www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - ImportC and color syntax highlighting

reply Walter Bright <newshound2 digitalmars.com> writes:
Now that I'm writing C test cases again (for #ImportC), the lack of color
syntax 
highlighting was annoying. Fixed it:

https://github.com/DigitalMars/med/commit/adcc084e21427236957a7f94f3547b22a89c7c1c

and did C++ while I was at it.
Sep 11 2021
parent reply jfondren <julian.fondren gmail.com> writes:
On Saturday, 11 September 2021 at 22:25:31 UTC, Walter Bright 
wrote:

 Now that I'm writing C test cases again (for #ImportC), the 
 lack of color syntax highlighting was annoying. Fixed it:

 https://github.com/DigitalMars/med/commit/adcc084e21427236957a7f94f3547b22a89c7c1c

 and did C++ while I was at it.


Neat, I wasn't aware of this.

It's a little bit annoying to compile from source as my distro's 
dmd/ldc2 packages don't include 32-bit libraries but this doesn't 
build under -m32, and gdc doesn't come with gdmd. But with gdmd, 
it builds out of master with

```
make DMD=gdmd -f linux.mak
```

It only has one v1.0.0 tag from 2014, so `dub run med` fails with 
a bunch of ancient errors:
```
.dub/packages/med-1.0.0/med/src/med/buffer.d(254,16): Error: 
instead of C-style syntax, use D-style `dchar[6 + 1] b
```

meanwhile `dub run med ~master --compiler=gdc --arch=x86` pulls 
the latest code but can't build due to -Werror and some 
"statement is not reachable" warnings.

When running it, it pegs a CPU at 100% because it's constantly 
running ttkeysininput(). That's enough to cause a fan to run at 
high speed on this laptop. But, that's easily fixed with select() 
or similar against stdin.
Sep 11 2021
next sibling parent Walter Bright <newshound2 digitalmars.com> writes:
On 9/11/2021 4:04 PM, jfondren wrote:

 On Saturday, 11 September 2021 at 22:25:31 UTC, Walter Bright wrote:

 Now that I'm writing C test cases again (for #ImportC), the lack of color 
 syntax highlighting was annoying. Fixed it:

 https://github.com/DigitalMars/med/commit/adcc084e21427236957a
f94f3547b22a89c7c1c 


 and did C++ while I was at it.


 
 Neat, I wasn't aware of this.
 
 It's a little bit annoying to compile from source as my distro's dmd/ldc2 
 packages don't include 32-bit libraries but this doesn't build under -m32, and 
 gdc doesn't come with gdmd. But with gdmd, it builds out of master with
 
 ```
 make DMD=gdmd -f linux.mak
 ```
 
 It only has one v1.0.0 tag from 2014, so `dub run med` fails with a bunch of 
 ancient errors:
 ```
 .dub/packages/med-1.0.0/med/src/med/buffer.d(254,16): Error: instead of
C-style 
 syntax, use D-style `dchar[6 + 1] b
 ```
 
 meanwhile `dub run med ~master --compiler=gdc --arch=x86` pulls the latest
code 
 but can't build due to -Werror and some "statement is not reachable" warnings.
 
 When running it, it pegs a CPU at 100% because it's constantly running 
 ttkeysininput(). That's enough to cause a fan to run at high speed on this 
 laptop. But, that's easily fixed with select() or similar against stdin.


Thanks. I'll see what I can do. I've never heard of select().
Sep 12 2021
prev sibling next sibling parent reply Walter Bright <newshound2 digitalmars.com> writes:
On 9/11/2021 4:04 PM, jfondren wrote:

 When running it, it pegs a CPU at 100% because it's constantly running 
 ttkeysininput(). That's enough to cause a fan to run at high speed on this 
 laptop. But, that's easily fixed with select() or similar against stdin.


I'm not too familiar with I/O on Linux. If you could modify ttkeysininput() to 
use select(), I'm interested to learn how it's done. Thanks!
Sep 12 2021
next sibling parent reply Walter Bright <newshound2 digitalmars.com> writes:
On 9/12/2021 2:23 AM, Walter Bright wrote:

 On 9/11/2021 4:04 PM, jfondren wrote:

 When running it, it pegs a CPU at 100% because it's constantly running 
 ttkeysininput(). That's enough to cause a fan to run at high speed on this 
 laptop. But, that's easily fixed with select() or similar against stdin.


