Man edits and more #274

N0NB · 2021-08-30T01:21:54Z

This is related to issue 272 (please do not close it yet!)

There are some script fixes and a simple implementation of Escape only stops TX on the first press which likely needs more refinement.

I'm requesting review and discussion.

Edit for recommended practices of formatting manual pages. Added formatting tips and documentation references at the top of the file. Completed up through the USAGE section.

Presumably this was a typo and should just be an alteration test for a local versus the file located in the user's PATH.

As installation places the soundlog script into ${prefix}/bin, avoid requiring the user to copy it to the ~/tlf/soundlogs directory. It works just fine from PATH when called from inside the directory.

The -w option on the rec commandline probably dates from the Open Sound System (OSS) version of rec. This option is not recognized by the SoX version of rec which is the only version available on modern Debian. This command line has been commented out so users can restore it easily if needed. Instead add an example command for Sox rec as the default. The -c 1 option forces single channel recording. This needs to be documented in the manual page so users can remove it as two channel recording is the default for this version of rec. Two channel recording may be useful for transceivers that output two channel audio, such as the Elecraft K3. The -q option makes the rec utility quiet, quelling its normal output.

The opendir() function is not in a context where it understands the value of $HOME, unlike the system() calls used elsewhere which operate in a user's shell environment. To make this work, query the environment for the value of HOME and concatenate it to the hard coded directory and pass that to opendir(). Add some error handling so if in the future the directory is exposed to user configuration, the user won't be completely in the dark if there is a typo in the entered path.

First try at only stopping TX when Escape is pressed the first time. This is very simplistic and likely will need more refinement for general deployment.

dl1jbe · 2021-08-30T05:55:55Z

src/audio.c

+		char *path = g_strdup_printf("%s%s",
+					     g_getenv("HOME"),
+					     "/tlf/soundlogs");
+


A point we should think about: Until now no other part of TLF requires a dedicated ~/tlf directory. Do we like to have a central one or keep our old tradition, that 'soundlog' should be a subdirectory of the actual contest directory in use?

zcsahok

Added some comments and a question.

zcsahok · 2021-08-30T17:09:32Z

src/main.c

@@ -258,6 +258,9 @@ int cwstart = 0;			/**< number characters after which
 int sending_call = 0;
 int early_started = 0;			/**< 1 if sending call started early,
 					   strlen(hiscall)>cwstart or 'space' */
+int escape_pressed = 0;	        	/**< 0. Escape not last key pressed.


I prefer using bool for actually boolean values. Also x = true or if (x) {../if (!x) {.. makes code a bit more readable.

Agreed. Originally I'd planned on the values 0-3 for counting all presses until all fields are cleared. Then I decided that isn't necessary.

As stated, this PR is mostly for review and discussion. I say cherry pick anything that's good and improve/discard the rest as needed.

zcsahok · 2021-08-30T17:12:26Z

src/callinput.c

@@ -147,6 +147,7 @@ int callinput(void) {
    int j, t, x = 0;
    char instring[2] = { '\0', '\0' };
    static int lastwindow;
+    extern int escape_pressed;


let's add globals to globalvars.h. This avoids having to list all used globals separately.

I think that should be useful.

zcsahok · 2021-08-30T17:22:41Z

src/callinput.c

@@ -796,9 +801,15 @@ int callinput(void) {
 	    // <Escape>, clear call input or stop sending.
 	    case ESCAPE: {
 		if (early_started == 0) {
+			if (escape_pressed == 0) {


Is the point of using escape_pressed is to stoptx() for the first ESC and cleanup() + clear_display() for the second?

Exactly.

I find the combination of stopping TX and clearing the exchange field to be rather irritating. Other programs I've used never did that, in fact they only used Escape to stop TX and another key/combination to clear the fields.

For CW this worked very well this past weekend. I gave up on stopping the voice keyer as per Tom and my discussion on the mailing list. I still think that should be implemented but there is likely a better way than the hack I came up with (not part of this PR).

OK, it took some time for me to figure out. :-)
How about putting it more explicitly, something like

if (x==ESC) { ++esc_count } else {esc_count=0} switch(x) { case ESC: if (esc_count==1) { stop_tx() } else if (esc_count>1) { clear... }

Would there be a way to avoid code duplication in callinput and getexechange? I mean some keys should behave the same, so it could warrant a common code. Not sure how to do it best.

