Settings placeholder substitution #164

rjwats · 2020-06-28T11:17:30Z

Thx for this, nice little change. The templates weren't working for me as PlatformIO was expanding them to empty strings before handing them off to the build process. I couldn't even get it working by escaping the dollars - strange. Removing the dollar from our placeholders trick though, look OK?

I opted to restore the settings ifndefs. IMO it's nice to have the source code build without any external defines needing to be provided (especially useful if you want to produce a project from the source code which won't be built with PIO).

Just finishing off testing this under various *nix configurations.

Header files contains duplicates of factory values defined in factory_settings.ini Removed them to simplify the code.

… setting.

Updated documentation

Header files contains duplicates of factory values defined in factory_settings.ini Removed them to simplify the code.

… setting.

…laceholders

…ders

Use text formatting for default factory values to produce dynamic names.

clang formatting use 2 spaces in ini files use ${platform}-${chip_id} for hostname put chip id in brackets in AP SSID

- this is so src can be built without an exaustive set build-time defines - standardize ordering of defines: factory settings, paths, config

experiment with removing $'s from our format strings (they are being substituted with empty values by pio)

rename FactoryValue to SettingValue, put in own header give example of direct usage of FactorySetting::format in README.md

kasedy · 2020-06-28T21:29:39Z

We can replace dollar sign ($) with a hash (#) to avoid bash variable expansion.

# Conflicts: # README.md # features.ini

# Conflicts: # factory_settings.ini

rjwats · 2020-07-06T23:21:55Z

Updated to use #{value} seems to work

# Conflicts: # factory_settings.ini # lib/framework/APSettingsService.h

rjwats · 2020-07-07T23:15:33Z

Have been struggling with odd issues while testing this branch on one of my ESP8266 devices (a nodemcuv2 board).

It was exhibiting an unresponsive access point when erasing flash before uploading (effectively a factory reset). Oddly the culprit appears to be the use of flash string helpers and the following change solves the issue (my older nodemcuv0.9 board works fine without this change):

I'm not entirely sure why it solves the problem and would love to know if anyone is more enlightened.

proddy · 2020-07-08T13:00:20Z

I'm facing a similar problem, esptool.py erase_flash -> upload firmware brings up an empty Captive Portal page on an ESP8266. I'll try the workaround.

rjwats · 2020-07-08T13:12:09Z

Thanks, it would be good to know it's not just my device.

…

On Wed, Jul 8, 2020 at 2:00 PM Proddy ***@***.***> wrote: I'm facing a similar problem, esptool.py erase_flash -> upload firmware brings up an empty Captive Portal page. I'll try the workaround. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#164 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKE4VDAPSB6HSERLCTAX6TR2RUXHANCNFSM4OKPOL2Q> .

proddy · 2020-07-17T07:38:37Z

Haven't been able to reproduce it unfortunately. I'm back on master now.

kasedy · 2020-12-30T05:28:35Z

Do we plan to merge this feature into main?

rjwats · 2020-12-30T11:29:20Z

Think so, it's seemed like a sensible change. IIRC last time I looked at this I had a strange issue with it on one off my esp8266 devices. I couldn't see how it was related to the change, perhaps a dodgy flash chip on that particular device. I'll have another look this eve.

…

On Wed, 30 Dec 2020, 05:28 kasedy, ***@***.***> wrote: Do we plan to merge this feature into main? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#164 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKE4VCE5ABRFQXKFAQDTILSXK3A7ANCNFSM4OKPOL2Q> .

# Conflicts: # factory_settings.ini # lib/framework/WiFiSettingsService.h

rjwats · 2020-12-31T00:40:44Z

I couldn't re-create the issue I saw last time around so this is probs OK to merge. Have caught it up with master and applied the fix from before for good measure. Have run out of time tonight.. will review the code and merge this tomorrow

kasedy · 2020-12-31T04:40:04Z

I also had this issue in the past. It is not related with the change. I believe the issue is in esp8266-Arduino SDK as it has lots of open tickets. For a long time I used ESPAsyncWiFiManager as it is the most issue free WiFi manager for ESP8266. It has many workarounds to make WiFi work properly. For example it resets wifi before connecting to it [1]. Also it disconnects wifi on reset [2]. I think we should do something similar. By the way I skimmed the other projects. JustWifi and Tasmota have a workaround for WiFi connectivity issue [3] and [4].

[1] https://github.com/alanswx/ESPAsyncWiFiManager/blob/master/ESPAsyncWiFiManager.cpp#L575
[2] https://github.com/alanswx/ESPAsyncWiFiManager/blob/master/ESPAsyncWiFiManager.cpp#L691
[3] https://github.com/xoseperez/justwifi/blob/master/src/JustWifi.cpp#L65
[4] https://github.com/arendst/Tasmota/blob/development/tasmota/support_wifi.ino#L121

use chip id and unique id for more secure secret

rjwats · 2021-01-01T13:13:09Z

Just need to replace "chipId" with "salt" to secure the JWT secret properly. It's a pretty significant security hole as it stands. I'm sure I can get PIO to generate a random value for a salt at build time... will look this eve.

Arduino uses the ESP random number generator for "true random" numbers on both esp32 and esp8266 This makes a better JWT secret and may be useful for other factory defaults too In addition a modification has been made to force the FSPersistance to save the file if applying defaults

rjwats · 2021-01-02T00:13:24Z

I think this might be ready now, another user of the framework spotted an issue with using efuse/chipId for the hostname which this PR also resolves:

#212

In addition I have resolved a security issue, where the default JWT secret was previously derived from the station MAC address which may be sniffed from the WiFi network (not great if you care about security!)

