ESP8266 Community Forum

- Sat Jan 03, 2015 6:31 pm #6202

Just a FYI: I have developed httpd and the mkespfs stuff, but I can only develop/test it under Linux; I don't do Windows. If someone figures a way to make things work in Windows, please post them here; same goes for patches that make Windows-support easier. Don't go expecting Windows-support from me, though; I'll just advise you to install Linux

Hey all ... I'm already doing it with Windows64.
very efficiently. ...!

Could you please tell me how I can make the "webpages.espfs"-File with Windows?
Thank you

- Sat Jan 03, 2015 6:35 pm #6203

Just a FYI: I have developed httpd and the mkespfs stuff, but I can only develop/test it under Linux; I don't do Windows. If someone figures a way to make things work in Windows, please post them here; same goes for patches that make Windows-support easier. Don't go expecting Windows-support from me, though; I'll just advise you to install Linux

Hey all ... I'm already doing it with Windows64.
very efficiently. ...!

Could you please tell me how I can make the "webpages.espfs"-File with Windows?
Thank you

- Sat Jan 03, 2015 10:47 pm #6215

I've asked the same question/researched on these forums and +-180kB size is the max suggested.

0x40000 minus 0x12000 = 0x2E000... or 188416 bytes in decimal.

And the problem for me is that I can not take the size of the webpages.espfs 'one byte or more' than 128kB (131 072 bytes),
after that the serial puts out junk and every thing needs to be reflashed.

Regards - FeK9

Ps I've seen the post/response by Sprite_tm to your question, I be-leave he's referring to line 157 of the makefile.
if [ $$(stat -c '%s' webpages.espfs) -gt $$(( 0x2E000 )) ]; then echo "webpages.espfs too big!";

My guess is that the size difference is due to the fact that the webpages.espf is compressed using heatshrink. This means than the resources when uncompressed may be too large for the stack, clobbering your serial interface. I haven't looked at the details of the espfs, but that would be my guess. The heatshrink author mentions expecting a reduction in image data of 25/80% depending on resource type. If you don't have images, I would expect that ~128k to be a rather large dataset when uncompressed. Anyways, food for thought.

- Sun Jan 04, 2015 10:51 am #6251

My guess is that the size difference is due to the fact that the webpages.espf is compressed using heatshrink. This means than the resources when uncompressed may be too large for the stack, clobbering your serial interface. I haven't looked at the details of the espfs, but that would be my guess. The heatshrink author mentions expecting a reduction in image data of 25/80% depending on resource type. If you don't have images, I would expect that ~128k to be a rather large dataset when uncompressed. Anyways, food for thought.

That shouldn't matter: Heatshrink/espfs does on-the-fly streaming decompression, so there's never more than a buffer of about a K allocated for the decompressed file data, regardless of the filesize.

FeK9: The 128K limit is weird... it's not like there's a storage type that overflows at 128K or so. I'll try to look into it when I have some time.