Namely, define it if only `MG_ENABLE_HTTP_STREAMING_MULTIPART` and
`MG_ENABLE_FILESYSTEM` are on
PUBLISHED_FROM=3094738a4b76ceb09a3c2374e8d9d914b71c3441
And use it for `FW_ARCHITECTURE`. This is needed because with
Microchip's xc32, `-DFOO=\"bar\"` results in `FOO` being `bar`, not
`"bar"`, so we have to stringify differently.
PUBLISHED_FROM=d0aeebb55ef9e0e92a7f750104cd6600e0d9c636
Also fixed a memory leak on reconnect: previously, each reconnect was
creating a `struct mg_connection` which was never reclaimed
PUBLISHED_FROM=eefdcf557e032ac81d5ed3aba55ac912e400148e
The only client of `mg_http_common_url_parse` (namely,
`mg_connect_http_base`) expects `port_i` to be the index in the address
string at which the port was added.
PUBLISHED_FROM=b095926b5485e4674e3c59ff8481171831fb61ae
`MG_ENABLE_...` macros are not yet renamed because it involves some
logic change: e.g. instead of defining `MG_ENABLE_MQTT`, one should
define `MIOT_ENABLE_MQTT`, which should imply `MG_ENABLE_MQTT` as well.
Macro refactoring is going to be done in another PR.
PUBLISHED_FROM=f8b66d5772846126b94223a729a6bd87897f8de6
It wasn't checked for `NULL`, and on CC3200 NULL dereferencing addresses
doesn't cause a crash, so it worked by pure luck: ctx->ssl_key was 0.
After `mg_` to `miot_` refactoring it's not the case anymore (presumably
because linker arranged objects in a different order), so this bug shown
up.
PUBLISHED_FROM=0f1cc73a078c18432c68ae0f9b14dd06b3bb4279
Now all warnings are disabled on SDK files, and enabled on our
application's files.
Fixed a couple of things in nRF52 which were unnoticed because all
warnings were disabled there from the beginning.
PUBLISHED_FROM=a33ab22d0ce85efea364b80478986b88f4b0a3a2
Unfortunately it turns out Nordic's examples were able to build with GCC
just because they suppressed all warnings; I had to do the same here
because there are some macro redefinitions of arm-none-eabi-gcc and
lwip.
TODO: resolve that and do not suppress warnings.
PUBLISHED_FROM=e65c7b0f2451f85765d958d7f67eb9262f7bc929
Keil actually saves project on disk only when the project is closed,
that's why these changes were missing before
PUBLISHED_FROM=d8ebdaa009ff5676bf6c16aaa2fe8356b03d3b8b
By making dir hierarchy replicate that of examples provided with the nRF
IoT SDK.
Also, example for Keil uVision IDE is added.
PUBLISHED_FROM=1a3d9a54d15eb1acfb1b5d162138576a387ed21c
On my way, fixed a couple of cases where we had `()` in the header, and
non-empty argument list in the source file.
PUBLISHED_FROM=5519526cf84e2bbd425a726fcc112fea1a95cbf1
We haven't actually _fixed_ evaluation of precompiled bcode from
non-mmapped memory: we just never free it. For sane use case (do not
evaluate the same file more than once) it should be acceptable.
PUBLISHED_FROM=a9ca2d8fa72e9c40c93415d060ee8db697caa225
`v7_mk_undefined()` and `v7_mk_null()` are left unchanged, but marked
deprecated, and `V7_UNDEFINED` and `V7_NULL` are public now.
Xtensa code size with instrumentation: 590400 -> 590432, i.e. 32 bytes.
Xtensa code size without instrumentation: 530848 -> 531760, i.e. 912
bytes.
As objdump reveals, inside v7.c, `v7_mk_number` and others are inlined,
so the extra v7 argument is eliminated; outside of v7.c it's obviously
not inlined.
----
Also, v7_get_int() is added, which currently just casts stored double
value to int
PUBLISHED_FROM=e984b7308faf2380b3de388f238e0fae0aea545d
It would be probably good idea to also remove tools/docgen.py and
asciidoc.mk, but asciidoc.mk is still mentioned under `cloud/doc`, which
may contain some useful info which we'll need at least to review before
removing.
PUBLISHED_FROM=faf454d4c52a2f07ea8ac084cf0bd11a0c9c9b3b