* Remove warnings, errors during host tests in CI
Debug strings often included format parameters which did not exactly match
the passed in format parameters, resulting in warnings in the host test build
process like
````
/home/runner/work/Arduino/Arduino/libraries/ESP8266WiFi/src/ESP8266WiFiAP.cpp:107:20: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 3 has type ‘size_t’ {aka ‘long unsigned int’} [-Wformat=]
107 | DEBUG_WIFI("[AP] SSID length %u, too long or missing!\n", ssid_len);
| ~~~~~~~~
| |
| size_t {aka long unsigned int}
````
Fix by applying casting or PRxxx macros as appropriate.
Also, fix one debug message which was trying to use a `String` as a `char *`:
````
/home/runner/work/Arduino/Arduino/libraries/ESP8266WiFi/src/ESP8266WiFiMulti.cpp: In member function ‘wl_status_t ESP8266WiFiMulti::connectWiFiMulti(uint32_t)’:
/home/runner/work/Arduino/Arduino/libraries/ESP8266WiFi/src/ESP8266WiFiMulti.cpp:331:34: warning: format ‘%s’ expects argument of type ‘char*’, but argument 3 has type ‘String’ [-Wformat=]
331 | DEBUG_WIFI_MULTI("[WIFIM] Connecting %s\n", ssid);
````
* Clean up SpeedTest.ino host build
esp_yield() now also calls esp_schedule(), original esp_yield() function renamed to esp_suspend().
Don't use delay(0) in the Core internals, libraries and examples. Use yield() when the code is
supposed to be called from CONT, use esp_yield() when the code can be called from either CONT or SYS.
Clean-up esp_yield() and esp_schedule() declarations across the code and use coredecls.h instead.
Implement helper functions for libraries that were previously using esp_yield(), esp_schedule() and
esp_delay() directly to wait for certain SYS context tasks to complete. Correctly use esp_delay()
for timeouts, make sure scheduled functions have a chance to run (e.g. LwIP_Ethernet uses recurrent)
Related issues:
- #6107 - discussion about the esp_yield() and esp_delay() usage in ClientContext
- #6212 - discussion about replacing delay() with a blocking loop
- #6680 - pull request introducing LwIP-based Ethernet
- #7146 - discussion that originated UART code changes
- #7969 - proposal to remove delay(0) from the example code
- #8291 - discussion related to the run_scheduled_recurrent_functions() usage in LwIP Ethernet
- #8317 - yieldUntil() implementation, similar to the esp_delay() overload with a timeout and a 0 interval
- =default for default ctor, destructor, move ctor and the assignment move
- use `std::unique_ptr<WiFiClient>` instead of raw pointer to the client
- implement `virtual std::unique_ptr<WiFiClient> WiFiClient::clone()` to safely copy the WiFiClientSecure instance, without accidentally slicing it (i.e. using the pointer with incorrect type, calling base WiFiClient virtual methods)
- replace headers pointer array with `std::unique_ptr<T[]>` to simplify the move operations
- substitute userAgent with the default one when it is empty
(may be a subject to change though, b/c now there is a global static `String`)
Allow HTTPClient to be placed inside of movable classes (e.g. std::optional, requested in the linked issue) or to be returned from functions. Class logic stays as-is, only the underlying member types are changed.
Notice that WiFiClient connection object is now copied, and the internal ClientContext will be preserved even after the original WiFiClient object was destroyed.
replaces #8236resolves#8231
and, possibly #5734
If a server returns "HTTP/1.x -8 OK", for example, it can misguide an application developer into freeing less-important memory so the request can be retried and succeed, when the problem is in the server.
_returnCode is never used anywhere else, but it could still contain a negative value returned by a broken server and therefore could cause troubles in the future (if _returnCode is in fact used)
Response bodies are ignored when _transferEncoding == HTTPC_TE_IDENTITY and there is no Content-Length header. The added code here fixes that issue.
Add logic to writeToStreamDataBlock to only read what's available so as to avoid timeout, and adjust formatting.
* Because of git problems, start from a new fork and create a new PR. This was PR #6457
* Style update to pass Travis
* Update ReuseConnectionV2.ino
* fix + enforce testing http code
per @earlephilhower review
* Close connection before ::connecting on HTTP/1.0
HTTPClient never actually closes the TCP connection on its own. It will leave the TCP connection open unless you explicitly do a getString which makes a StreamString and stuffs it with the HTTP server response, at which point the HTTP server itself will close the connection.
If you check the HTTP error code and find failure, unless you do a getString and throw it away, it won't disconnect. Even in HTTP/1.0 or in cases when you haven't enabled _reuse.
Change the logic in ::connect to only reuse the connection when it is specifically allowed. Otherwise, fall back to re-connection.
* Adjust example per request
Do single URL get in each loop, avoid infinite for loop at end.
* Fix astyle
* Clean up final pass notice
* Fix example syntax error
Editing code in a web textbox without running it is a painful process.
Co-authored-by: Earle F. Philhower, III <earlephilhower@yahoo.com>
* Add way to force follow redirections in `HTTPClient`
* Follow other client implementations about `HTTP_CODE_FOUND`; Small rewrite of `sendRequest` function of `HTTPClient`
* Better names for follow redirection modes in `HTTPClient`
Also changed a bit order of the enums (0 element to be DISABLED)
* Rewrite `sendRequest` to remove recursion
Also got rid of unnecessary `redirectCount` field. Now redirect counting and limiting is handled in `sendRequest` directly.
