 0 then print "mytable is found" else "mytable is not found"
  

• Database connection in Haskell Connection code

• C++ build script cpp_compile.sh
• Add include or multiple includes paths.

• \( \mathbf{-I} \Rightarrow \) specifify a <include> directory

  g++ -I/myinclude -o myfile myfile.cpp
  # multiple includes
  g++ -I/include1 -I/include2 -o myfile myfile.cpp
  
      # Use environment variable $clib
  echo $clib # /Users/cat/myfile/bitbucket/cpplib
  g++ -I/$clib -I/include1 -I/include2 -o myfile myfile.cpp
  

• \( \mathbf{-l} \Rightarrow \) link to shared library or shared object file such as .dll in Window and .so in Linux .dylib in MacOS

  g++  -I$clib -std=c++14 -O3 -Wall -pedantic recurveFileBoost.cpp -lboost_filesystem -lboost_system
  

• gcc link to object file with -lm Ref Code Example

  gcc linkedObject.c -o linkedObject -lm
  

• \( \mathbf{-L} \Rightarrow \) specificy a <lib> directory
• \( \mathbf{-g} \Rightarrow \) turn on debug(so GDB gives more friendly output)
• \( \mathbf{-c} \Rightarrow \) output an object file (.o)
• \( \mathbf{-o} \Rightarrow \) name the output, executable output directory
• \( \mathbf{-Wall} \Rightarrow \) turn on most warning
• -pendantic or -Wpendantic \( \Rightarrow \) Issues all warning required by ISO C and ISO C++ standard, it issues all warning whenever there are compiler extensions non compilant to ISO C or C++ standard.
• When you use new C++ library, you need to add Header search paths and Library search paths in XCode.
• I try to use C++ boost library and add Header search paths and Library search paths in XCode.
  • I still get linking error, symbol not found
  • When you get this kind of error, generally you try to Google the error and hopefully you will find the solution.
  • It is hard to know what the Xcode is doing behind the sense when you get weird error.

• I use home brew to install C++ boost library on my MacOS, the head location and libraries location are:

  # header location, *.h 
  /usr/local/Cellar/boost/1.61.0_1/include
  # libraries location *.dylib and *.a files
  /usr/local/Cellar/boost/1.61.0_1/lib
  
  # 27-12-2020
  /usr/local/Cellar/boost/1.72.0_2/include
      /usr/local/Cellar/boost/1.72.0_2/lib
  

• The recurveFileBoost.cpp can be built with following g++ command line

  # file:///Users/aaa/myfile/bitbucket/cpp/recurveFileBoost.cpp
  cd $b/cpp
  g++ -std=c++14 -O3 -Wall -pedantic recurveFileBoost.cpp -lboost_filesystem -lboost_system
  

• XCode add include header, library and linker in setting. xcode boost

  

• Add shared libraries to XCode. xcode boost

  • /Users/cat/myfile/bitbucket/cpplibtest/CppLibTest Running Example.
  • boost_system, boost_system are added to XCode Build Setting.

  

• Find Xcode tool chain in MacOS and g++ libraries paths

  g++ -print-search-dirs
  

• gcc compile c file, gcc --help > $t && vim $t with filter g!/output/d

  gcc hello.c -o /tmp/hello
  

• Change cpp compile in Xcode

  

• Compile cpp to dynamic libraries MacOS

  • Ref

  g++ -dynamiclib -o mydylib.dylib mylib.cpp
  g++ client.cpp -L/dir/  -lmydylib
  

• Install boost on MacOS or Linux
• Download boost boost₁₇₅₀ (current boost version) boost₁₇₅₀
• Unzip or untar it
• Nothing to build, most of Boost libraries are header-only

• Goto boost root directory which is the following

  $b/cpplib/boost_1_75_0/boost
      #include <boost/filesystem.hpp>
  

• If you want to use any of separated-compiled Boost libraries, you'll need to acquire library binaries.

  cd $b/cpplib/boost_1_75_0
  ./bootstrip.sh --help
      ./bootstrip.sh  # => generate $b/cpplib/boost_1_75_0/stage/include, $b/cpplib/boost_1_75_0/stage/lib
  
      ./b2            # build
      ./b2 install    # install
  

  • include header and lib dynamic libraries are generated
  • How to use include and lib in your CPP code?
  • Here is g++ for line project
  • It Use dynamic libraries: boost_system and boost_filesystem
  • \( \mathbf{-I} \Rightarrow \) specificy a include directory
  • \( \mathbf{-L} \Rightarrow \) specificy a lib directory
  • \( \mathbf{-l} \Rightarrow \) specificy dynamic libraries

  $boostinclude/stage/include
  $boostinclude/stage/lib
  outdir="$ff/runbin/line"
  run="g++ -I$clib  -I$boostinclude -std=c++14 -O3 -Wall -pedantic line.cpp  -L$boostinclude/stage/lib  -lboost_system -lboost_filesystem -o $outdir"
  

