|
I once ran across a description of a Civil War-era cypher and it was very strange. It used a crossword pattern with and without dots for the first eighteen letters and then an X pattern with and without dots for the last eight. I can see how messages like that could take a while to decipher.
"They have a consciousness, they have a life, they have a soul! Damn you! Let the rabbits wear glasses! Save our brothers! Can I get an amen?"
|
|
|
|
|
|
Interesting that this was sent less than 2 weeks before the pivotal Battle of Gettysburg - one of the few times that the Confederate Army marched into the northern states. Hill was part of that army that marched north.
Edit: Mistake on my part - I am confusing Confederate General A P Hill with Confederate General D H Hill. D H Hill stayed in Richmond VA during the battle of Gettysburg. I apologize.
|
|
|
|
|
Yes; Gettysburg. There are dispatches (from both sides) about their relative movements towards that big event ... and then this pops up.
"Before entering on an understanding, I have meditated for a long time, and have foreseen what might happen. It is not genius which reveals to me suddenly, secretly, what I have to say or to do in a circumstance unexpected by other people; it is reflection, it is meditation." - Napoleon I
|
|
|
|
|
I get reminded of Sherlock Holmes adventure "The Dancing Men". A cryptography based adventure.
|
|
|
|
|
This.
This is why all manner of security gets broken. Security is hard work. Security must be maintained, no matter if it's passwords or door locks.
If Seddon were alive today, he would certainly reuse passwords.
|
|
|
|
|
I was going to make a small demo of loading a jpg from the web and consuming it on an ESP32
The image should be small (320x240 ish) because i can't resize it, it should be retrievable by an end user for verification and hopefully it would be a prominent recognizable figure on the web.
Here's what I've checked:
Code Project's logo (gif)
Google's logo (png)
Twitter (png, and huge)
Reddit (not easily retrievable)
Github (not easily retrievable)
Youtube (not easily retrievable)
Here's what I don't want to promote - even by way of a demo:
Microsoft
Facebook
I feel like maybe I have some tunnel vision. Surely there is site with a prominent logo in jpg that's easily accessible from a browser, no?
Am I being unrealistic?
Should I implement png? (I really don't want to for reasons)
To err is human. Fortune favors the monsters.
|
|
|
|
|
PNG makes a lot of sense: it's a lossless compression format which JPG isn't, and (under some circumstances and not often) smaller as well.
But importantly, PNG has a transparency layer, which JPG doesn't - so "floating" a logo over a background works better as the apparent logo doesn't have to be rectangular - it can be circular or irregular, and still not affect the stuff behind it.
That's handy - you don't need different images for each background colour for example.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
Yeah, I've considered that. It's hard to find code to load it progressively like JPG does though. I'll have to keep looking, or wind up dramatically modifying something out there. Maybe I'll try to add that now. It gives me something to do.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Nobody who uses images on websites cares about that - it's "just a image stream" and if one format has an advantage for them, they'll use it. And a transparency layer is an advantage to many site builders.
They don't care about poor old us, after all! 
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
Yeah, I think finally implementing PNG might be the way to go here.
To err is human. Fortune favors the monsters.
|
|
|
|
|
|
Because it's of limited use for IoT stuff, and I'd rather not waste the flash space on it.
To err is human. Fortune favors the monsters.
|
|
|
|
|
What space? All you would need to do is parse it into a BITMAPINFOHEADER structure and write the DIB directly to your screen. The only space required might be your BITMAPINFOHEADER header definition. 
|
|
|
|
|
And without a proper use case there's no reason for it, even if it's only a BITMAPINFOHEADER that has no business being in there without a use case.
JPGs have dozens of use cases - heck I use them in my commercial projects w/ this lib.
Arguably, so do PNGs for similar reasons.
But a web "ICO" which is not a real windows ICO, no. Literally the only thing they're good for is displaying a little picture in a web browser for a particular url. There's no use case there for IOT, not really.
That's why I won't include it.
Now, there are some fairly good use cases for implementing BMP support which is similar, but the actual spec for that is nasty, and way too complicated to do me any good for what I'm targeting. MS should have kept it simple.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Hmmm,
It's been a long time (months) since I last looked at what you are doing. I think last I looked you were ripping out parts of the TrueType parsing code from STB[^] or something.
I think it has JPEG and PNG decoders too, have you looked?
|
|
|
|
|
Yeah, but they're not geared for IoT. I had to gut the truetype stuff to make it all work from a stream instead of RAM. I wound up changing most of that file.
I don't want to go through the same thing with the PNG loading. I'd like to look for a different source - preferably something that loads it progressively.
The STB stuff is PC gaming code so it's geared for having gobs of RAM.
To err is human. Fortune favors the monsters.
|
|
|
|
|
honey the codewitch wrote: I'd like to look for a different source - preferably something that loads it progressively.
I just glanced over the PNG spec and progressive decoding only works on PNG images that were 'progressive encoded'. I think that means multiple IDAT sections. With a normal (non-progressive) PNG having a single IDAT.
So you would only benefit here if you control the image encoding too. Progressive encoded PNG images are rare today from my understanding.
|
|
|
|
|
I think I can still do it line by line
To err is human. Fortune favors the monsters.
|
|
|
|
|
Sweet,
Can you point out which step in decoding PNG images uses lines? I'd like to see how it's done.
Here's an overview[^], show me which decoding step uses lines.
|
|
|
|
|
I couldn't say off hand. I was looking at some code on github some months back on doing this.
To err is human. Fortune favors the monsters.
|
|
|
|
|
Any reason you can't convert a suitable image to jpg and add it to your GitHub project? It would then be easily accessible and your preferred format.
Keep Calm and Carry On
|
|
|
|
|
That would defeat the purpose of the demo, which is to load a picture off a network stream.
Edit: Never mind. I see what you mean. I'm a bit slow today. Yeah, I could do that.
To err is human. Fortune favors the monsters.
|
|
|
|
|
honey the codewitch wrote: Should I implement png?
Yes.
And WebP[^].
And AVIF[^].
And don't forget APNG[^] too. 
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
I'm not making a web browser. I need bare minimum functionality. This isn't a Cortex-A I'm targeting. 
To err is human. Fortune favors the monsters.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
