enhance internal help

[jelmd]

wtmpdb should emit command relevant options and operands, only. Also for convenience all sub commands should accept the -h and -v option. Last but not least because all sub commands already accept the -f file option it should be displayed as a generic aka common option as well.

[thkukuk]

-f is no common option as it does not work without any command.
Global options are the ones you can use directly without any command. That's common wording for many tools, we should not derivative here, this will only confuse people.

About -v and -h for all sub commands: -h makes sense if you display only the options for this command, but -v does not.

[jelmd]

-f is no common option as it does not work without any command.

Can you elaborate on that or at least show an example that supports this statement?

Global options are the ones you can use directly without any command.

OK. But users reading this would also expect: which you can use with any commands.

That's common wording for many tools, we should not derivative here, this will only confuse people.

Actually global means worldwide available/accessible and does not depend on which place of the sphere/world/country/code the caller lives. So one could call them common options, too - doesn't make a real difference.

About -v and -h for all sub commands: -h makes sense if you display only the options for this command, but -v does not.

As said: for convenience (i.e. making its usage comfortable for the customer alias make them happy) it makes a lot of sense, because one just needs to add -v to the command previously executed or just being in the CLI buffer - no need to navigate to the right place and remove a word (which is really annoying). Last but not least: why does the customer deserve getting tortured with (bothered to read) options which do not apply anyway? And similar, why should he get tortured to read endless long man pages which just tell, what he already knows: no wtmpdb command which does anything [useful] except showing help and version requires the db (and therefore allows -f ...). Did I miss something?

[jelmd]

Just forgot another important thing since wtmpdb has the potential to replace related legacy tools like last, lastlog, etc. and is already supported wrt. last: one would need to file another bug/RFE: last should provide an option to show usage help and version info.

[jelmd]

see #45

[thkukuk]

Just forgot another important thing since wtmpdb has the potential to replace related legacy tools like last, lastlog, etc. and is already supported wrt. last: one would need to file another bug/RFE: last should provide an option to show usage help and version info.

wtmpdb is no lastlog replacement and will never be.

[thkukuk]

-f is no common option as it does not work without any command.

Can you elaborate on that or at least show an example that supports this statement?

global option means:

wtmpdb -f file is a valid command and will not fail, but it will fail.

Global options are the ones you can use directly without any command.

OK. But users reading this would also expect: which you can use with any commands.

This are users who only rarley read manual pages of complex commands.

That's common wording for many tools, we should not derivative here, this will only confuse people.

Actually global means worldwide available/accessible and does not depend on which place of the sphere/world/country/code the caller lives. So one could call them common options, too - doesn't make a real difference.

Correct, worldwilde, so wtmpdb -f file needs to work, but it does not since it doesn't make any sense.
-f is no global option, and the global option section is not a comman option section.

About -v and -h for all sub commands: -h makes sense if you display only the options for this command, but -v does not.

As said: for convenience (i.e. making its usage comfortable for the customer alias make them happy) it makes a lot of sense, because one just needs to add -v to the command previously executed or just being in the CLI buffer - no need to navigate to the right place and remove a word (which is really annoying). Last but not least: why does the customer deserve getting tortured with (bothered to read) options which do not apply anyway? And similar, why should he get tortured to read endless long man pages which just tell, what he already knows: no wtmpdb command which does anything [useful] except showing help and version requires the db (and therefore allows -f ...). Did I miss something?

A crystal ball what might come in the future?
A good design following best practices is cucrial for further development. Adding hacks over hacks because some few people are very lazy is the best way to make a project unmaintainable.

[jelmd]

wtmpdb is no lastlog replacement and will never be.

Well, the wtmp database maintains the login/logout entries/log of the system and wtmpdb[lib] is the tool to maintain/query it for required information. So it makes sense, that it provides an easy way to show lastlog infos as well. Actually it is just a slightly modified representation (subset) of the last output/result.

[jelmd]

-f is no common option as it does not work without any command.

Can you elaborate on that or at least show an example that supports this statement?

global option means:

wtmpdb -f file is a valid command and will not fail, but it will fail.

You didn't test the PR, did you? The wtmpdb produced using the #45 does not fail.

Global options are the ones you can use directly without any command.

OK. But users reading this would also expect: which you can use with any commands.

This are users who only rarley read manual pages of complex commands.

Ultimately, it’s all about usability, efficiency, and consistent usage patterns.
E.g. if people had to read the Tesla manual before they could drive one, Tesla probably wouldn’t sell many ;-)
The reality is that most users avoid reading manuals — especially when they’re too long, cluttered with redundant details, or describe confusing and unintuitive behavior that just doesn’t make sense. So let’s make users happy by helping them work efficiently and effortlessly. Don't blame the users.

That's common wording for many tools, we should not derivative here, this will only confuse people.

Actually global means worldwide available/accessible and does not depend on which place of the sphere/world/country/code the caller lives. So one could call them common options, too - doesn't make a real difference.

Correct, worldwilde, so wtmpdb -f file needs to work, but it does not since it doesn't make any sense. -f is no global option, and the global option section is not a comman option section.

That’s not correct. It doesn’t apply to your case because you’re definitely talking about something else — not the wtmpdb produced by applying the PR. Try out the PR yourself, and you’ll see that it works fine and actually makes a real difference.

About -v and -h for all sub commands: -h makes sense if you display only the options for this command, but -v does not.

As said: for convenience (i.e. making its usage comfortable for the customer alias make them happy) it makes a lot of sense, because one just needs to add -v to the command previously executed or just being in the CLI buffer - no need to navigate to the right place and remove a word (which is really annoying). Last but not least: why does the customer deserve getting tortured with (bothered to read) options which do not apply anyway? And similar, why should he get tortured to read endless long man pages which just tell, what he already knows: no wtmpdb command which does anything [useful] except showing help and version requires the db (and therefore allows -f ...). Did I miss something?

A crystal ball what might come in the future? A good design following best practices is cucrial for further development. Adding hacks over hacks because some few people are very lazy is the best way to make a project unmaintainable.

Crystal ball? Not needed. What we have is logic and experience (30+ years of Linux/UNIX in my case). Maintaining the same information across multiple databases with different tools only creates inconsistencies, confusion, complexity, and copied bugs. No hacks here — just logic, well-thought-out decisions, and a DRY approach (efficiency, if you will).

If you think lastlog doesn’t belong in wtmpdb, consistency demands that last and the others be split into separate tools too. “Unmaintainable”? That’s a code-structure problem. Factor out each main_* function into its own file, move common functions to util.c, ... — standard C patterns, best practices ;-). Simple, clean, maintainable.