Everyone Has a Story

About

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Sat, 20 Jun 2015 14:02:37 +0200

My name is Bhavesh S. Gupta. I’m an Associate Tech Lead at IDX and I work out of Vadodara.

I’ve been writing code professionally for eleven years, most of it across the stack: NodeJS, React and Next.js, TypeScript, MongoDB, Redis, Python, and a good deal of AWS. These days most of my time goes to live video and real-time systems, and to the far less interesting work that keeps them standing once real people show up: load testing, version upgrades, security audits, disaster-recovery plans.

It only fails in Safari: two HLS engines, and how to debug across hops

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Tue, 06 Oct 2026 09:03:00 +0530

When a live stream breaks in Safari and plays in Chrome, the tempting conclusion is that Safari’s player is the problem. Often it isn’t. Both browsers receive the same files from the same CDN. They differ in what they do when one of those files is wrong, so a delivery fault that Chrome quietly recovers from becomes a Safari bug report.

In short: Chrome, Edge and Firefox usually play HLS through hls.js, which retries failed requests and skips small gaps; Safari uses its native player, which surfaces the same faults as errors or stalls. Both get identical bytes, so a problem that only shows in Safari is usually a delivery problem. Find it by measuring each hop separately: compare responses at the origin and the CDN, probe from many regions, group logs by browser, and inspect the segments.

Slide sync over HLS with ID3 timed metadata

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Tue, 06 Oct 2026 09:02:00 +0530

A webcast with slides has to change the slide at the moment the presenter does, as the viewer sees it. Viewers are 10 to 30 seconds behind live, each by a different amount, so a message sent “now” from the presenter’s page arrives at the wrong time for everyone. The fix is to put the slide change inside the stream itself, at the right timestamp. In HLS that means ID3 timed metadata.

In short: To change slides at the exact moment the presenter did, put the change inside the stream as ID3 timed metadata. Each tag rides in the segment on its own data stream with a presentation timestamp, so the player fires it at the right frame however far behind live the viewer is. Put every tag in every quality, re-send the current slide regularly for late joiners, and read the tags through the video element’s metadata text track, which works in hls.js and Safari.

Putting a live HLS origin behind a CDN

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Tue, 06 Oct 2026 09:01:00 +0530

A CDN in front of a live HLS origin should mean each segment leaves your server a handful of times, however many people watch. Whether it does depends on details that are easy to get wrong and hard to see: the URLs your origin hands out, the headers on each kind of file, and every layer between the origin and the viewer that has its own opinion about caching.

In short: A CDN only protects a live origin if every viewer requests the same URLs: no per-viewer session ids, and no query strings in the cache key. Give media playlists about a one-second lifetime, segments about a minute, and errors no-store so an early 404 isn’t served to everyone. Check headers at every hop, send CORS headers when the player is on another domain, and confirm request collapsing and an origin shield with the CDN that pulls from you.

HLS for backend engineers: playlists, segments and keyframes

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Tue, 06 Oct 2026 09:00:00 +0530

If you run the servers behind a live stream, you don’t need to know much about codecs. You do need to know what files a player asks for, how often, and what decides their size and timing. That’s what this post covers, with commands you can run against any HLS stream.

The series is about standard HLS, with MPEG-TS or fragmented MP4 (CMAF) segments. The same principles carry over to Low-Latency HLS, though its partial segments and blocking playlist reloads aren’t covered here. Real-time streaming, such as WebRTC at under a second of delay, works differently and isn’t covered at all: it has no playlists or segments for a CDN to cache.

In short: A live HLS stream is a master playlist of qualities, one media playlist per quality, and short segments. Each segment must start on a keyframe, so a fixed keyframe interval of 1 or 2 seconds keeps segments even. The target duration is the ceiling on segment length; players reload about once per target duration and start three target durations behind live, so it sets your latency. Every quality needs the same boundaries and timestamps for switching to be smooth.

Exploring Backend-as-a-Service (BaaS) Platforms for Faster Development!

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Wed, 30 Apr 2025 17:39:08 +0530

Backend-as-a-Service (BaaS) platforms hand you the backend pieces you would otherwise build yourself: authentication, a database, real-time data sync, hosting. You write the parts of the app that are actually yours.