That looks clean to me.

Sometimes there are subtle differences depending on whether a key is input in the call input or exchange field. I suppose that can be worked out but it would be a big refactor and I'm not sure what the gain would be.

Factor out the handling of common keys is something we should go for.

For the keys which are handled differently we could even add a state variable telling us the actual input field we are in.

Also provide commented play command for OSS.

N0NB · 2021-09-09T12:46:27Z

Tom, please merge all of the commits except f9be31b, especially the two for the manual page updates.

I have also modified doc/Makefile.am to distribute station-sample.cbr.

N0NB · 2021-09-09T12:46:58Z

My apologies for mixing PR topics.

dl1jbe · 2021-09-13T18:14:52Z

Seems mail mail from Saturday did not made it as a reply :-(, sorry. Will pick up the commit shortly. I think it is ok to merge related commits (e.g. the man page updates).

dl1jbe · 2021-09-13T18:29:10Z

A question came up for ef0696c: The change does not simply allow having 'soundlog' in PATH - instead it is now REQUIRED to be in PATH.

Before I commit it - do we want to keep it so?

N0NB · 2021-09-13T21:31:44Z

Tom,
That could be changed in the same way play_vk is handled with a test to see if it is in the working directory and then use the installed version if not.

dl1jbe · 2021-09-14T10:58:33Z

Ok, I'll pick up the actual solution and take the extension on the todo list.

dl1jbe · 2021-09-14T14:51:10Z

Cherry-picked, squashed and applied all commits (besides f9be31b) as requested.

dl1jbe · 2021-09-21T19:25:05Z

Hi Nate, while working on the $HOME problem did you have success with playing the recorded sound log files?
It seems that the command used in the code are also from old OSS play command.

The whole sound handling is in a bad shape. Some parts use hardcoded script names, some use hard coded command strings and some use external command given by logcfg.dat. We should choose a common mode for determining the play and record commands to use.

What makes it more demanding is that we need to record from different sources (microphone or rig) and also have to play to different sinks (radio or speaker). Any idea how to specify that in SoX?

N0NB · 2021-09-21T21:09:36Z

Hi Nate, while working on the $HOME problem did you have success with playing the recorded sound log files?
It seems that the command used in the code are also from old OSS play command.

Once I changed things in the code, yes, it worked. I don't think I pushed that stuff out, though.

The whole sound handling is in a bad shape. Some parts use hardcoded script names, some use hard coded command strings and some use external command given by logcfg.dat. We should choose a common mode for determining the play and record commands to use.

I would think that opting for a default that can be overridden by a setting in logcfg.dat would be preferable. Let the commands default to the SoX utilities and the user can change them as needed.

What makes it more demanding is that we need to record from different sources (microphone or rig) and also have to play to different sinks (radio or speaker). Any idea how to specify that in SoX?

That, I don't know. As I run Gnome and PulseAudio and run Tlf in an Xterm, I use the pavucontrol GTK+ utility to direct the sound streams as I need. There are probably ways to write a configuration file for PA and then there will be the cases where PA is not used and the user will need to rely on the ALSA configuration.

dl1jbe · 2021-09-22T08:57:44Z

Hi Nate, while working on the $HOME problem did you have success with playing the recorded sound log files?
It seems that the command used in the code are also from old OSS play command.

Once I changed things in the code, yes, it worked. I don't think I pushed that stuff out, though.

Ok, I will look into it.

The whole sound handling is in a bad shape. Some parts use hardcoded script names, some use hard coded command strings and some use external command given by logcfg.dat. We should choose a common mode for determining the play and record commands to use.

I would think that opting for a default that can be overridden by a setting in logcfg.dat would be preferable. Let the commands default to the SoX utilities and the user can change them as needed.

Same preference here. I will go that way.

What makes it more demanding is that we need to record from different sources (microphone or rig) and also have to play to different sinks (radio or speaker). Any idea how to specify that in SoX?

