APDL Public Domain 1

home *** CD-ROM | disk | FTP | other *** search

/ APDL Public Domain 1 / APDL_PD1A.iso / comms / a_l / binkley / ChangesLog < prev next >

Wrap

Text File | 1993-10-28 | 63.4 KB | 1,249 lines

This is the Binkley Fidonet Mailer Changes Log. This should list all the modifications that the developers have done over the last few years. The developers, and their initials from which changes are identified, and their fidonet email addresses are as follows : Stephen Davies (SLD) [?????????????????????] Geoff Riley (GIR) [2:250/208.0 @ fidonet] Hugo Fiennes (HF) [2:252/102.0 @ fidonet] Steven Haslam (SRH) [2:250/116.3 @ fidonet] Keith Hall (KEH) [2:250/219.0 @ fidonet] This programmer's log is in reverse-chronological order (jeez!) meaning that the newest stuff is at the top and the oldest at the bottom. Entries with the same version number but different dates are internal versions; i.e. they were never released. a *R* by the developers initial indicates that version was a release version. ------------------------------------------------------------------------------- 27 Oct 1993: V2.16 *R* When sending ArcMail without aliases (xxxxxxxxM1 files), it will now substitute the correct day code (Su/Mo/Tu/We/Th/Fr/Sa). Last caller display made bigger so you can see the full address rather than just "Last: 2:250/219 @ F" etc. New config option 'DomainType' to alter the way Arc-Binkley displays domains... e.g. DomainType # will display addresses like : Fidonet#2:250/219.0 Spherenet#20:925/101.0 whereas DomainType @ (default) displays : 2:250/219.0 @ Fidonet 20:925/101.0 @ Spherenet NoPickup flag placed in flags menu so you can call a system and not pick anything up (EMSI only) When rescanning the outbound, the window is cleared properly! Outbound window has a minimum size so it will always 'look' right when only a few nodes are in it. Ensured DTR is enabled and XON/XOFF disabled before an init - could have caused problems before, symptoms are extremely slow quitting, problems mainly occured on non-Internal serial ports. When rescanning outbound, now displays 'Rescanning' in outbound window. Selecting Exit from the task manager should now shut up shop properly! Mask out 'locked' and 'reserved' bits on file reception attributes. * * * Completely reviewed (rewrote most of) the outbound handling routines, now should be faster, and more detail. Mail on hold for other domains and zones which you do not have nodelist data for, will now be displayed in the outbound window. 6d support added to outbound scanner (actual sessions won't work yet). * * * Complete redesign of templates, including a File Transfer window and System Info window to take the load off the Current Status window. ##### Copy the 'Templates' and 'Sprites' files into the same directory as !RunImage (i.e. !Binkley), the RiscOS200, RiscOS300 and RiscOS310 files in !Binkley.System are no longer required. NOTE: I will *NOT* be creating special windows for RISC OS 2 users, it takes too long to recreate them for such a minority of users. Obviously you can try to alter them if you wish, as long as the icon numbering or validation lengths get changed (see file IconNums). Outbound window now displays mail type (Mail, Flow, Freq etc.) and the relevant size (bytes or files for freqs) of the item. LogLevel replaced by a more flexible LogChars option, from which you specify the 'log characters' (e.g. +,:,*,! etc) which you want present in your logfile (and recent activity window). LogChars menu option now a dialogue box from which you can enter your own logchars or select them via the old method by clicking on 1,2,3,4,5 for loglevels 1-5, and Deb, Hyd and Ems for DebugLog, HydraLog and EmsiLog. Removed LogLevel, DebugLog, EmsiLog and HydraLog from configuration file. Removed Nodelist from config file (this is now irrelevant with the use of NListMgr). Removed a lot more config commands from the config file parser routine, for details see the example Config_0 file in the System dir. Tidied up source, removed redundant/inappropriate code/config lines. ##### The task name has now been changed from BinkleyTerm_x to Arc-Binkley_x (where x is the port number). To change ARCbbs you need to edit the file !Config, and change the line shown as : from Set ARCbbs$MailerTaskName "BinkleyTerm_%d" to Set ARCbbs$MailerTaskName "Arc-Binkley_%d" RiscBBS users need to edit their SysBBS files and change the line from binkley `<Binkley$Dir>.!Run 0` BinkleyTerm_0 