Our MySQL env is a pxc/galera cluster, and connections require tls. When I asked the developer about adding tls support (about 4 lines of code if you’re using mysqli as he is) I just got a flat “nope, not doing that”.
At which point I lost all reservations about deobfuscating the “protected” files and simply replacing the db setup call in every single god damn file with a single setup that accepts tls options.
I’ve worked on some garbage code projects, but sendy is almost deliberately bad. In one place it: disabled all error reporting; tries to connect to the database; fails silently if the connection doesn’t succeed.
There is zero logging (besides an ungodly number of notices and warnings about undeclared variables, etc) and the author has apparently zero interest in actually fixing or improving any part of it at a technical level.
For those who work in php and want a comparison: it makes Wordpress in general look like a well architected application.
Still, the big selling point is: Inexpensive. Easy to install (simple PHP). Looks relatively polished.
All the "competitors" are either much more expensive, or you need to be an enthusiast in the tech stack they have chosen.
We've been using it for years without any issues.
A lot of popular (but expensive) services like ConvertKit have moved to a tagging approach many many years ago.
It's so much easier to manage your list of emails when it's a single list and then you add tags to specific users like "purchased X". This way the email address only exists in 1 spot and you can segment on the tags.
The Sendy approach (and from the looks of it your app too) becomes very unwieldy with having to manage, parse and merge multiple lists on a regular basis.
Performance is extra helpful for making real-time segment size calculations. We use elastic search and it's costly.
The multi-list approach has several other benefits. When manager / sender (and other) permissions get introduced, it will be straight forward to restrict users to managing certain lists. In addition, multiple lists allow subscribers to selectively subscribe / unsubscribe from lists.
Internally, the structure is simple. There is only one subscribers table and subscriber data is not duplicated anywhere. List (foreign key) relationships are in a separate table.
So in theory could you have 1 list and each subscriber has many tags, and then you can segment on those tags?
Also, do you have any public success stories beyond your own Zerodha campaigns? Have you compared delivery / bounce / etc. rates vs Sendy and other tools using the same email providers? Also how fast can you send emails out through SES?
Like GrapesJS: https://github.com/artf/grapesjs-mjml
Background: I write PHP for fun, I don't compete with Sendy, I don't have any stake in any service that competes with Sendy. I also don't expect PHP to be written in the most OOP-everything Symfony-esque way either. But this is a bit too much for me even.
It doesn't follow any modern standards or guidelines; it doesn't use any templating system, just PHP hardcoded inline with HTML in hundreds of files (no modern framework, lightweight or not), concatenated variables into SQL queries and HTML. The author seems to hate using { } for if/else statements, which has the potential of introducing fun bugs if not extremely careful. 1 and 2 letter variable names are common (and I don't mean in the way of $i or $j); parts of it border on insane or bizarre; there are 500 character lines where instead of an error generation feature they just die() with a full HTML document as a string for error pages; functions are written inline all over the place and they all use the keyword global; there are no parameterised queries - it's covered in repetitive mysqli_real_escape_string and query concatenation. There is a massive amount of copy-pasted code: instead of some kind of single database management layer, there's a re-definition of a function that connects to a database with the same name in dozens of files, all of which try to read global variables (doesn't take any parameters)