These are chat archives for esp8266/Arduino

20th
Jul 2015
probonopd
@probonopd
Jul 20 2015 07:12
Can it be that OTA in git esp8266/Arduino is entirely broken at the moment?
i get crashes during OTA
Me No Dev
@me-no-dev
Jul 20 2015 07:15
Network is "broken" currently
Read this thread from friday to see why
probonopd
@probonopd
Jul 20 2015 07:20
thanks @me-no-dev
probonopd
@probonopd
Jul 20 2015 07:30
git reset --hard 1154545fa8db7f1ce183865b776cadffc4e8d03fdid the trick for me
is it intentionally that NodeMCU 1.0 does not have the OTA menus enabled?
probonopd
@probonopd
Jul 20 2015 09:34
is it possible to call SDK functionality like system_phy_set_max_tpw from Arduino code?
Ivan Grokhotkov
@igrr
Jul 20 2015 09:39
yes, just include the "user_interface.h" header file
extern "C" {
#include "user_interface.h"
}
probonopd
@probonopd
Jul 20 2015 09:40
Thanks igrr
Me No Dev
@me-no-dev
Jul 20 2015 09:49
@probonopd OTA is added to the generic, because it's a work in progress
when it is made to work properly it will work for all
probonopd
@probonopd
Jul 20 2015 10:13
excellent!
Ivan Grokhotkov
@igrr
Jul 20 2015 10:31
@Makuna why is ETS_INTR_WITHINISR checking PS.UM bit?
i'm asking because it's normally set to 1, which makes ETS_INTR_WITHINISR return 1 most of the time
Ivan Grokhotkov
@igrr
Jul 20 2015 10:53
@probonopd regarding esp8266/Arduino#573
any particular reason to call it "Native" functionality? :)
as far as I remember, "native" was used in .NET to describe code running outside of a VM as opposed to code running in a VM
we don't have a VM here, so SDK code is as native as Arduino code
Perhaps calling this example something like "CallSDKFunctions" would be less ambiguous?
also typo in word Espressif in line 26
probonopd
@probonopd
Jul 20 2015 11:02
Thanks for catching this, will do
done
Ivan Grokhotkov
@igrr
Jul 20 2015 11:33
@Makuna also it seems that PS.EXCM bit is not set when we are in a GPIO interrupt
probonopd
@probonopd
Jul 20 2015 11:49
@me-no-dev is there an example somewhere on how to run both a http server and the ota service? I seem to be able to only run one or the other, stupid me.
Ivan Grokhotkov
@igrr
Jul 20 2015 12:50
@me-no-dev https://github.com/igrr/Arduino/tree/optimistic-yield could you please give it a go and see if it resolves network issues?
i did a test with mDNS + web server, seems to be fine
Russ Mathis
@RussMathis
Jul 20 2015 14:29
Has anyone had a chance to review the emw3165 yet?
Ivan Grokhotkov
@igrr
Jul 20 2015 14:31
not me, but regarding the "esp8266 killer"...
https://twitter.com/EspressifSystem/status/622710136096698369
Russ Mathis
@RussMathis
Jul 20 2015 14:51
Much smaller that's for sure!
Ivan Grokhotkov
@igrr
Jul 20 2015 14:52
this is esp8266ex in this pic — so we don't know what the new chip will look like yet
chad cormier roussel
@chadouming
Jul 20 2015 16:10
esp8366 xD
Harrison Mclean
@h4rm0n1c
Jul 20 2015 16:15
I'm having some trouble with memcpy_P after Makuna's changes, getting lots of: undefined reference to memcpy_P(void, void const, unsigned int)
I think it's gone, I can't find an actual function to go with it in pgmspace.c
Harrison Mclean
@h4rm0n1c
Jul 20 2015 16:28
Yeah, it's definitely missing.
Michael Miller
@Makuna
Jul 20 2015 16:56
@igrr PS.UM is supposed to be set only when the Level 1 exceptions vectors are being run (IO PINS, some timers, etc).
@igrr where you finding it being set while in the main loop? I was not seeing this.
Ivan Grokhotkov
@igrr
Jul 20 2015 16:58
I tried printing it from setup() and loop() methods and it was set
not at the computer now, will post my sketch later
Michael Miller
@Makuna
Jul 20 2015 17:02
memcpy is missing!? Merging with git without using more powerful tools will be the death of me, but the more powerful tools tend to require too much understanding of all the details of how git works which adds another layer of making mistakes.
I will get it added back now.
Harrison Mclean
@h4rm0n1c
Jul 20 2015 17:02
I made a pull request already
If that's ok, not wanting to tread on toes or anything.
Michael Miller
@Makuna
Jul 20 2015 17:06
@h4rm0n1c looks good. thanks and no problems from me, I contribute, I don't own.
Michael Miller
@Makuna
Jul 20 2015 17:18
@igrr I think I know why PS.UM is still set, it seems that some of the return from interrupts instructions do not reset them, expecting the code to do this in the right cases, well, expecting the compiler to do this.
@igrr I will look at changing this to use the PS.IntLevel instead.
Ivan Grokhotkov
@igrr
Jul 20 2015 17:22
ok,
Michael Miller
@Makuna
Jul 20 2015 17:40
@igrr I am going to be out for the rest of the day, the change to use intlevel is as follows if you want to give it a try
inline bool ETS_INTR_WITHINISR()
{
    uint32_t ps;
    __asm__ __volatile__("rsync; rsr %0,ps":"=a" (ps));
    // PS.EXCM and PS.UM bit checks
    // return ((ps & ((1 << 4) | (1 << 5))) > 0);
    return ((ps & 0x0f) != 0);
}
I have not had time to test it yet
probonopd
@probonopd
Jul 20 2015 19:34
someone knows a more professional yet cheap solution to connect to ESP-12Es without soldering? https://cloud.githubusercontent.com/assets/2480569/8785241/9d0c7ebe-2f26-11e5-9c29-dbe2f3f63636.jpg
Ivan Grokhotkov
@igrr
Jul 20 2015 19:54
pogo pins?
probonopd
@probonopd
Jul 20 2015 19:57
good idea
Me No Dev
@me-no-dev
Jul 20 2015 23:05
@igrr will give it a go :)