to binkley `<Binkley$Dir>.!Run 0` Arc-Binkley_0 Note that they are case-sensitive. A symptom of an incorrectly set up bbs host is that when a user presses escape to enter the bbs, and nothing happens. Outbound window now has a Node Info entry, if you move across this a window will appear, from which you can see details of the node such as name, sysop, phone number or password. A nice feature of this window is the option to edit details of the node such as the password or phonenumber before polling. The same window is used to replace the Manual Poll option on the main menu, and hence you can edit the address (node number) of the system you want to poll, press return (or the down/up arrow keys to edit the phone number or password), the details of the node will appear, and you can either cancel or poll. Clicking on Cancel or pressing escape will remove the window, and clicking on Poll or pressing return a few times will poll the node in question, so long as a phone number is present. Dial Prefix menu has been added, so you can set the dial prefix up before making a call without having to quit and reload binkley after changing the config file. No config changes are required, simply alter the Dial Prefix from the Commands menu and poll! (Don't forget to change it back afterwards)!!! ;-) Outbound window menu now has Re-flavour, Re-direct and Delete options, to change the flavour (Normal, Hold, Crash or Direct) of an item in the outbound directory, then you press menu over the required item, then move across Re-flavour and select the desired new flavour. This is so you can easily change the status from Normal to Hold (for example) without having to interfere with the outbound directory at all. Re-direct means you can re-route mail and files to another node, for example, there is a netmail to 2:250/219.0 and you want to reroute it to your hub, click menu over the item in the outbound, move across Re-direct and enter the address of the system you want to route to. Delete is plain enough, select the item, press menu, move across delete and click on 'Do it' to delete the item. A new command line option 'errlog <fsp>' has been added, for details of errors in config lines etc. It is also used for more specific debugging which normally cannot be activated until the Binkley Icon is on the icon bar. To enable it insert it before the %*1 at the end of the line which looks like this: Run <Binkley$Dir>.!RunImage config <Binkley$SystemDir>.Config%0 errlog <Binkley$Dir>.Error_Log %*1 The FTSC product code list is now read in at startup rather than being hardcoded in the !RunImage - this is located in the System directory, called 'FTSC_Prods', and is the list of all known Fido programs. A new menu command 'Ensure events' is used when the event file has been altered. After you have altered your events file, click on Ensure events to make sure that the new events are registered. If you simply change the event file and save it back, the changes are not registered by Binkley unless you delete the BinkSch and BinkDay files too, so this is why 'Ensure events' has been added. There is no need to delete the sch and day files any more. Split up the Display=>Inbound menu into three separate submenu items: Default, Listed and Secure, for the three separate inbound directories (if you have it configured like that, otherwise just Default is the option). 10 Sep 1993: V2.15 Thanks to Geoff, Arc-Binkley now uses NListMgr to lookup nodelists, this means that you must ensure that the NListMgr *module* is loaded after the nodeserver module. The use of NListMgr is a very useful thing indeed. It means that new nodelist formats added by the author will instantly work with all software using nlistmgr (binkley, netway etc), eventually, every program which requires access to the nodelist should use it. It also means that you can process diffs/recompile nodelists etc, whilst binkley is loaded, without having to quit and reload again. Adding the following line after the line which loads NodeServer, in the !Run file, will ensure that binkley will function correctly. The new !Run supplied has this change among others. RMEnsure NListMgr 0.54 RMLoad NListMgr:NListMgr * * * If you do not have NListMgr it is available for File Request from * * * 2:250/219.0 using the magic name NLISTMGR. The use of NListMgr means that the configuration file must be altered. The commands 'ArchyList' and 'QuickNodelist' are no longer supported, instead, you