• How to use boost library in your CPP code

• file:///Users/cat/myfile/bitbucket/cmake/CodeAnalysis/CMakeLists.txt
• Working example at /Users/cat/myfile/bitbucket/cmake/CodeAnalysis
• Add include header files $b/cpplib/AronLib.h to CMakeList.txt

• Add shared libraries to CMakeList.txt

  # header location, *.h                     
  /usr/local/Cellar/boost/1.61.0_1/include         
  # shared libraries location *.dylib and *.a files 
  /usr/local/Cellar/boost/1.61.0_1/lib       
  

• Best CMake Tutorial

• Edit the http.config to load mod_alias
• It seems it does not work so far.

• Recurve from a given directory and use depth first search to traversal.
  • If file is found, then use file.replace(' ', '') to remove whitespace.
    • If directory is found, then recurve into the directory.
    • If return from a directory, then dir.replace(' ', '')

• You can comment out block of code in Python?
• Do not understand the rational behind that.

• Python os.stat, Python standard library 3.9.a02
• Lib/os.py

import os
statinfo = os.stat('/tmp')
statinfo.st_mtime

os.stat_result(st_mode=33188, st_ino=7876932, st_dev=234881026,
           st_nlink=1, st_uid=501, st_gid=501, st_size=264, st_atime=1297230295,
           st_mtime=1297230027, st_ctime=1297230027)

• Convert byte string to utf-8 string

• The module itertools contains takewhile, dropwhile and other functions. Those functions are very common \( \text{red}{operation} \) in programming languags.
  • \(\text{red}{takewhile}\) and \(\text{yellow}{dropwhile}\)

import itertools as it
        ls = [0, 0, 1, 0, 1]
        print(list(it.takewhile(lambda x : x == 0, ls))) # => [0, 0]
        print(list(it.dropwhile(lambda x : x == 0, ls))) # => [1, 0, 1]

• List Comprehension

  s1 = [1, 2]
  s2 = [3, 4]
  # Cartesian Product of s1 and s2
  ls = [(a, b) for a in s1 for b in s2]
  print(ls)
  

• Why do we need the list() to wrap the map and filter functions.
• Python return iterator instead of a list when use map, filter etc.

list(map(lambda x: x > 1, [1, 2, 3]))
list(filter(lambda x: x % 2 == 0, [1, 2, 3]))

• reduce is a bit better, or fold in Haskell terminology.

  from functools import reduce
  reduce(lambda acc, x: acc + x, [1, 2, 3], 0) => 7
  

• map can be used like zipWith in Haskell

  list(map(lambda x, y: x + y, [1, 2, 3], [3, 4, 5]))
  

• List is awesome. Python list, drop list, take list, reverse list

  ls = [1, 2, 3]
  ls[:1] => [1]
  ls[1:] => [2, 3]
  ls[:1] + ls[1:] => [1, 2, 3]
  ls[:-1]  => [1, 2]
  ls[-1:]  => [3]
  ls[::-1] => reverse [3, 2, 1]
  

• Install Virtualenv

pip3 install --user pipenv

• Initialize project

cd my_project
pipenv install requests

• pipenv is in following dir

/Users/cat/Library/Python/3.8/bin/pipenv

• Create simple project, main.py

import requests
response = requests.get('https://httpbin.org/ip')
print('Your IP is {0}'.format(response.json()['origin']))

• Run the script using pipenv run:

pipenv run python main.py

• Expect similar output

Your IP is 8.8.8.8

• let \( c \in \mathbb{C} \)

• \( x + y i \) is rectangular form.
• \( r(\cos \beta + i \sin \beta) \) is the polar form

• Create an haskelldoc_index.html file under $b/public1file/haskelldoc_index.html
• Symbol link the file to $ht/../htmlhaskelldoc/index

• Keybinding is supper easy. VSCode configure file

  

  • If you want to add ctrl+k to move your cursor up, then you need to edit your keybinding.json file as following

    {
    	"key": "ctrl+k",
    	"command": "cursorUp",
    	"when": "textInputFocus"
    }
    

  • Edit the keybinding.json file.
    • Path to keybinding.json

  ~/Library/Application\ Support/Code/User/keybindings.json
  

• Java autocomplete is working for SDK and external jar files.
  • I have been the vim user for many years but I never use Java auto complete because it is painful to install those plugins.

  • VSCode supports autocomplete out of the box so far but I have not tried other libraries. but Java SDE API is working so far. All the following can be autocompleted.

