I've noticed various helpfiles that are incorrect in Smaug and I figure we should fix them for FUSS. They can cause issues such as I thought the ifcheck ispassage was not working due to the helpfile was wrong. This was the helpfile:

Turns out you don't use door numbers, you use door keywords. So it needs to be this:


Anyways, I think this topic should be a place to post incorrect helpfiles. So here are some more I can remember.

HELP MPCLOSEPASSAGE

You have to use door numbers, not door keywords. So add the line:


<Post will be edited over time>

Maybe it might be a good idea to choose a consistent path on things like this. Either use direction names, or use numbers, but don't keep switching back and forth. I personally prefer names since they're much easier to remember. Maybe the MPCLOSEPASSAGE one would be better suited as a code change?

I'd say so. It'd be easier for everyone unless someone can tell me a reason why the numbers should be used over actual exit names.

Oh and according to help doordirs:

Those exits were finished, right?

So far as I know they should have been finished up long ago. There are likely some helpfiles in there that are outdated and could stand to be cleaned up or even removed.

Also there are many helpfiles that have bad keywords, such as "2.help" and so on that have no help desc. These would cause the similar helpfile snippet to crash the mud. I've removed most/all of these "bad keyword" helpfiles so I cannot recall any right now. Just letting everyone know there are quite a few.

Off the top of my head I can recall a few of these that gave me trouble when I put in Senir's similar helpfiles snippet. I've since been able to fix a few loopholes with it (with help from others, of course) but anyway, here's what I recall:

"blackhands" "immauth" (dupe entry) "animate dead" (dupe entry) "names" (dupe entry)

Also if you had two ' instead of a single " as the start of the helpfile (one of the two IMM AUTH's had this) it caused a crash.

Further, the "acid blast" / "burning hands" helpfile is somehow improperly formatted and displays rather badly with the similar helpfiles snippet. I think it might have actually displayed bad all along, but I'm not sure. Anyway, it's all one help entry for several spells so you can't miss it. I think it has some invalid characters or something in it perhaps, but I don't recall exactly.

Hope someone finds this useful.

Maybe someone should ask the writer of the simular help files snippet to allow it to become standard in the release?

In all honesty, no offense to Senir, but I find Remcon's version of it much more friendly: Remcon's helpfile code

It performs the same function and has not been shown to be unstable. He has allowed me to include it in AFKMud, and he uses it on his own Smaug deriv. The code doesn't seem to have received nearly the attention it deserves for some reason.

However, it is not technically considered a bugfix and there just isn't the kind of demand behind this feature like there was to get hotboot added as a standard.

Hmm. I believe there was an issue with part of the original code, not just the similar helpfiles snippet, that led to the problems encountered when the snippet searched through the helpfiles to find similar entries.

Basically the change was here (act_info.c, get_help function):

if ( isdigit( argument[0] ) && !is_number(argument))

I believe the !is_number(argument) was the addition.

Though I don't fully recall and didn't read through the whole thread again, I think the issue was level requests being confused with numbers in the helpfile names somehow. This particular fix was supposed to prevent helpfile names like "123" from causing a loop because the name was being taken as a helpfile level.

Anyway, I thought this information might be useful. I would think this could apply to both snippets, but I've not looked at Remcon's so I'm not sure if it would cause a problem there. In any event, hope the info is useful to someone.

Check out my topic for what the gdb loops were:
http://www.gammon.com.au/forum/bbshowpost.php?bbsubject_id=5042

#0  get_help (ch=0x94374d0, argument=0xfed04570 "2" at act_info.c:2042
2042        if ( isdigit(argument[0]) )
(gdb) bt
#0  get_help (ch=0x94374d0, argument=0xfed04570 "2" at act_info.c:2042
#1  0x0805489f in do_help (ch=0x94374d0, argument=0xfed04570 "2"
    at act_info.c:2277
#2  0x08054b2f in do_help (ch=0x94374d0, argument=0xfed075b0 "2"
    at act_info.c:2329
#3  0x08054b2f in do_help (ch=0x94374d0, argument=0xfed0a5f0 "2"
    at act_info.c:2329

I believe this kind of looping was part of how the above fix came about in the dicsussion on MudPlanet. Have you checked out that old thread yet? Since the fix isn't part of the snippet but part of the base code itself, it would probably happen regardless of whose snippet you used when it ran into requests for helpfiles with numbers in them and mistakenly thought you wanted a helpfile at that level, or something to that effect (still fuzzy on all the details, was a while ago).

All of those discussions pertained to Senir's code and not Remcon's. Thus far on our own mud we have not been able to make Remcon's code crash anything, even when sending numerical arguments.

Eh? In the topic I linked to (Gammon) I did try with Remcon's and it did crash.

This is true, but I changed the get_help function to add that !is_number check, which is not part of either snippet. I'm not sure how this would have affected Remcon's snippet since I had Senir's in and this was to fix the bug with this (have since upgraded to Remcon's. Seems cleaner and I rather like the way the output looks too).

Still, as it's part of the SMAUG code and Zeno reports some issues with this particular chunk of code in get_help anyway (presumably in conjunction with Remcon's snippet, right?), I thought this might be relevant information.

If not, my apologies, just trying to pass along what I thought might be useful. Also, thanks for pointing me toward Remcon's snippet.

Yes, I know what was reported, and just like back then, I am not able to duplicate the scenario now either. I understand it may have something to do with bad helps but FUSS still contains those and my test copy for snippet checking does not exhibit this crash behaviour. I'm not saying he's not having an issue, but it might be the result of other things and not the get_help function.

I have also gone and dug up the old thread on MUD Planet and am not able to duplicate the scenario their either where typing with the trailing single quote would crash the code.

I thought I'd try my hand at this and tried a variety of helpfile entries to reproduce the original problems with numeric help entries like "123 crash" or what not, couldn't get it to crash no matter what I did, regardless of the level of the person trying to access the help and the level of the help itself. This is, again, using Remcon's code now.

Still curious about whether or not there are holes in this code though. If anyone has some specific things that are causing crashing, I'd certainly be willing to plug them in and see if I can reproduce it. Whatever I can do to help.

Not to sound rude or whatever but you basically just made my point

The crashes experienced were with Senir's code, not Remcon's, that's all I was trying to get across. As you have just noticed, his code doesn't barf on the same test conditions. Doesn't mean it's flawless, but until someone finds data it chokes on and it can be verified elsewhere, I have to assume the bug is not in Smaug but in Senir's code.

I'll put in Remcon's code in stock FUSS and test it out, see what I can come up with. Tomorrow perhaps, after I get out of my new job.

Oh I understood that the first time you said it.

I had upgraded to Remcon's snippet when you mentioned it earlier in the thread. After that, I decided to try and reproduce the kinds of problems I had before and that Zeno said he was having, presumably with Remcon's snippet since that's apparently what he's having problems with at the moment. I understand he had Senir's in and now he's changed to Remcons and still having the problem(s). If I'm wrong here, please correct me.

So yes, I did confirm what you said, but not because I didn't understand what you were saying.

Like I said before, only trying to help figure out what the problem may be. To that end, I'm still up for helping to try and reproduce / verify things if anyone comes up with something.

I looked at both persons code and saw that Remcon's snippet might have been made after Senir's and thus had the bugs chinked out. I am debating using Remcon's snippet now that I see how it is progressed.

Thank you for all your insights.