should now use 'NewNodelist'. The 'Nodelist' config line must now state the directory in which the Domains file exists, e.g. '<Binkley$Dir>.Nodelist' rather than '<Binkley$Dir>.Nodelist.Fido'. Domain support welded in for 5d outbounds, configuration keywords OutFormat <type> added to select the type of outbound, currently all bbs software uses 4d addressing, so use : OutFormat 4d 4d uses the format Outbound[zz].nnnnoooo## or Outbound[zz].nnnnooooPN.pppp## 5d uses the format Outbound.dddddddddd.zzzznnnn.oooopppp## where d=domain, z=zone, n=net, o=node, p=point and ##=flavour The dddddddddd is read in from the Domains file in the Nodelist dir. When bbs's/packers etc start to use the 5d format more, change the 4d to 5d. Bug fixed whereby setting 'EMSIlog' in the config_x file *after* the 'baud' config line would cause havoc. Domain stripping works properly now, returning 'Fidonet' from 'Fidonet.Org'. Will now close inbound file request files so they can be deleted, if the caller hangs up, rather than giving files to the next caller! Bug fixed in the creation of _Z (BadWaZOO complement) files. Bug fixed in the scanning of point directories. If there were four points with FLO (0001FO, HO etc) and one of them had a packet (OT) as well, only the point with the packet would be displayed. Recent activity window has been widened so you people who are dying to see what was at the end of a line can now !! ;-) The little line which appeared next to the '+ Begin, Arc..' status line has been removed now (I'm glad to see the back of that)! Also, the window contents no longer 'scroll up' if you close and open it again. Outbound window no longer stays where you left it when you close and re-open, but it opens up at the original place as defined in the templates file like all the other windows. New config command 'openwindows' will open all the windows at startup, for those people who like to have the windows open all the time. FTS-1 file receive routine altered slighty so it now works with BBCscan systems properly. WaZOO update requests implemented, this now means that you must change your nodelist flag from XB to XA. e.g. ,27,Arcade_Bbs,London,Dave_Coleman,44-81-655-4412,9600,H16,V32B,CM,XB to ,27,Arcade_Bbs,London,Dave_Coleman,44-81-655-4412,9600,H16,V32B,CM,XA and yes, all USR 16800 bps DS modem users should have the above flags and nothing else. Hint hint Dave... yes, all arcbinkley users using USR 16k8 DS modems should have : ,9600,H16,V32B,CM,XA at the end of their phone number ;-) ... File requests would eat about 70 bytes of memory per request line, and if you already had the Avail config line set, it would override the ~AVAIL line in the OkFile rather than the other way round. Fixed now! Abort poll will now drop DTR (to force carrier drop) regardless of whether or not there is a connection. DietIfna file requests have been added - still waiting to see if they work or not though ! Hydra is now implemented - the bidirectional file transfer protocol. 8 new configuration options have been added for this new feature: HydraLog : Enable hydra debugging log NoHydra : Disable Hydra filetransfer protocol Hydra_Speeds <min> <max> : Min and Max speeds for full duplex (bidirectional) to be enabled. Hydra_NoFdx <string> : Force 1 way HYDRA for connections which report <string> Hydra_Fdx <string> : Force 2 way HYDRA for connections that only report <string> and 1 way for ANYTHING else Hydra_RxWindow <bytes> : Segmented streaming (forces an ACK after <bytes> bytes have passed - default 0 Hydra_TxWindow <bytes> : as for RxWindow, but for xmitting. Hydra_Options <telenet> : Escape Telenet (CR-'@'-CR) <xonxoff> : Escape XON/XOFF <ctlchrs> : Escape all control characters <highbit> : 7 bits mode, use 7-bit encoding rather than 8-bit <highctl> : Escape ctlchars with high bit set too e.g. Hydra_Speeds 1200 16800 : 2 way HYDRA between 1200 and 16800 bps Hydra_NoFdx Hst : use unidirectional HYDRA on HST connects (when modem reports Connect xxxxx/Arq/Hst/Hst/V42bis etc.) The Rx and TxWindow and Options commands are for advanced use and should be left commented out. The Hydra_Fdx is used as the opposite of Hydra_NoFdx, i.e. where the Hydra_NoFdx string means use two way xfer on all connects EXCEPT the one specified, and the Hydra_Fdx string means use two way xfer on the specified string ONLY. N.B. The Fdx and NoFdx strings ARE CASE SENSITIVE. So for a connect string of 'Connect 14400/Arq/Hst/Hst/V42bis' you want