* Use new `setFollowRedirects` of `HTTPClient` instead deprecated one.
* More explanatory comment for `followRedirects_t` in HTTPClient
sendRequest has a major problem when sending a big payload, the comparator in the IF loop has its two operators changed, so the last part of payload is never sent
* Reduce temporary string creation/reallocation in HTTPClient
This improves both performance due to fewer memory allocations/copies
as well as reduces code size by ~ 25% (150 bytes)
* Add more correct reservation calculation
Co-authored-by: Develo <deveyes@gmail.com>
Use the proper api (::clear(), isEmpty()) instead of doing
comparisons/assignments of empty strings. Also fix mixture
of tabs and spaces in the source code.
This is all @dirkx , whose PR unfortunately got borked when we were
trying to update it to the new format. As @dirkx said:
When sending POST responses of well over a K - _write() may not sent it
all. Make sure we do -- but cap the individual writes - as somehow large
2-3k blobs seem to cause instability (on my 12F units).
Supercedes #2528
* Cleanup base64::encode functions
The implementation choice here using libb64 is generally good as it
is a relatively fast implementation, however the adaptation to
use PROGMEM for the translation function was a bad choice, as reading
randomly PROGMEM with byte-wide access is very very very slow.
Doing a naive if-snake is between 20% and 55% faster and uses less
flash (about 120 bytes less) and also for reasons I don't understand
8 bytes less data RAM (maybe the removal of static?).
In addition the base64::encode function was allocating for larger
input a huge amount of memory (twice the total size). we can reduce
that by doing a chunk-wise conversation to base64.
* Create authorisation base64 encoded string without newlines
Rather than first creating a string with newlines and then
stripping it away in the fast path of constructing the query,
we can call the right method and trust that the result does
not have newlines anymore.
There are actually several instances where we pass in read-only
parameters as pass-by-value, where in the case of String() that
is inefficient as it involves copy-constructor/temp string creations.
We can avoid that, similarly to single character string concatenations
done via string literals instead of char literals.
* Put longer string literals into PROGMEM
* Use Flash Strings for Debug output
This is hopefully very infrequently used, so it shouldn't
be in main memory.
The function accidentally compared the current location instead
of the passed in (new) location, which didn't match intended
logic and the code comment.
* fix DEBUG macros
All fmt strings in flash
fix#5658
This also allows to avoid warnings and easy mistakes with (no brace):
if (something)
DEBUGV("blah");
* use newlib unaligned-compatible printf for DEBUGV
* remove useless putprintf since ::printf already uses ets_putc
* optionally allow redirects on http OTA updates
* Refactored HTTPClient::begin(url...) & setURL functions, now only beginInternal parses URL, sets ports
Added HTTPRedirect example.
* fix indentation for style check
* add space after while for style check
* don't use deprecated begin method in redirect example
* moved redirect handling code to HTTPClient.
only GET and HEAD requests are currently handled automatically
Redirects that fail to be automatically handled return the redirect code as before
* added support for POST/303 redirect
added device redirect tests
* add missing getLocation() implementation
* if the new location is only a path then only update the URI
* WIP compile examples on host with 'make examples'
* WIP bufferize tcp input
* WIP Makefile
* WIP network to rework, tcp/udp to factorize, udp addresses broken
* minor changes to the core
* WIP basic udp working
* WIP mdns
* WIP mcast receiving, not sending
* WIP mdns OK
* beta version
* SSL + doc
* update travis host test command
* licenses
* typo
* doc: arduino builder is not around: declare functions before calling them
* fix with latest SSL PR, compile in 32 bits mode
* fix make clean
* make -m32 optional
* 32bits compiler ability tester
* WIP
* WIP (fix 1 vtable error, still another one to hunt with using spiffs)
* example astyle
* fix os_printf_plus
* load / save mock spiffs
* fix style
* fix using spiffs/mock
* don't mess ram
* update doc
* remove leftover
* optimization -Os except for CI, rename ARCH32 to FORCE32
* revert useless cast (not even compiled)
* remove unused function
* use proper type for pointer arithmetics
* makefile: sketch object and cpp file moved to bin/ directories
easier to clean, and IDE don't like them
* changes for review
* make use of %zd
* less verbose makefile by default (option)
* update readme
* update HTTPClient API usage
skip the second POST as end() has different semantics and nulls the client pointer
use bearssl in ssl tests
add delay in python side when shutting down http web server so MacOS does not complain about address already in use
* fix crash if GET/POST was called after end() without a new begin()
update double POST test to insure no crash if POST called after end()
test now are for both AxTLS and BearSSL
* fix small comment typo
* Removed _client->stop() from destructor; some minor changes
* Changed BasicHttpsClient.ino to allocate BearSSL::WiFiClientSecure object on the heap in stead of stack
* Removed unnecessary code
* Correcting bad fix for #5216
* Minor formatting to pass Travis tests
* Changed client * to std::unique_ptr<> client
* Updated example
* Removed _client->stop() from destructor; some minor changes
* Changed BasicHttpsClient.ino to allocate BearSSL::WiFiClientSecure object on the heap in stead of stack