While looking at options for rebuilding one of our platforms, we evaluated a few of them:

Amazon Web Services (AWS) Amplify (Amplify Backend includes hosting)
Supabase
Pocketbase (still in beta)
Convex
Firebase

If the goal is to ship an MVP or a prototype soon, any of these will get you there faster than rolling your own, and a small team or a startup can spend its time on the product instead of on infrastructure.

Resignation Etiquette

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Mon, 08 Feb 2021 20:50:32 +0530

One follow etiquette about several things in life but rarely people talk about resignation etiquette.

The Coin Weight Problem

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Wed, 03 Feb 2021 11:06:14 +0530

The problem resolves around find which of the coin is uneven in weight of all other coins. Suppose there are around 8 coin out of them one is heavy/light then others and we have one scale where we can compare two things

The Fizzbuzz Problem 2

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Tue, 02 Feb 2021 08:56:01 +0530

Last time I had stopped optimizing this the problem more because we were having a discussion where string concatination is a costly opperation. So yesterday I was talking with Nishchay how can about the previous blog and other things at that point we started on optimizing this further avoiding string contactination as well.

The Fizzbuzz Problem

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Sun, 31 Jan 2021 16:42:25 +0530

Recently I have been giving few interviews and I have faced this problem twice or thrice I guess where an interviewer asks me this problem.

First and Last post of 2020

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Thu, 31 Dec 2020 23:11:46 +0530

2020

2020, has been amazing year by far. Have experience so much this year which is unbelievable. Although I have not written anything then these two lines. But this would a memorable year in lot of sense.

Cheers, Happy New Year everyone, Happy 2021

Convert Excel to Json

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Sat, 14 Sep 2019 19:54:55 +0530

Some time in our projects we need data in json files(mostly web projects). But collecting and managing data in json files is harder(At least this is what I feel) compared to excel. I had recently ran into a similar situation where I need some data. Data is going to be provided by a non-technical person/client. And there going to be several rounds of it.

So I was looking for options what can be done here, where I found several online tools which could be used for same. But I don’t want to use that because of course privacy. I don’t want to upload an excel file to a unknown application which would process it server side give me results.

Another Attempt at Wag3

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Sun, 25 Aug 2019 16:57:46 +0530

Until now we had done as of following code. We would create couples of more functions which would help us in managing our application. As discussed earlier we would now we creating different directories and files present inside .git directory along with directory itself. More or less we would something like following directory structure after doing wag init

.git is the git directory itself, which contains:
- .git/objects/ : the object store, which we’ll introduce in the next section.
- .git/refs/ the reference store, which we’ll discuss . It contains two subdirectories, heads and tags.
- .git/HEAD, a reference to the current HEAD (more on that later!)
- .git/config, the repository’s configuration file.
- .git/description, the repository’s description file.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47


import sys
import argparse
import os

argparser = argparse.ArgumentParser(description="Another Content Tracker")
argsubparsers = argparser.add_subparsers(title="Commands", dest="command")
argsubparsers.required = True
argsp = argsubparsers.add_parser(
 "init", help="Initialize a new, empty repository.")
argsp.add_argument("path", metavar="directory", nargs="?",
 default=".", help="Where to create the repository.")


def main(argv=sys.argv[1:]):
 args = argparser.parse_args(argv)
 if args.command == "init":
 cmd_init(args)


def cmd_init(args):
 repo_create(args.path)


class GitRepository(object):
 """A git repository"""
 worktree = None
 gitdir = None

 def __init__(self, path, force=False):
 self.worktree = path
 self.gitdir = os.path.join(path, ".git")


def repo_create(path):
 """Create a new repository at path."""
 repo = GitRepository(path)
 # We check whether there exsits a path of worktree
 if os.path.exists(repo.worktree):
 # if path exsits its a directory or not
 if not os.path.isdir(repo.worktree):
 raise Exception("%s is not a directory!" % path)
 # if path exsits its empty or not
 if os.listdir(repo.worktree):
 raise Exception("%s is not empty!" % path)
 else:
 # Create directory as nothing exist with same path
 os.makedirs(repo.worktree)

Now let’s create function which would gives us path of directory after verifying that the directory exist and if it doesn’t it should gives us an error. We would also require mechanism to check whether we need to create the directory or not if it doesn’t exists.