a Hydra_NoFdx string of Hst, not HST or hst. ** FOR PROPER USE OF HYDRA YOU MUST CONFIGURE A TRANSMIT BUFFER OF AT LEAST 2048 BYTES, using *SerialBufferSize 16384 2048 for a recommended 16k input/2k output buffersize. We need a dynamic SerialBufferSize command, one that can be altered whilst online. Flags menu has an on/off option to reflect the addition of Hydra. Wimp messages updated in accordance to the Mailer message spec. The following messages are implemented : FTNMailer_Rescan (0) FTNMailer_Outgoing (2) FTNMailer_Freeze (3) FTNMailer_UnFreeze (4) FTNMailer_KillSession (5) (& 6) FTNMailer_Exit (7) FTNMailer_ExternalCall (16) FTNMailer_PostSession (18) All other messages are not (yet) implemented. Disable and Enable changed to Freeze and Unfreeze. Now closes logfile before doing AfterMail, and also whilst Frozen. New config option 'FreezeTimeout' added - sets the default timeout value for when binkley is frozen. e.g. FreezeTimeout 30 ;sets a default of 30 seconds Whilst frozen, the floating window display changes to 'Frozen' and returns to 'Waiting' after being unfrozen or the timeout occured. EMSI data (sysop name, location etc. are now prefixed with the '~' character) rather than ':' Removed a lot of the xfer status lines so they all appear on one line (a lot neater IMO) * I THINK * I've definitely found the hanging problem... (the code which waits for something to be sent, or something to be received didn't have any carrier status checking in)! EMSI sessions from 252/110 caused havoc as the PUP flag (pickup mail for primary address only) wasn't working - fixed now. Started coding outbound window menu. 27 Jul 1993: V2.14 (KEH) *R* Pathnames truncated so they fit on the recent activity window. File request counter should not be incremented if the 'about' file was sent. File requests to point systems fixed - problem was xxxxRQ file being sent rather than 0000xxxx.REQ file :-) File sending wording changed... No longer tells you the file it's deleting as it's obvious! MailPkt log entries changed too. Extra chance given to complete the yoohoo/2u2 session in case of a line hit, now waits for up to five incorrect YOOHOO characters before dropping carrier. File requests do not return a 'Received file' errorlevel. Internal exit() handling revised. Looked again at the misterious loop when carrier is dropped during a transfer. I need more info - does it only to it when sending, or receiving, or both? When it happened here - it was sending. Implemented Rescan, Quit, Disable and Enable wimp messages. ARCHmail updated to use these messages. Disable can have a timeout for automatically enabling - in case the program which disabled it crashes and forgets to re-enable etc. Disable and enable placed in commands menu. Poll wimp message added to poll systems. Abort poll to abort poll. Now with EMSI which couldn't be implemented this fast without help from Steven Haslam! Due to EMSI, seven new configuration commands have been implemented. These are: NoEMSI (when present, disables EMSI) EMSIlog (when present, includes EMSI debug log in main logfile) NLsystem (if present - overrides default 'system' configuration during EMSI sessions - e.g. TPS - MailNode) NLsysop (if present - overrides default 'sysop' configuration during EMSI sessions - e.g. Geoff Riley) NLcity (if present - transmits the system's location in EMSI sessions - e.g. Lymm, Cheshire) NLphone (nodelist format phone number, e.g. 44-925-757920) NLbaud (maximum connect baudrate e.g. '14400','16800','2400' etc) NLflags (nodelist format flags e.g. 'CM,XB,V32B,H16' etc.) Note that any NLxxx config keyword (except system) will be substituted with 'Unknown' if not present in Config_0. EMSI_CLI requests should be interpreted for fast BBS access. Useless time delays removed and changed, so there is a 6 second delay before sending the intro line of your system if someone called you and you have a modem which doesn't have ARQ capability - i.e the connect message returns before the remote system's connect message. 2 second delay otherwise. ZedZip/EMSI_Zmodem should work now - config commands 'nozedzap' and 'nozedzip' added to disable them. 