I'm trying to port some code over to OSX - it appears that setsockopt has a few differences from Linux. The one I've found and corrected so far is using TCP_KEEPALIVE instead of TCP_KEEPIDLE.
What is the equivalent to TCP_USER_TIMEOUT for OSX?
Using TCP_USER_TIMEOUT fails to compile and prints an undeclared identifier error
I find a macro named "TCP_CONNECTIONTIMEOUT" in tcp.h. I use it to replace TCP_USER_TIMEOUT on Mac. It works well. You may try it.
Another possible option is TCP_RXT_CONNDROPTIME from the same header:
#define TCP_RXT_CONNDROPTIME 0x80 /* time after which tcp retransmissions will be
* stopped and the connection will be dropped
*/
I've found this weird commit where there are three implementations provided for TCP_USER_TIMEOUT:
Linux: TCP_USER_TIMEOUT is set manually (seems like this code is quite outdated);
Darwin: TCP_RXT_CONNDROPTIME is used;
Win: no implementation provided.
This commit might be a clue, but I haven't tested it myself. I just hope that this information could be helpful to someone.
Check TCP_USER_TIMEOUT and TCP_CONNECTIONTIMEOUT descriptions in the Linux man and BSD man - they doesn't seem to be very similar.
Related
I need to create ECDSA keys and used old OpenSSL code until now, which dosn't compile any more under OpenSSL3, there are now lots of deprecation errors. I invested now several days with search & try, but I can't solve it. I am trying the simple example [https://fm4dd.com/openssl/eckeycreate.shtm][1], but already the first lines create these errors:
WorkerThread.cpp(34,2): error C4996: 'ERR_load_BIO_strings': Since OpenSSL 3.0
WorkerThread.cpp(47,19): error C4996: 'EC_KEY_new_by_curve_name': Since OpenSSL 3.0
WorkerThread.cpp(54,2): error C4996: 'EC_KEY_free': Since OpenSSL 3.0
So the first idea would be to eliminate these compiler errors somehow. I found the switches
#define OPENSSL_API_COMPAT 30000
#define OPENSSL_NO_DEPRECATED
but they create even more errors, because now suddenly even the EC_KEY is unknown to the compiler. No solution.
So there stays the hard way to try to migrate the code somehow to OpenSSL3, because there seems to be no example code in the whole internet for simply creating an ECDSA key with OpenSSL3, and extract the private and public part. But just changing single functions with an OpenSSL3 matching function doesn't solve it, because the whole concept changes (switching from low level API to high level, with a completely different programming model).
I have tried fiddling around with lots of examples, but it looks like my programming task that seems so simple is a very complicated thing. Any help is appreciated, but I want to stay in C/C++ because my application is performance critical.
Environment: MS Win10, MS Visual Studio 2019, latest OpenSSL 3
[1]: https://fm4dd.com/openssl/eckeycreate.shtm
WorkerThread.cpp(34,2): error C4996: 'ERR_load_BIO_strings': Since OpenSSL 3.0
You don't need to call any of those initialisation functions. Just remove them from your code. They are not needed with any version of OpenSSL from 1.1.0 onwards.
In the example, all of the code for getting an EC_GROUP, creating an EC_KEY, generating a key, setting the OPENSSL_EC_NAMED_CURVE flag and assigning it to an EVP_PKEY can be completely removed. Replace it with a simple call to EVP_EC_gen(). See the man page here:
https://www.openssl.org/docs/man3.0/man3/EVP_EC_gen.html
E.g. all of that code can be replaced with something like
pkey = EVP_EC_gen("secp521r1");
if (pkey == NULL) {
BIO_printf(outbio, "Error generating the ECC key.");
abort();
}
Trying to run this USB Serial example (bottom) to learn MBED, but I get the following compilation error:
class "USBSerial" has no member "printf"
Is it possible it isn't implemented for the STM32F411? Or is this a problem with MBED itself?
Seems like this should be basic functionality. Not finding much useful info on google when searching for this error. Has anyone else seen this error before?
potentially useful details:
IDE: vscode/platformIO
platformio.ini:
[env:nucleo f411re]
platform = ststm32
framework = mbed
board = nucleo_f411re
monitor_speed = 115200
MBED version: 6.2 (as I recall from memory, though I doubt it matters since I checked the docs for a few versions and the API and example appears unchanged)
The method printf() (which is a C, not C++ concept anyway) does not exist, simple as that. Use sprintf() if that's what you're familar with, then USBSerial.write(), perhaps.
I red the following article and it seems like I should be able to define a variable of type varying that's size limit will be 16mb instead of 65535 characters.
http://www.mcpressonline.com/programming/rpg/v6r1-rpg-enhancements.html
This forum post suggests that it is posible to do this in V6.1.
http://www.code400.com/forum/forum/iseries-programming-languages/rpg-rpgle/11426-character-string-max-length
D BigVarying S a Len(25000000) Varying(4)
When I try to implement this I get the following errors and it seems like I cannot use Len(25000000) Varying(4)
Maybe I don't understand what exactly is meant by V6.1 but I checked on the Green screen using the command DSPSFWRSC and get the following Release V6R1..
I also checked by using I systems navigator and I checked the servers properties and it is : i5/OS version Version 6 Release 1 Modification 0.
The IDE I use to compile is IBM Websphere development studio
Version: 7.0.0
Build id: 20070202_0030
Trying to compile a RPGLE function.
Am I checking the correct version or is there a RPG version and how do you check it.
can you please help me out of my confusion.