Thankfully the Arduino random() function was adapted for these devices to use the built-in hardware true random number generator so it offers a good solution for the default JWT secret (with added bonus that it's random rather than derrived and is re-generated on factory reset). Nice.

fix default defines

kasedy and others added 19 commits May 31, 2020 21:21

Use text formatting for default factory values to produce dynamic names.

496c214

Header files contains duplicates of factory values defined in factory_settings.ini Removed them to simplify the code.

Merge branch 'master' into ft-placeholders

1e213f0

Use text formatting for default factory values to produce dynamic names.

490f2f6

Header files contains duplicates of factory values defined in factory_settings.ini Removed them to simplify the code.

Configured the WiFi host name to contain the device id by default

441410a

Removed possibility to use placeholders for FACTORY_WIFI_SSID factory…

ba057f3

… setting.

Update README.md

6f04d69

Updated documentation

Use text formatting for default factory values to produce dynamic names.

a6d4a17

Header files contains duplicates of factory values defined in factory_settings.ini Removed them to simplify the code.

Configured the WiFi host name to contain the device id by default

385fa45

Removed possibility to use placeholders for FACTORY_WIFI_SSID factory…

c355f6e

… setting.

Merge remote-tracking branch 'remotes/esp8266-react/master' into ft-p…

7fae43c

…laceholders

Merge remote-tracking branch 'remotes/kasedy/master' into ft-placehol…

91a7d82

…ders

Added a space to the end of the file to comply project code style

8aed287

Merge pull request #140 from kasedy/ft-placeholders

646f6cf

Use text formatting for default factory values to produce dynamic names.

fix typos

5c995ef

clang formatting use 2 spaces in ini files use ${platform}-${chip_id} for hostname put chip id in brackets in AP SSID

restore (and update) factory setting ifndefs

eefe9fc

- this is so src can be built without an exaustive set build-time defines - standardize ordering of defines: factory settings, paths, config

format and modify comment

4009316

escape spaces in pio defines

cb34dda

experiment with removing $'s from our format strings (they are being substituted with empty values by pio)

fix formatting in readme

0ba530d

rename FactoryValue to SettingValue, put in own header give example of direct usage of FactorySetting::format in README.md

auto format

2539c47

rjwats added 3 commits July 7, 2020 00:00

Merge branch 'master' into ft-placeholders

8619575

# Conflicts: # README.md # features.ini

use hash to delimit placeholders

afddead

Merge branch 'master' into ft-placeholders

6d21e73

# Conflicts: # factory_settings.ini

rjwats added 2 commits July 7, 2020 09:18

fix factory_settings.ini

9224897

Merge branch 'master' into ft-placeholders

7fe2770

# Conflicts: # factory_settings.ini # lib/framework/APSettingsService.h

rjwats added 2 commits December 31, 2020 00:08

Merge branch 'master' into ft-placeholders

5557282

# Conflicts: # factory_settings.ini # lib/framework/WiFiSettingsService.h

remove flash string helpers

115d9be

format ini file

5c76d6d

rjwats added 3 commits December 31, 2020 19:54

use MAC address instead of chip id for properly unique identifier

025ecb5

use lower case hex encoding for unique id

451805c

use chip id and unique id for more secure secret

fix comment

92b2249

rjwats linked an issue Jan 1, 2021 that may be closed by this pull request

Chip ID is not unique enough to be used as the secret or the hostname #212

Closed

rjwats added 2 commits January 2, 2021 00:16

Don't use spaces in default AP SSID

87e4508

restore helpful comment in factory_settings.ini

7650efa

fix default defines

rjwats merged commit e771ab1 into master Jan 3, 2021

rjwats deleted the ft-placeholders branch January 16, 2021 19:12

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Settings placeholder substitution #164

Settings placeholder substitution #164

Uh oh!

rjwats commented Jun 28, 2020

Uh oh!

kasedy commented Jun 28, 2020

Uh oh!

rjwats commented Jul 6, 2020

Uh oh!

rjwats commented Jul 7, 2020

Uh oh!

proddy commented Jul 8, 2020 •

edited

Loading

Uh oh!

rjwats commented Jul 8, 2020 via email

Uh oh!

proddy commented Jul 17, 2020

Uh oh!

kasedy commented Dec 30, 2020

Uh oh!

rjwats commented Dec 30, 2020 via email

Uh oh!

rjwats commented Dec 31, 2020

Uh oh!

kasedy commented Dec 31, 2020

Uh oh!

rjwats commented Jan 1, 2021

Uh oh!

rjwats commented Jan 2, 2021

Uh oh!

Uh oh!

Settings placeholder substitution #164

Settings placeholder substitution #164

Uh oh!

Conversation

rjwats commented Jun 28, 2020

Uh oh!

kasedy commented Jun 28, 2020

Uh oh!

rjwats commented Jul 6, 2020

Uh oh!

rjwats commented Jul 7, 2020

Uh oh!

proddy commented Jul 8, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rjwats commented Jul 8, 2020 via email

Uh oh!

proddy commented Jul 17, 2020

Uh oh!

kasedy commented Dec 30, 2020

Uh oh!

rjwats commented Dec 30, 2020 via email

Uh oh!

rjwats commented Dec 31, 2020

Uh oh!

kasedy commented Dec 31, 2020

Uh oh!

rjwats commented Jan 1, 2021

Uh oh!

rjwats commented Jan 2, 2021

Uh oh!

Uh oh!

proddy commented Jul 8, 2020 •

edited

Loading