'NoWaZOO' config command will disable yoohoo/2u2 on incoming rather than just outgoing call now. After a few days of testing - I think the mysterious lockup after drop carrier on zmodem send is fixed! Thanks for Edward Hobson for his patience in helping me test EMSI sessions with a properly written program - FrontDoor! All this testing costs me too much. Just a few too many modem initialisations are done when aborting a dial etc, I've removed some so it initialises, but not lots of times. The problem with the block sizes not always being displayed should have been fixed. Geoff helped sort out password problems with EMSI, as well as helping test EMSI too, and adding a really nice LogLevel menu - good one Geoff! You can now have the 'magic-magic' file requests 'FILES' and 'ABOUT' listed in the OkFile using the following 'magic-magic' okfile names: ~ABOUT <pathname> [<alias>] ~AVAIL <pathname> [<alias>] e.g. ~ABOUT <Binkley$Dir>.FileReqs.MySystem Mysystem.Txt ~AVAIL <Binkley$Dir>.FileReqs.Avail25219 Filelist.Txt Note that these 'special' magic names MUST appear at the very start of the OkFile, and aliases are optional. Flags menu so you can turn off EMSI, WaZOO, ZedZap and ZedZip (EMSI Zmodem), at the click of a button! (Quite neat me thinks!). Info box will no longer 'freeze' binkley whilst open due to a little fix. If anyone using RiscOSLib wants to know - you know where I am! Disable will send the 'Busy' config string to the modem, and 'Init' once enabled again. I've changed all references to 'BinkleyTerm' to 'Arc-Binkley' as that is what it is really called. BinkleyTerm lives on the PC. I'll be changing the Wimp task title in the next release too, so now is the time to change bbs host software which uses the task name to decide when to startup/toss mail etc. Properly written software should by now be using the Wimp Messages broadcasted by Arc-Binkley. Existing users of RiscBBS will simply need to change the name in the SysBBS file under the configuration keyword of 'binkley' when 2.16 is out. All references to FTSC-0001 have been changed to FTS-0001 as I have been informed that is correct - thanks Birger !! :-) Working on a way to get DirectZap (EMSI Zmodem variant) working. Removed inits from 'Poll completed' message and changed that message to a debuglog only option. An 'R' flag in an event containing 'M' should now not allow calls to be made to non-cm nodes during ZMH. Inits are made as soon as a FAILED connect appears - just in case the failure could have been a modem prob. Changed a little thing in binkley to allow EMSI to operate in non-fido networks without using separate outbounds (coming soon tho)... Yoohoo/2u2 probs incurred during EMSI implementation have been ironed out completely (I hope)! 07 Jul 1993: V2.13 (KEH) *R* Magic names can now have aliases in the okfile! Problem with dupicate 'BBSbase filename (xxxxxx) generated' and reduction in maxreqlim should be fixed now. Added magic name message "not known - sending 'about' file" if a freq for a magic name not present in OkFile was received. Version 2.12 was never released - just passed to a few people for testing etc. hence the date of Apr/May 1993. Added support for antialiased fonts in binkley if you want to change the templates :-) I tried it but didn't like it - the icons don't get updated properly unless they are filled! Updated NodeServer module supplied (3.04) - requires change to the !Binkley.Nodelist.Domains file (see domains file in archive). Log files kept closed most of the time now, opening only when anything is to be written to the file. It's closed when binkley is 'Waiting'. It also gets closed every five seconds while in Waiting 'mode'... :-) The filetype is changed to Data when open and Text when closed. Thanks to Stefan Brueck for the logfile idea Apr/May 1993: V2.12 (KEH) LockFossil short changed to long to allow >19200!!! Increasing block sizes on Zmodem xfers similar to FrontDoor, so that s***ty little slow modems don't end up with a CRC error or something after 1024 bytes of data (about 9 seconds for 1200 baud), or it could be worse still! Points 'fixed' for the time being - I've commented out some of the key code - will fix properly when I get time... Wimp Messages added, 0x800, 0x801 - sent when exitting to bbs or normal exit respectively. Overwriting of phone numbers and passwords fixed, fault was due to removing the 'set page 0 byte 0 to zero' 'bug' that a few people moaned about :-) Increasing block sizes on xmit now start from 256 rather than 64... 