Another Attempt at Wag 2

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Sat, 24 Aug 2019 15:32:32 +0530

Last time we had created cmd_init(args) function. We add more functionality to cmd_init(args).

This was our code till last time.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18


import sys
import argparse

argparser = argparse.ArgumentParser(description="Another Content Tracker")
argsubparsers = argparser.add_subparsers(title="Commands", dest="command")
argsubparsers.required = True
argsp = argsubparsers.add_parser("init", help="Initialize a new, empty repository.")
argsp.add_argument("path", metavar="directory", nargs="?", default=".", help="Where to create the repository.")

def main(argv=sys.argv[1:]):
 args = argparser.parse_args(argv)
 if args.command == "init":
 cmd_init(args)
 

def cmd_init(args):
 print("Hello from init")
 

We had added an argument to your sub parser called path. If the value of path is not specified it would be ‘.’ current directory or it would whatever specified. This would help us

Another Attempt to write a git

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Thu, 22 Aug 2019 20:50:56 +0530

Second Try

We would be trying to create a mini git clone from this tutorial. I aspect this to go long. In general git is awesome source code manager which is wieldy use.

So first we would be creating a lib file and executing file.

Let’s create a file called wag which would have following code.

1
2
3
4


#!/usr/bin/env python3

import libwag
libwag.main()

Then we would create another file which would be libwag.py with just a main function in it.

An attempt to Write Yourself a Git

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Mon, 19 Aug 2019 21:13:07 +0530

So I tried to follow wyag at here but I was not able to. This tutorial is great but I was not able follow it because may be I am not still not much clear with python and how it works. So after much thought I have decided that I would try to create my own blog post series where I would take learnings from wyag and try to do it in my way.

Learning python through pym 14-08-2019

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Wed, 14 Aug 2019 16:57:26 +0530

Day 4

PS: These are notes, there could be grammar error and spelling errors.

for loops

For loop over a list goes as follows:

1
2
3
4
5
6
7
8


>>> a = ['Fedora', 'is', 'powerful']
>>> for x in a:
... print(x)
...
# these ... are automatically created by python interpreter
Fedora
is
powerful

range() function

range function creates lists based on parameter passed in function``py

1
2
3
4
5
6


>>> list(range(1, 5)) # when two numbers are passed n,j List starts at n ends at j-1
 [1, 2, 3, 4]
>>> list(range(1, 15, 3)) # when three numbers are passed n,j,k. List starts at n ends at j after stepping at every k steps
[1, 4, 7, 10, 13]
>>> list(range(10)) # when only single number n is passed list starts from 0 and ends after n elements
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

Continue statement

Just like break is use to break loop, continue is used to continue loop towards iteration without further execution A combined example.

Learning python through pym 13-08-2019

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Tue, 13 Aug 2019 20:57:26 +0530

Day 3

PS: These are notes, there could be grammar error and spelling errors.

Last time we were just starting with If else, the control flow,

if-else , the control flow

Control flow is used when we need to do some decisive task

if, if else, if elif else statement

If statement is simple as it is other language

1
2


if expression:
 do this

1
2
3
4


if expression:
 do this
else:
 do this

1
2
3
4
5
6


if expression:
 do this
elif expression:
 do this
else:
 do this

Things to note:

Learning python through pym 12-08-2019

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Mon, 12 Aug 2019 20:57:26 +0530

Day 2

PS: These are notes, there could be grammar error and spelling errors.

Last time had got until here,

We don’t need specify data type with python something similar to javascript where we use var for every type of variable.(const and let are new ES6 standards, although they are not new but new)

So basically

1
2
3


a = 10
b = 1.0
c = "test"

in above a become integer, b becomes float and c becomes string.

Learning python through pym 10-08-2019

BhaveshSGupta@disroot.org (Bhavesh S. Gupta) — Sat, 10 Aug 2019 17:59:22 +0530

Day 1

Installing python

one can install python using package manager depending upon the operating system you are using. As I have a windows machine and ubuntu using linux on windows subsystem. I can install python from https://python.org . And Ubuntu I can use below command to install python.

`1`	`sudo apt-get install python3`

Running python

To run python you would call python3 which open Python interpreter. On that you can type print(“Hello World”) to run your first python code which would print Hello World in next line.