Support Flash Mode Dout

[arendst]

I would like to advertize your tool for my Tasmota project as it seems to be a single container executable.

To get rid of all kinds of issues regarding different hardware (ESP8266 vs ESP8285) I decided to only support Flash Mode Dout as it works fine on both chips and all Tasmota supported hardware.

Would you consider implementing the Dout option in your Gui?

[marcelstoer]

Sure, why not. For consistency reasons I'd have to support all four flash modes rather than just three. Yet, I actually see no need for QOUT.

Note to self: https://www.esp32.com/viewtopic.php?p=5523&sid=c86b6012ad1ab15eec49e0158c649696#p5523

[marcelstoer]

@arendst what's your take on the question I raised in the above comment?

[arendst]

I'm not interested in QOUT. I'm only interested DOUT as that is the only way to go in my future supported hardware with Tasmota as the manufacturers have massively chosen for esp8285 which only supports DOUT.

[marcelstoer]

I know but the big picture is bigger than your specific use case.

[arendst]

Well I asked with regards to my use case. If it's too much of a problem alright with me. Will look for another tool or clone yours and tune to my likings.

Thnx.

[marcelstoer]

It ain't "too much of a problem" and there's no reason to get annoyed. Every change offers the opportunity to (re-)evaluate an existing feature slightly beyond the one specific problem that triggered the change in first place.

[arendst]

Agree.

[ThomDietrich]

@marcelstoer good thinking on your side and I'd also not see the need. From what I read, @arendst is of course not against QOUT, it just doesn't matter to him. Maybe he got the question wrong.
Anyhow, I'd tend to say that you could just add all four methods as option and be done with it. Either choice is fine and depends on your willingness to implement and test everything. Thanks for your effort.

[marcelstoer]

Looking forward to your feedback, here's a binary for testing:
NodeMCU-PyFlasher-dout-support.exe.zip

[arendst]

Thnx. Will look in to it soon and let you know.

[arendst]

Hi Marcel,

looks good but...

To provide a single binary file for all supported hardware my source is ALWAYS linked with 1MB no SPIFFS and I store settings in the 8k area just before the end of SPIFFS. If my source is linked with 4MB no SPIFFS it works fine too.
I noticed that I cannot tell your program to use a fixed flash size of 1MB and therefore on a Wemos it changes to 4MB (probably caused by esptool and lack of flash parameter -fs 1MB) which will fail as I linked it with a 1MB and expect to store my settings at the linkers SPIFFS end -8k.

Would it be possible to also provide the possibility to select the desired flash size in the GUI.

While thinking and probably asking too much you might even provide a GUI derivative with fixed DOUT and 1M flash selected without GUI control and call it Tasmota-Flasher.exe (Without the Py as you managed to include Python virtually without the user to know) ;-)

[marcelstoer]

Thanks for the feedback.

Would it be possible to also provide the possibility to select the desired flash size in the GUI.

It would but it's highly unlikely that I'll implement that. I actually see value in providing as few knobs to turn as possible, see https://github.com/marcelstoer/nodemcu-pyflasher#why-this-project-exists to understand my motivation. Ideally it should just work w/o the user having to worry about the gory details.

Related to #11.

[ThomDietrich]

"as few knobs as possible" -> How about abstract presets then? Currently the UI asks for Baud rate and Flash mode but the uninitiated user will not even be able to decide that. Apart from these "Custom" settings, there could be a first selection element offering presets for famous devices or projects like "NodeMCU devkit 1.0", "WeMos D1 mini", or "Sonoff-Tasmota".
I feel like this would be a huge step in terms of user experience and the implementation effort seems reasonable and #11 could be part of that. wdyt?

[marcelstoer]

I've been thinking about something similar in the past but didn't spend too much time developing the idea in my head. This time I did and I think it could turn out to be quite cool. Please open a separate issue so we can continue the discussion there.