14 Apr 1993: V2.11 (KEH) *R* Large stinging bug found which meant that no version of binkley compiled on my machine would work with any serial device other than the internal serial port! Logical port number can be set for those with 'mixed' serial cards, eg those which result in two different ports with the same number. For example - if you have a single II card and also a single SP card then you will have two port 0's and two port 1's, but one is the II card and the other SP... Now an extra bit has been added to the Port config line for binkley which should get around this problem by assigning a logical port number to the port being used by binkley. This is achieved by using the # character followed immediately by the port number to be used in BinkEvtx, BinkDayx, BBS_x and the port number displayed in the status window. e.g. : Port @SP_Dual #4 1 Port @<driver> #<logical port no> <actual port no> which means that BBS_4 files will be created when binkley exits to an ARCbbs system, rather than BBS_1 files, whilst still talking to the Serial Port dual serial card port 1. The Port display field in the Status window has been updated to accommodate this change. Now is displayed the actual port number followed by the logical port in brackets... In the above example it would display: Event:xx /xx Port:1 (#4) Driver:SP_Dual Port numbers now are correct port numbers, none of this Port+1 rubbish which was taken from the PC's Com1 etc... Erm, extra information in Config_0 file and !Run updated for block drivers. RISC OS 2 compatibility improved for BlockDrivers - Paths aren't handled in the same way as in RISC OS 3.1 Driver name now displayed in the status window, instead of the Status field, which was really superseded by the floating window anyway. The 'bug' which people are starting to moan about - the overwriting of byte zero to stop another 'fault' in porting from the PC is being looked at... 01 Apr 1993: V2.10 (KEH) *R* Hmmm.... Decided to go to 2.10 now as I think we're getting a bit out of hand with the letters, and anyway, you can't tell differences if someone calls you, since only the numerical version number is passed during YOOHOO handshake. Silly bug which printed Can't find address XXXXX for a load of nodes which are not in the compiled nodelist has been made so it only gets displayed during debuglog sessions. Major source reorganisation - may improve speeds, very marginally. c.com and h.com files no longer there - moved into c.port_sprt and h.port... Hugo also spotted another short baudrate instead of long :-) Strange bug where FTSC-0001 sessions spark up on incoming mailer calls may have been fixed, only time will tell [grin] ;-) Added a short note to explain about driver names in Port, in Config_x. 25 Mar 1993: V2.09p (KEH) *R* Ahem! I rushed things a bit, and nearly distributed, totally, the version of block driver binkley which doesn't set any speeds other than 19200 :-) ooops... I didn't like the 'o' on the version anyway! 24 Mar 1993: V2.09o (HF) Fixed serial block driver support (added support in Zmodem). Baud's all over the place are now longs, prepare for the future! Main block driver problems were (a) not used in zmodem assembler and (b) flow control was set to xon/xoff only when xon/xoff was requested by Bink, whereas it should have been xon/xoff *and* hardware. 14 Mar 1993: V2.09o (GIR + KEH) Block driver support welded in - keeps missing characters off for some reason... Sent copy to Hugo to look at. 07 Mar 1993: V2.09n (KEH) *R* Binkley's FTSC-0001 only mode works correctly now, but filerequests still do not work. Config option FTS_0001 used for this purpose. File attaches now work when sending, with aliases. Added a bit of debugging code to Zmodem to show up where the transfer cancelled message actually came from. Mail me if you have this occuring. Outbound name case insensitivity added. This was to stop names such as '01020018hO causing major problems. Started to write a bit of code to show up everything in the outbound window, not just 'selected' nodes. 17 Feb 1993: V2.094m[dev] (KEH) FTSC-0001 sessions fixed, should no longer crash, and also files should be sent and received. The main problem was in using the incorrect CRC calculation routine :-). Thanks to Edward Hobson for his assistance with testing these yucky FTSC-0001 connects. NetFile exit now changed to '4' to stop confliction with ARCbbs's '7' errorlevel, which exports mail. File requests using FTSC-0001 don't work. Started to make Binkley poll out as an FTSC-0001 only mailer. 06 Feb 1993: V2.093m[dev] (KEH) Exits after ArcMail received are no longer 'stopped' if a NetFile is received AFTER the ArcMail. 