V6.1 is your operating system version and you can refer to the V6.1 ILE RPG Language Reference here to see if it's possible. I think what you're looking for is on page 185.
I've just tested this in 7.2 and I am not getting the same errors.
D BigVarying S a Len(25000000) Varying(4)
RNF0501E Length of character item exceeds 16773104; length defaults to 16773104.
RNF0231E Length of varying length character item exceeds 16773100; length defaults to 16773100.
//Dcl-S BigVarying Varchar(25000000);
BigVarying = 'Hello world';
Return;
So, as the (duplicate) error mentions
Length of character item exceeds 16773104; length defaults to 16773104.
If you'd like a bigger length you'll need to update to a newer version of IBM i - but the max length is also 16773104, meaning 25000000 is invalid.
While RPGLE supports 16MB variables from 6.1 forward and you are on 6.1..
You have two problems:
25000000 (23.8MB) is bigger than 16773104 (~16MB)
Websphere Development Studio v7 (WDSc) is older than IBM i 6.1 (your build date is 2007, the article you're referencing came out in 2008). So your IDE doesn't recognize the new keywords and new max size.
Problem #2 isn't a deal breaker, you can simply ignore the errors in the IDE and compile on ther server successfully. If you were using the green screen editor SEU, you'd have to do the same as IBM stopped enhancing SEU at 6.1.
I am trying to get started with fastcgi development, therefor I download the reference implementation of libfcgi and tried to get a testprogram to run with lighttpd. Since fcgi_stdio allows for cgi backwards compatibility, I decided to start with that.
However I could not get the examples/tiny-fcgi.c to work with lighttpd, it yielded an internal server errror 500, the same configuration allows to run the lighttpd example program (http://redmine.lighttpd.net/projects/lighttpd/wiki/Docs_ModFastCGI, below C/C++ FastCGI on lightty named socket) flawless. However it is totally unclear to me why the default supplied example would not work.
Questions:
What is wrong with the examples/tiny-fcgi.c example from the reference implementation causing lighttpd to return error 500?
Which implementation of fastcgi is preferable for c++ development (fcgi_stdio, fcgiapp, Other)? (There is something for streams in the fcgi pack, however I failed to find good/any documentation)
I have tested the unmodified example (Ubuntu/13.10/amd-64, Apache/2.4.6, libapache2-mod-fastcgi/2.4.7~0910052141-1.1, libfcgi-dev/2.4.0-8.1ubuntu4) and it runs ok:
manuelz#garibaldi:~$ curl habrich/tiny-fcgi
<title>FastCGI Hello! (C, fcgi_stdio library)</title>
<h1>FastCGI Hello! (C, fcgi_stdio library)</h1>
Request number 2 running on host <i>habrich</i>
Answers:
Hard to tell without knowing your configuration, but here's a shot in the dark: getenv will return NULL if SERVER_NAME is undefined.
fcgiapp is the base implementation: I would use that one for new development. fcgi_stdio is a wrapper for compatibility with CGI: use that for migrating legacy CGI projects. Quote:
fcgi_stdio is implemented as a thin layer on top of fcgiapp
You can find decent documentation for fcgiapp in the fcgiapp.h header.
I do not know whether your problem has the same origin as mine had, but at least it has the same symptoms.
There are different versions of the tiny-fcgi example. I first tried the one given here (example 1). This failed with an internal server error. However, the example given here (which does basically the same) works. The small but crucial difference is that the working code uses
getenv("SERVER_NAME")
instead of
getenv("SERVER_HOSTNAME")
not exactly the same for me, but if i avoid using getenv the example runs without errors, have to discover why getenv leads into error
I'm working on a software which uses FFMPEG C++ libs to make an acquisition from an UDP streaming.
FFMPEG (1.2) is implemented and running but I get some errors (acquisition crashes and restarts). The log displays the following message:
*Circular buffer overrun. To avoid, increase fifo_size URL option. To survive in such case, use overrun_nonfatal option*
I searched online for documentation about how to use this option, but I only got informations about how to use when running directly ffmpeg executable.
Would someone know how to set the correct option in my C++ code to:
- increase fifo_size
- use overrun_nonfatal option
Thanks
The same option works from command line or C++ libraries, you need to modify your UDP URL as follows:
If you original URL looks like this:
udp://#239.1.1.7:5107
Add the fifo_size and overrun parameters like this:
"udp://#239.1.1.7:5107?overrun_nonfatal=1&fifo_size=50000000"
Remember to escape the URL with quotes.
overrun_nonfatal=1 prevents ffmpeg from exiting, it can recover in most circumstances.
fifo_size=50000000 uses a 50MB udp input buffer (default 5MB)
The only documentation is in the source code:
http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavformat/udp.c;h=5b5c7cb7dfc1aed3f71ea0c3e980be54757d3c62;hb=dd0a9b78db0eeea72183bd3f5bc5fe51a5d3f537
I don't have enough reputation to comment the other answer, but if I did I would say that studying the source linked in the answer:
fifo_size is measured as multiples of 188 Byte (packets) according to the line:
s->circular_buffer_size = strtol(buf, NULL, 10)*188;
so whilst Grant is roughly correct that "default 5MB", because of the line:
s->circular_buffer_size = 7*188*4096;
If you want a circular buffer of 50MB you should really set the fifo_size parameter to something closer to 50*1024*1024/188 otherwise 50000000 will give 50000000*188 bytes which is closer to 8965MB!