 
 I'm not too familiar with I/O on Linux. If you could modify ttkeysininput() to 
 use select(), I'm interested to learn how it's done. Thanks!


I'm looking at:

  #include <stdio.h>
  #include <sys/select.h>

  int main(void) {
     fd_set s_rd, s_wr, s_ex;
     FD_ZERO(&s_rd);
     FD_ZERO(&s_wr);
     FD_ZERO(&s_ex);
     FD_SET(fileno(stdin), &s_rd);
     select(fileno(stdin)+1, &s_rd, &s_wr, &s_ex, NULL);
     return 0;
  }

from

 
https://stackoverflow.com/questions/6418232/how-to-use-select-to-read-input-from-keyboard-in-c

Comparing it with the man page makes sense.

   https://man7.org/linux/man-pages/man2/select.2.html

I'll give it a try.
Sep 12 2021
parent reply Brian <bcallah openbsd.org> writes:
On Sunday, 12 September 2021 at 09:32:38 UTC, Walter Bright wrote:

 On 9/12/2021 2:23 AM, Walter Bright wrote:

 On 9/11/2021 4:04 PM, jfondren wrote:

 When running it, it pegs a CPU at 100% because it's 
 constantly running ttkeysininput(). That's enough to cause a 
 fan to run at high speed on this laptop. But, that's easily 
 fixed with select() or similar against stdin.


 
 I'm not too familiar with I/O on Linux. If you could modify 
 ttkeysininput() to use select(), I'm interested to learn how 
 it's done. Thanks!




Speaking of I/O: did you know that med will work just fine on 
OpenBSD (and every other Posix system, and probably macOS too) if 
all the version(linux) are replaced with version(Posix)? This is 
because med is using generic Unix I/O.

A PR doing the mechanical change: 
https://github.com/DigitalMars/med/pull/9
Sep 12 2021
parent Walter Bright <newshound2 digitalmars.com> writes:
On 9/12/2021 5:14 AM, Brian wrote:

 A PR doing the mechanical change: https://github.com/DigitalMars/med/pull/9


Pulled. Thanks!
Sep 13 2021
prev sibling parent reply jfondren <julian.fondren gmail.com> writes:
On Sunday, 12 September 2021 at 09:23:06 UTC, Walter Bright wrote:

 On 9/11/2021 4:04 PM, jfondren wrote:

 When running it, it pegs a CPU at 100% because it's constantly 
 running ttkeysininput(). That's enough to cause a fan to run 
 at high speed on this laptop. But, that's easily fixed with 
 select() or similar against stdin.


 I'm not too familiar with I/O on Linux. If you could modify 
 ttkeysininput() to use select(), I'm interested to learn how 
 it's done. Thanks!


I went with adding a tt* function:

```d
/**************************
  * Return when there are unread chars ready in the input.
  */

void ttwaitkeys()
{
     import core.sys.posix.sys.select;
     fd_set readfds;
     FD_ZERO(&readfds);
     FD_SET(0, &readfds);
     select(1, &readfds, null, null, null);
}
```

and running that in the while loop at the outset of getkey():

```
     ttyield();
     while (hasmouse && !ttkeysininput())
     {   c = mouse_command();
         if (c)
             return c;
         ttyield();
         ttwaitkeys();
     }
     c = term.t_getchar();
```

This fixes the CPU use and results in normal key input. But, it 
can just be replaced with a 'break;' in the same place of that 
loop. And it should break mouse handling.
Sep 12 2021
parent Walter Bright <newshound2 digitalmars.com> writes:
On 9/12/2021 12:38 PM, jfondren wrote:

 This fixes the CPU use and results in normal key input. But, it can just be 
 replaced with a 'break;' in the same place of that loop. And it should break 
 mouse handling.


I added this. It turns med from burning all the CPU to using negligible
amounts. 
The mouse still works, too! Thanks!
Sep 13 2021
prev sibling parent Walter Bright <newshound2 digitalmars.com> writes:
On 9/11/2021 4:04 PM, jfondren wrote:

 [...]


Should be good to go now, including adding a v1.1.0 tag.
Sep 13 2021