03 Jan 1993: V2.093m[dev] (GIR) *R* Cumulative elapsed time now recorded in the history file, and displayed in the 'Today at a glance' pane. Plus elapsed time for each call and time of !Binkley being active are recorded in log file. (Just in case anyone wants it - it takes hardly any code space...) 28 Dec 1992: V2.09m[dev] (KEH) Fixed Outbound window updating, it does it correctly now - it failed to update if there was nothing in the outbound area at all i.e. a call removed the last item in the outbound. Wow! I think I've found the _slow_ problem - a call to fill_in_status at every wimp poll to update the time elapsed etc. I think this was the culprit which slowed it down! Apart from RO3.1 serial port ;-( I'm looking for a non-flicker method of icon updating at the moment ;-) Starting to put all the binkley words into a messages file. This includes log format etc. 11 Dec 1992: V2.09l[dev] (KEH) Fixed (I think) the efficiency percentage prob - it should now give the effiency relative to the _connection_ baudrate, not the baudrate you talk to the modem at. Removed 'BINK' characters from log - will save you 5 bytes per line of log - actually amounts to quite a lot! 01 Dec 1992: V2.09l[dev] (KEH) *R* Removed bug whereby if you click on Exits instead of quit from the iconbar menu it would say 'Junk character from keyboard:continuing' in the recent activity window ;-) Unlinking <Binkley$Dir> .... changed to Removing <Binkley$Dir> .... New config option EventFile <name> where <name> is the name of the event file for the port. This allows separate event files, like separate config files for different ports. Examples of use of EventFile : EventFile <Binkley$Dir>.BinkEvt0 ; for config 0 (port 0) EventFile <Binkley$Dir>.BinkEvt4 ; for config 4 (port 4) EventFile <Binkley$Dir>.System.BinkEvt0 ; for port 0 Files BinkScd and BinkDay like BinkEvt will be created, where is the port no. Now exits correctly with event E2 for non-compressed mail and E3 for compressed mail (used to always exit with E2), so you need to alter the event file so E2=5 and E3=6 Put seconds back in time at top, but now updated every 5 seconds, so flicker isn't really noticed. Working on a flicker free method like !Alarm uses! Abort Poll menu option added, to abort the dialling when immediate call or Manual poll has been activated (in fact it will abort dialling at any point). A bash at immediate originate has been made - not quite working yet, and needs some thought. Unattended mode can now be assumed whether it exists in Config_ or in !Run, since not having it caused binkley not to initialise! Immediate originate completed - but have to enter the node of the remote computer you wish to communicate with! Errorlevel E4 in events used for NetFiles (not Mailpkts or ARCmail). '<Binkley$Dir>.' no longer assumed for config and event files anymore, must supply <Binkley$Dir>.<file> instead of just <file> to the config command line option in !Run and include & eventfile in config_0. Added modem command characters '>' and '<' to turn off and on the requirement of DSR (to allow characters to be sent to the serial port). Some (old) modems don't turn the DSR line on until a connect, therefore disallowing all communication to the modem. The fix is to turn DSR requirement OFF and then send stuff to the port, i.e. a '>' character in the Modem Init string. Exits menu expanded to allow 'standard' exits (i.e. 5, 6 & 7) to be easily executed. Removed the need for share in the !Run file - aaarrggghhh (if it wasn't present the serial port was disabled whenever binkley had to exit, including going to the bbs, therefore having the effect that users cannot log in due to no input from the serial port). 24 Nov 1992: V2.09k[dev] (KEH) Separate template files used for different operating systems, e.g. RiscOS200, 300, 310 are for RISC OS 2, 3.00 and 3.10 respectively! These are now placed in the <Binkley$Dir>.System directory. The <Binkley$Dir>.System.Nodelist directory has been moved back to <Binkley$Dir>.Nodelist... 22 Nov 1992: V2.09j[dev] (GIR) *R* Removed useless M'Task label from Status display: replaced it with elapsed time display. If you redefine the colours in the outbound window, the colours will be used for the text (background and foreground). I have done this predominately to allow for the new optional template files. RiscOS 2 has the old style templates; RiscOS 3 has the new style 3d icons. 