That, I don't know. As I run Gnome and PulseAudio and run Tlf in an Xterm, I use the pavucontrol GTK+ utility to direct the sound streams as I need. There are probably ways to write a configuration file for PA and then there will be the cases where PA is not used and the user will need to rely on the ALSA configuration.

I did some more research and I found some ways to use different sound devices at the same time. Just have
to make some tests which way to go best.

dl1jbe · 2021-09-23T18:17:03Z

Just found how to control the device for recording and/playing sound.
You can stat it on SoX command line:

sox <file> -t alsa hw:1,0 or
sox <file> -t pulseaudio devicename........

will do. I will finally add these to the man page.

zcsahok · 2021-11-20T09:47:40Z

tlf.1.in

+.B QSO
+or
+.B DXPED
+modes).
 .
 .TP
 .B IGNOREDUPES


Just came across this: it's spelled in singular, IGNOREDUPE.

Good catch. Fixed.

zcsahok · 2022-02-26T19:09:28Z

Could this be merged? Master is drifting away...

dl1jbe · 2022-02-26T21:04:31Z

Could this be merged? Master is drifting away...

Accept the commit mentioned by Nate

Tom, please merge all of the commits except f9be31b, especially the two for the manual page updates.

all got merged in already in commit e26dfde in September last year.

I kept the PR open to kept the discussion about the audio recording around until PR #276 got finished. That one is overdue and I had planned to complete it in next weeks after finishing the move to use qso_array instead of qsos[] (PR #305).

So please let us wait with closing the PR here for some days longer.

N0NB · 2022-02-26T23:55:16Z

* On 2021 30 Aug 05:49 -0500, Thomas Beierlein wrote:

@dl1jbe commented on this pull request. > - if (sounddir == NULL) + /* Must query the environment for the value of $HOME + * and build the path to the soundlogs. + */ + char *path = g_strdup_printf("%s%s", + g_getenv("HOME"), + "/tlf/soundlogs"); + A point we should think about: Until now no other part of TLF requires a dedicated ~/tlf directory. Do we like to have a central one or keep our old tradition, that 'soundlog' should be a subdirectory of the actual contest directory in use?

I think that making soundlog a subdirectory of the working directory is a good idea. Such audio logs should be contest/event specific and putting them all together makes it easy to back them up, etc. As an example, in my home directory is a subdirectory named 'logs'. Under that I have directories for each event, e.g. 'KSQP-2021', etc., under each of those are all of the files/directory(ies) for an event. 73, Nate P.S. I guess I did not catch this when it first hit my email box almost six months ago!

…

-- "The optimist proclaims that we live in the best of all possible worlds. The pessimist fears this is true." Web: https://www.n0nb.us Projects: https://github.com/N0NB GPG fingerprint: 82D6 4F6B 0E67 CD41 F689 BBA6 FB2C 5130 D55A 8819

dl1jbe · 2022-02-27T08:48:51Z

Nate, there is an unfinished branch stop_tx in PR #276 which I plan to complete next.

In it I made the soundlog directory configurable by keyword in logcfg.dat. At the moment the default is '~/tlf/soundlogs' but I agree that just './soundlogs' would be better. That can be easily done.

N0NB · 2022-02-27T12:14:33Z

* On 2022 27 Feb 02:49 -0600, Thomas Beierlein wrote:

Nate, there is an unfinished branch stop_tx in PR #276 which I plan to complete next.

I noticed that things were still as before when I did some of the CQ 160 last month. I appreciate the careful evaluation of features.

In it I made the soundlog directory configurable by keyword in logcfg.dat. At the moment the default is '~/tlf/soundlogs' but I agree that just './soundlogs' would be better. That can be easily done.

Good. I'd not thought of that directory before and I think yours is a reasonable approach. 73, Nate

…

-- "The optimist proclaims that we live in the best of all possible worlds. The pessimist fears this is true." Web: https://www.n0nb.us Projects: https://github.com/N0NB GPG fingerprint: 82D6 4F6B 0E67 CD41 F689 BBA6 FB2C 5130 D55A 8819

zcsahok · 2022-02-27T16:00:20Z

I was not closely following this PR, sorry.
A local dir ('./soundlog') would fit surely better.

dl1jbe · 2022-03-04T18:21:42Z

With PR #317 and #318 in the pipe we can close these PR now. Thanks for all the discussion.

