You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Please note that the behaviour of String(F("text")).c_str() is undefined, since the lifetime of the String object is shorter than the function call. In single threaded systems which don't do much memory allocation you are probably fine, but not a good coding practice.
I think i saw this construct in our code base already, but unfortunately GCC 4 (or 5, for that matter) doesn't know how to generate warnings for such cases.
When igrr/newlib-xtensa#5 is merged, it will be possible to pass progmem strings directly to sprintf though, so the String(F(...)).c_str construct will not be needed.
It's clear as the white snow. But we do not see a strong influence on the work. It is clear that RAM is much faster, but it is terribly lacking. After loading the initial settings, 15,000 bytes remain, and at less than 10,000 bytes, the WEB starts to fail.
PS who can from conducting "commit" in Russian to communicate, you will contact me. [email protected]
The String().c_str() are indeed a bad construct, given that the returned pointer is provided by a temp object.
I am going to leave this open to cover the changes needed to *printf() format strings once the newlib changes are merged.
not found how to attach file ESP8266WebServer.cpp.patch
before 41424 (50%) = 40496
after 41408 (50%) = 40512
need commit
The text was updated successfully, but these errors were encountered: