pinacolada64 / imagebbs Goto Github PK

Printed out +.IM and +.NM/config and evaluated. Working on renumbering and converting +.NM/config to IM module format.

Hopefully this is a gentle introduction to issue tracking for @x-tec2017.

There are a few issues with the idle screen variable display module +/lo.varlist.

• Displays string variable name instead of string contents:
  

• Goes into infinite loop when displaying array variable contents:
  

It could be advantageous to already know the Image device/drive assignments, that way you don't have to spend time copying/moving files from within BASIC.

+.MU (the message move utility, pity it doesn't work on 1.x) will load the message base directory, let me "get" a message to move, but then I can't "drop" it in the second message base I defined.

Workaround: Search for letter before desired search result.

It would be nice not to have to manually set all the flags on the sysop account when you're first setting up your BBS.

We need a better port of this. The 8/3/89 version for 1.2 allowed you to overwrite if exist, and use alternate filenames at the time of saving. It would also be useful to give the user a choice to stay in editor or return to main.
Also, using WF to create a file doesn't work, since it doesn't ask for filename on .save

Some massaging of source code, and there's a halfway intelligible display. Not sure what it's all supposed to mean, but it could be an interesting exercise to fix.

Not sure what is causing it, but e.access seems to constantly add lines to the file until eventually it is completely overwritten.

Need an updated user editor that can access the new flags and stats portions of the 2.0 user files.
We don't have access to change the user's terminal parameters, or maintenance flags.
Another issue with this version is single digit entry for "Downloads per day" field.

If UU is the suggested editor, it's broken. See #29.

From @x-tec2017:

NMI is not getting disabled during serial port operation. If a program is loading and something comes on the SwiftLink, the program load will either get stuck and the drive activity light stays on, or if it manages to complete the load, the program that loaded will be corrupted. Try making the BBS load a large +.file like +.IM or +.EM and start hitting the Return key rapidly while logged on through the SwiftLink. That should duplicate the problem. The file rs232/swift.s has the routines for disabling and enabling NMI.

I didn't discover this until I went to add the changed lines in +/IM.access. I have the i/IM.access from the December set, but I'm not sure what has been changed since then.

After typing HELP, the message is displayed, followed by the prompt to create a temporary account. On answering NO, the time remaining is reduced to 0. Recommend rewriting 3188 to ifa=.then3136
If anything, I would offer raising tr by 2 minutes since they probably read the message slowly.

--original lines--
3186 lp=1:&"‹‹˜Would You Like To Log On With A œTemporary˜ Account? ":gosub1902
3188 ifa=.thentr%=.:goto3136

This didn't start happening until I set the "lines per screen" setting for sysop user. I wonder if this was somehow responsible for issue #15 where we got unwanted prompts.
Suggest setting a temporary lines/screen variable until the user profile is loaded.
See screenshot:

This is size from the Christmas 1994 NISSA disk. Try running it multiple times (sys 52992) with the same program in memory. The block size keeps increasing (one less than the total blocks, it seems) between runs. The block size should be initialized to zero between runs.

Add more user flag configuration options:

• Graphic Menu mode

• Expert mode

• Change ld$ to d1$
  

  ImageBBS/v2/core/plus_new.lbl

  Line 13 in 0ca67ac

  jn$="":l=l1:ac%=.:id=.:cr=nc:f$=cc$:ld$="1"+"8001011201":p$=""

Started after the e.idle/e.access issues started cropping up.
Went back to zero day mods and fresh data (3 users), problem still persists.

On the R2 disk (that I'm using) there were a few plus files missing from the distro. We either need to update the distro to include the missing files, or the existing ECS data on the distro should be modified to have them removed from the list.

+.TU
+.CD
+.word
+.NM/chess
+.OR
+.NW
+.TAG

Is this by design that the data field only allows one character? I can't think of a reason to give users less than 10 downloads per call. If that's the case, I'll just give them unlimited downloads and modify the credit ratio to allow 20-30 for most generic access levels.
Subops and Maint users get unlimited anyway. At least that's how I roll.
Otherwise, fix field 20 input to allow more than one character.

When configuring a new BBS, the current process has some deficiencies:

• Mess up an answer, there's currently no way to go back and correct it without rebooting
• (Mention in new version "You can change answers later.")
• IM menu system could be used to answer these questions

Hardware:

• Using Lt.Kernal device?
• Configure modem/telnet bridge options (nimodem.dat)
• Dallas Semiconductor DS12C887 RTC chip at $d500 (VICE)

Sysop config:

• Graphic Menu mode available, enabled? lightbar options
• Ask which style of menus the sysop prefers: text or graphic menu
• Email vs. phone # mod of X-Tec's is something to consider, too
• Configure BBS time zone
• Configure Prime Time
• Add user 1 to QuickList (instant mode login)

The beginning of line 3484 of +.lo should be changed
from:
3484 on1goto3490:
to:
3484 onigoto3490:

When editing a sysop's flags, there is no way to exit the menu. Options exist for flags 1-8, but neither [CR], Q, or backarrow will get you out.
I had to break and goto1100 just to get back to the previous module.

Not sure if it's just me, but here's how to recreate:

R from main prompt, type in a file that is not there, or select a drive that does not exist.
Results in break: load error in 2006

That said, something got really screwy with my etc files somehow. It got to where I could not load menus anymore, and then various modules started misbehaving. I noticed my e.Sig and e.Sub files were 121 blocks each, and going into SB would eventually result in bad subscript and other things.

I restored the etc drive from a week old backup, and now I can see menus again, and the subs work as expected (after I used the updated consig files Larry sent).
Just a heads up, and I don't know the root cause yet. I'm going to recreate what I did this week to make it happen again.

Additions for 1.3 manual:

• Additional & commands (and examples)
• Memory map with locations in decimal and hexadecimal
• Additional module types

I am working on an idea to add a sub-module to IM. Moving the "Remote Time Editor" into its own sub-module will knock a few blocks off IM and add two features to the `Time Set" menu:

• Set prime time start/end times, toggle prime time
• Change clock set method/external device number

Maybe the menu item could be renamed to "Time Functions" or something.

All issues fixed plus some cosmetic changes. Updated +.CP uploaded to the Image 2.0 support boards.

e.zz log (local maint file) and e.nu log (blacklist logon file) aren't created/read from anywhere that I see. New routines will have to be written if we want these features.

e.em log (mail sent file) is created by im.misc2 and +.EM.

Is it just me, or does > and < not switch between sub/boards like UD and UX?

This isn't particularly desirable from a security standpoint.

Issue 1: Ignores the menu mode. After selecting the user, the data is displayed in cursor menu mode.

Issue 2: No data can be selected
When selecting (either by cursor/return or letter) any item, the screen clears twice, then returns to view mode. No entry prompts appear, and no records are changed.

In non-cursor menu, the commands for editing and adding are broken. Z enters maint menu, but the only command that works in there is O for logoff.
I was able to add and view libraries using the cursor menu however.

This is a 2 part problem when the More Prompt is set to Messages Only

1. Doesn't pause messages in the Sub Boards.
2. Entering MP causes "Illegal Quantity Error in 1870"

Not sure where this is happening. Run +.alpha/ind to fix.

Writing data skewed by one record.
On the screen when editing a login instance, it looks correct, but after saving, if you go back and look at the changes, the data is skewed. Example. add 3 modules, and configure:

# name acs
1. news 1023
2. last 1023
3. grf 1023
Save the changes, then go back and edit and it will look like

# name acs
1. news 1023 0
2. last 1023 0
3. ^

The content of e.loginmods is
3
last 1023
grf 1023
news 1023

Before I edited it, the content (as is on the R2 disk set)
2
last
1023
news
1023

Update: it appears to be the write routine in +/IM.logon, because all 3 areas are affected in this manner.
Maint mods, login mods, and quicklist. Haven't tested wall ads, but I imagine it would be the same.
In my humble opinion, I think just adding another print#2 statement in between the variables will cure it.
Change 4040 fori=1tobr:print#2,tt$(i)r$a%(i):next:close2:return
to 4040 fori=1tobr:print#2,tt$(i):print#2,a%(i):next:close2:return
I'm guessing someone is using r$ as a soft carriage return but doesn't get translated that way when writing the file for some reason. Either that, or r$ was never loaded with the chr$(13) during the startup routine. Hell, perhaps the second print#2 statement can be mitigated by adding chr$(13) in it. Either way, that's beyond my pay grade.

** UPDATE (again)
I changed my +/IM.logon line 4040 to read
4040 fori=1tobr:print#2,tt$(i)chr$(13)a%(i):next:close2:return
and it worked like it should. I'm thinking someone just forgot to set r$=chr$(13) during init. I went further to just declare it in the master file +.IM at line 3000 with r$=chr$(13) and it cured all 3 of the areas mentioned. I know we're all about saving memory space, but cmon. :)

In an area where dialog menus (yellow overlays) are prominent (i.e.: GF maint or options menus) the only option is to use the letter/number item to access the options. The cursor appears at the bottom of the screen, and no marker seems to be present anywhere else on the screen.

Pressing Ctrl + F1-F7 spit out garbage upon first press, second press then jams VICE @ $dc05

I love this BBS software, but wow this source tree is terrible. This isn't meant as a personal attack or anything but as someone who likes to be able to just build my programs without having to fuck around in the makefiles too much, or fix errors, this is awful

I'm focused on the V2 tree here, (in master, NOT the v2 Branch)

First off you're relying on using wine on specific beta versions of freeware non-open source software (C64list), i recommend not doing that as a future update could fuck shit up, and since we don't have the source, it can't be unfucked. Not to mention they're distributing the binaries through links of facebook, ew.

Second of all, building. I believe the makefile is incomplete, but here are some pointers. I understand Makefile IS a major pain in the arse to work with, but it makes things easier for the user.

First off, there seem to be minor syntax errors in:
v2\asm\swap3.asm on line 112, thee's a missing space, it should be "ldy #0"
v2\asm\irqhn.asm on line 315 (I think just changing the #',' to #$2c shoould be equivelent?

Now, by using the latest version (4.03 at the moment) of C64 list it seems to build okayish, but not all the way into d64s or a d81.

I get ecs.bin garbage-collect.bin swap1.bin swap3.bin
editor.bin struct.bin swap2.bin wedge.bin

I haven't gone any further yet, but i'm hoping this is at least what i need to put into a disk with different names to get something to work?

Sorry if i sound super critical, i get this is tough work, but i do think it's important to at least get the project building or at least offer the latest working binaries in the Releases section so people can actually access the software easier without hunting down binaries elsewhere

Linefeeds don't reset to default at login screen.
Steps to recreate:

1. Connect via ASCII or ANSI terminal that would normally need LFs
2. Attempt to login as a user with no linefeeds (C64 user)
3. Fail authentication with bad password
   at this point, the setting from that user's lf is used
4. Return to login prompt and attempt login as anyone else
   settings (not sure how many) appear to be in effect for the attempted failed user

Would clear arrays be appropriate here? At this stage of login, we've already established baud, translation, and graphics status.

Probably more of a feature request, but I don't have access to write into that area. I'm pretty sure the C-Net 128 software has Ymodem, and I wonder if it can be ported over for our purposes. Since so many people are running emulators to call and download from us, it seems almost a shame to limit them to Xmodem only.

From @xtec2017:

? I'm stuck on something that maybe you can help with in Image 2.0.
Go into Image Term by holding down the CONTROL key and hitting the left arrow.
Select 3) Term Parameters
Select 3) U/D Protocol
The system gets stuck in line 2500

I've traced it this far:
in +.t
3210 gets an$' '3210 gosub3006' ' 3006 converts an$ to a (a=3)' '3212 on(3)goto3228' '3228 gosub1666' ' 1666 sets im=4: goto2406' ' 2406 sets a$=params:goto2500' ' 2500 gosub2510' ' 2510 loads "im.params"' ' 2512 return' ' 2500 gets stuck here'

it should continue in line 2500 with is=is+1:im$(is)=im$ then fall to 2502
2502 gosub60000:new60000:is=is-1:if is then a$=im$(is):gosub 2510 2504 return