example, you can expect a config file like .blazer_config in home directory, there you can store defaults like threadcount, save download history option etc

Use Delimiter "|"

blazer -url=https://speed.hetzner.de/100MB.bin t=1
Fetching file meta..
Can't initiate download Head "https://speed.hetzner.de/100MB.bin": EOF

Lot of threads doesn't guarantee optimal concurrency, need to find out the ideal concurrency for a given situation and select that as thread count as default value

blazer -url=https://schooleverywhere-elquds.com/lib/uploadbook/179121479125786First_Children__s_Encyclopedia.pdf -t=200 -out=test.pdf -checksum=ea008f32a265d322bb3956afbd7c39d1b010b28dfc659c7b7ad3ed20bc0de339

https://datatracker.ietf.org/doc/html/rfc7233

Get "https://url/test.pdf": read tcp 192.168.0.113:56940->75.98.168.118:443: read: connection reset by peer
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x40 pc=0x646d0a]

goroutine 773 [running]:
github.com/arvpyrna/blazer/network.DownloadSegment(0xc0008eef00, 0x1af, 0xd8d0b8, 0xd9517f)
/home/arv/lab/blazer/network/downloader.go:47 +0x34a
github.com/arvpyrna/blazer/network.ConcurrentDownloader.func1(0xc0005940b0, 0xc0008eef00, 0x1af, 0xd8d0b8, 0xd9517f)
/home/arv/lab/blazer/network/downloader.go:32 +0x75
created by github.com/arvpyrna/blazer/network.ConcurrentDownloader
/home/arv/lab/blazer/network/downloader.go:30 +0x30f

• Introduce an optional config ~/.blazerc to enable more features of blazer like tracking history and path to store those history

For homebrew users like me it would be very cool if you can publish blazer as a package so that we can do

brew install blazer

and have it installed on our systems. Homebrew is majorly used for package management by mac users. It is also available for linux as well. :-)

• Apply clean code
• Add comments for exported functions
• Restructure the application to comply according to go standards
• Имена пакетов и функций неидиоматичны (пакет "data", к примеру).

-Get "https://logic.pdmi.ras.ru/~gravin/storage/Concrete_Mathematics_2e.pdf": dial tcp 83.149.197.121:443: i/o timeout

• Handle improper URL

• Get "https://mirror.yandex.ru/fedora/linux/releases/34/Workstation/aarch64/iso/Fedora-Workstation-Live-aarch64-34-1.2.iso": read tcp 192.168.0.113:38734->213.180.204.183:443: read: connection reset by peer

• Get "https://sample-videos.com/img/Sample-jpg-image-30mb.jpg": read tcp 192.168.43.62:41606->103.145.51.95:443: read: connection reset by peer

• TCP timeout incase of large file

• Try to download a large file like 10 or 20 gb and

• Get "https://releases.ubuntu.com/20.04.2.0/ubuntu-20.04.2.0-desktop-amd64.iso": EOF
  panic: runtime error: invalid memory address or nil pointer dereference
  [signal SIGSEGV: segmentation violation code=0x1 addr=0x40 pc=0x64366a]

A segment download can fail due to errors such as network backpressure, it makes sense to retry few times before failing the entire download process.

before merging the file we are not checking if all the segments were downloaded or not. And also moved the segment merging out of the network code for more clarity.

if a segment is not downloaded successfully, push it to a channel, listener of the channel will initiate the download again

• check bytes and then verify download completion

CI/CD pipeline seems to be broken, needs to be fixed

https://github.com/cheggaaa/pb/

The process doesn't clean up the temp directory created for segment storage when it receives SIGTERM / SIGINT signal.

gox

Probably, prompt the user with available space and confirm to go ahead with the download or not.

If a segment is not completely downloaded, need to retry that, create a new channel that will run those before merging files, do not merge files if there is not enough segments

Todo:

• show time elapsed
• print the performance graph as well to see perfromance difference, using different number of threads and
• show progress
• also need support for google-drive, dropbox, a simple youtube version too.
• use interactive terminal to show progress across each thread, so it more cooler
• upload it to free file sharing service upto some MB, you can build one for yourself and share
  it with others, write in go, or node js and also use it here, a simple service
  the life of the service can be just 1 hour, there you can learn so many interesting things
• download files after giving username and password(with authentication)
• export as package too, later somehow
• show current download speed and ETA (ETA depends on the number of threads, download speed, speed of server)
• maintain history of downloads, if there is a flag -h = true

Bugs:

• Handle the case where there is no internet
• while downloading file if there is interruption and if the downloaded file is incomplete, what will you do,
  you can find out incomplete file download by checking expected bytes to be written and bytes written, and then
  you can try to retry that attempt
• even if the program crashes, use exception handling and restart the download with same thread count
• Create a ruby server and try to serve a 10 gb file and try to do the same in go and other servers and
  estimate performance and also node js

Help wanted:

• unit testing
• Write a compatable, high speed server to compliment this client
• Support torrent based
• GUI maybe in electron or QTCreate