N0NB added 8 commits August 21, 2021 20:54

Add *.swp for Vim users.

31f1bc1

Manual page formatting and minor wording edits

a65a51d

Edit for recommended practices of formatting manual pages. Added formatting tips and documentation references at the top of the file. Completed up through the USAGE section.

Correct typo that causes a shell error

6a9d601

Presumably this was a typo and should just be an alteration test for a local versus the file located in the user's PATH.

Reword Sound menu item 3 slightly

be6cd13

Allow for soundlog script in PATH

ef0696c

As installation places the soundlog script into ${prefix}/bin, avoid requiring the user to copy it to the ~/tlf/soundlogs directory. It works just fine from PATH when called from inside the directory.

Implement first press of Escape stops TX

f9be31b

First try at only stopping TX when Escape is pressed the first time. This is very simplistic and likely will need more refinement for general deployment.

N0NB requested review from dl1jbe and zcsahok August 30, 2021 01:21

N0NB linked an issue Aug 30, 2021 that may be closed by this pull request

Rework sound recorder and playback. #272

Closed

Merge branch 'master' into man_edits

ee873b3

dl1jbe reviewed Aug 30, 2021

View reviewed changes

zcsahok reviewed Aug 30, 2021

View reviewed changes

N0NB added 3 commits September 9, 2021 07:35

Complete manpage updates and slight reorganization

3b97a0b

Make SoX play the default

af23403

Also provide commented play command for OSS.

Distribute station-sample.cbr

13e8209

zcsahok reviewed Nov 20, 2021

View reviewed changes

dl1jbe closed this Mar 4, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Man edits and more #274

Man edits and more #274

N0NB commented Aug 30, 2021

dl1jbe Aug 30, 2021

zcsahok left a comment

zcsahok Aug 30, 2021

N0NB Aug 30, 2021

zcsahok Aug 30, 2021

N0NB Aug 30, 2021 •

edited

Loading

zcsahok Aug 30, 2021

N0NB Aug 30, 2021 •

edited

Loading

zcsahok Aug 31, 2021

N0NB Aug 31, 2021

dl1jbe Sep 3, 2021

N0NB commented Sep 9, 2021

N0NB commented Sep 9, 2021

dl1jbe commented Sep 13, 2021

dl1jbe commented Sep 13, 2021

N0NB commented Sep 13, 2021

dl1jbe commented Sep 14, 2021

dl1jbe commented Sep 14, 2021

dl1jbe commented Sep 21, 2021

N0NB commented Sep 21, 2021

dl1jbe commented Sep 22, 2021

dl1jbe commented Sep 23, 2021

zcsahok Nov 20, 2021

dl1jbe Nov 21, 2021

zcsahok commented Feb 26, 2022

dl1jbe commented Feb 26, 2022

N0NB commented Feb 26, 2022 via email

dl1jbe commented Feb 27, 2022

N0NB commented Feb 27, 2022 via email

zcsahok commented Feb 27, 2022

dl1jbe commented Mar 4, 2022

Man edits and more #274

Man edits and more #274

Conversation

N0NB commented Aug 30, 2021

Choose a reason for hiding this comment

zcsahok left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

N0NB Aug 30, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

N0NB Aug 30, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

N0NB commented Sep 9, 2021

N0NB commented Sep 9, 2021

dl1jbe commented Sep 13, 2021

dl1jbe commented Sep 13, 2021

N0NB commented Sep 13, 2021

dl1jbe commented Sep 14, 2021

dl1jbe commented Sep 14, 2021

dl1jbe commented Sep 21, 2021

N0NB commented Sep 21, 2021

dl1jbe commented Sep 22, 2021

dl1jbe commented Sep 23, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

zcsahok commented Feb 26, 2022

dl1jbe commented Feb 26, 2022

N0NB commented Feb 26, 2022 via email

dl1jbe commented Feb 27, 2022

N0NB commented Feb 27, 2022 via email

zcsahok commented Feb 27, 2022

dl1jbe commented Mar 4, 2022

N0NB Aug 30, 2021 •

edited

Loading

N0NB Aug 30, 2021 •

edited

Loading