    Map<String, String> map = new HashMap<String, String>();
    Set<String> set = new HashSet<String>();
    

• Autocomplete can be used in many languages such as C++, Java and Python.
• It is very easy to configure keybinding.
• You can install many packages like Emacs. searching and click install.
• There are lots of nice themes.
• VSCode is not perfect so far.
  • I try to configure VSCode to search inlcude CPP header files.
    • First, I try add the includePath in the settings.json, but it failed to find my namespace under AronLib.h
    • Second, I try to add c_cpp_properties.json under .vscode directory and it failed to find header file AronLib.h

• Initially I though the .classfile is in the VS Code workspace directory and VSCode can not detect my own jar file.
• Finally I realize .classpath should be in home directory. I did not see anyone mention that on stackoverflow at all.
• The content of .classfile can be similar the following:
• You need to have xxx.code-workspace file in order to build and compile.
• Stupid thing about VSCode is the directory for jar files has to be called lib under your workspace. No clue.. why.

• You might be able to change the name somewhere. Ask MS.

  <?xml version="1.0" encoding="UTF-8"?>
  <classpath>
    <classpathentry kind="lib" path="/Users/cat/myfile/bitbucket/javalib/jar"/>
    <classpathentry kind="lib" path="lib/commons-io-2.5.jar"/>
  </classpath>
  

• Actually it is not ture for above .classfile. Apparently it does not need .classfile if there is lib/myjar.jar in the workspace directory.
• I can do autocomplete for my own jar file, but I got compiler error. The error indicates I use older JDK or Runtime environment with newer version class file.
• TODO: No clue how to fix it right now.
• The weird thing is I can use my current java compiler java_compile.hs to compile and run the same source java file.
• It means VSCode uses an older version of java compile which might be java 8 compiler.

• Find out all about your java stuff on your machine with following command

  java -XshowSettings:properties -version
  

• Output executable file, ex: Cpp option -o

  • g++ -o HelloWorld HellWorld.cpp

  "${fileBasenameNoExtension}",
  

• Cpp Compiler

  #MacOS 
  "command" : "clang++"
  # or
  "command" : "g++"
  

• Include library

  -I /usr/local/Cellar/boost/1.61.0_1/include
  

• Two Liking libraries: -lboost_filesystem and -lboost_system

  • Here is tasks.json file for MacOS

    {
    	"version": "2.0.0",
    	"tasks": [
    		{
    			"label": "Build with Clang",
    			"type": "shell",
    			"command": "clang++",
    			"args": [
    				"-std=c++14",
    				"-stdlib=libc++",
    				"-I /usr/local/Cellar/boost/1.61.0_1/include",
    				"-lboost_filesystem",
    				"-lboost_system",
    				"${file}",  
    				"-o",
    				"${fileBasenameNoExtension}",
    				"--debug"
    			],
    			"group": {
    				"kind": "build",
    				"isDefault": true
    			},
    			"problemMatcher": [
    				"$gcc"
    			]
    		}
    	]
      }          
    

• What is on the tasks.json
• What is on the c_cpp_properties.json
• It is really painful to setup VSCode to compile Cpp code.
• From SO, you need to setup c_cpp_properties.json and tasks.json
• I follow some tutorial in SO but itt is still not working. Unfortunately it is just a "Hello world" with including a library.
• I got following error Error
  • Actually when run it in "Run Task", VSCode will run whatever on your current window. VSCode does not know what file they should run.??
• Add linking library to VSCode, adding to tasks.json file with trial and error.
• It works at least.
• Working example under $b/cpp/vscode/tryreadfile

• The main objective is to autocomplete Haskell code.
  • It does not work for me when I install Haskero inside VSCode, it seems to me I have to install from command line.
  • Install Haskero Haskero 1.31

stack build intero  --copy-compiler-tool 

• Install intero right now. intero
• Find your package Haskero inside your VSCode package Window and Enable Haskero again.
• You need to add your stack project to as workspace in VSCode, this is the most important step.
• It can autocomplete source code so far.
• You can click on stack build, stack run and stack test

• Intero is the best package out there for Haskell. Intero
• autocomplete, defintion and type checking
• Intero can autocomplete paths too:)
• Add a few lines to dotemacs file.

(exec-path-from-shell-initialize)
(package-install 'intero)
(add-hook 'haskell-mode-hook 'intero-mode)

• Open your stack project and you are ready to go.

• Find an Integer solution for \( x^3 + y^3 + z^3 = 42 \)
• It seems to be an easy problem but no one can find a solution or prove there is no solution for the eqation.

x = -80538738812075974
y = 80435758145817515
z = 12602123297335631
if x**3 + y**3 + z**3 == 42:
   print("x^3 + y^3 + x^3 = 42")
else
   print("x^3 + y^3 + x^3 != 42")

\( f(n) = \begin{cases} n/2 &\quad \mbox{ if n is even} \\ 3n + 1 &\quad \mbox{ if n is odd} \end{cases} \)

• Terence Tao posted a partial solution on Collatz Conjecture

• If you want to relod the new CSS file from Chrome. Clear up browser data is not enough.
• I have no clue to reload the new CSS files from Chrome when you try to inspect your html source.