20 Nov 1992: V2.09i[dev] (GIR) Removed 'seconds' from time display: stops the flickering which annoyed Stefan (and me) somewhat! Fixed (again) the @@ArcBBS okfile filename generator - a slight oh dear occured when I added in the @@BBSBase filename generator... [sigh] 15 Nov 1992: V2.09h[dev] (GIR) 'Fake' points would occationally appear in the outbound if there were empty point directories after a point directory with a file in... This is now fixed. Another wonderful bug - the Binkley Schedule file always got saved with the time and date stamp of the last received file!! Hmm... Fixed. 14 Nov 1992: V2.09g[dev] (KEH) *R* The Number of Rings facility didn't work! (properly anyhow!) Instead of resetting the number of rings 2 times a minute, it will now reset only if there haven't been any RING's for ten seconds. So if someone calls, rings 4 times, hangs up and calls again, the rings counter will be reset, since it will be more than ten seconds since the last RING. 13 Nov 1992: V2.09g[dev] (KEH) If Rings is set to 4 and a caller rings 2 times, and then hangs up before binkley answers, the next caller would only need to ring 2 times for binkley to answer, since it was not resetting the value afterwards. This is now fixed, by resetting it 3 times a minute! Added Re-Init to floating window display when reinitialising the modem after 10 minutes of silence! Buggy RingString code rewritten to 'proper' specs! PreInit and PreDial config options implemented. 'Freqs from us allowed' and 'No freqs from us allowed' messages now only activated by the debug log. NetWay 'seen' by binkley now, as product code in binkley was one line out, forcing it to think NetWay was product code ED! 10 Nov 1992: V2.09f[dev] (KEH) Added configuration options 'RingString' and 'Rings' so that any modem (within limits) can use Binkley, since you can now define the message to look for when the modem detects a ringing signal. Also, the rings config option defines the number of rings between 1 and 15 before Binkley attempts to answer. 07 Nov 1992: V2.09f[dev] (GIR) Domains operate beyond the first address/aka now - slight mix up in the variable used to hold the current aka. All fixed up right now. Outbound window: Double click with select button will, as before, cause a repeated manual poll to the node clicked; Double click with the adjust button will now cause a single immediate call to the node clicked, provided that the current event allows calls to that node. File Requests at the system which is being polled are now displayed correctly. Filetypes now transfer along with load and exec addresses between like systems - extra info included in all file transfer options (when things other than ZedZap start working!). Immediate Answer now an additional menu option - guess what it does. 06 Nov 1992: V2.09e[dev] (KEH) Spawn event handler fixed. Not sure of the use though... Added <Binkley$Spawn> for the spawn file... Bug whereby points were ignored. A misplaced '.' was the cause! Tried fixing the colour when double clicking outbound entries but with no luck... Better try harder next time! Fixed MiniWin display to accept full info when connected e.g. 2:250/219 @ Fidonet 05 Nov 1992: V2.09e[dev] (GIR) New config option 'BlindDial'. This provides a method of inserting an extra pre-dial string which will not appear in the log file or on screen. This is mainly for use with the Mercury system: you put your Mercury pin number in as the BlindDial string, and put '%%' before each number which requires Mercury. This is only really useful if you do your dial translations in ParseLst - I'm working on it for those who do all dial translations in Binkley. 18 Oct 1992: V2.09d[dev] (GIR) Close log file before starting up the EndBatch task. Now call time_release during directory scan to prevent the apparent lockup during initialisation. Filename truncated to the base name when being displayed in the file transfer window. (No <Binkley$Dir>.Outbound.... rubbish now!) Get the banner resent after 2 second intervals... bit much maybe? 17 Oct 1992: V2.09c[dev] (KEH) *R* (Acorn User Show 1992) Exit level now set correctly, I was setting Binkley$Return instead of Binkley$Return, where is the port number. Also, the port number is now the first parameter passed to the EndBatch file. 12 Sep 1992: V2.09b[dev] (Keith Hall, or KEH) Tidied up main window, placing in RO3.1 border validation strings. Two lines instead of one for address. Gap created on other side due to this is now filled with a version number. Seconds placed on the time at the top right. Should not slow the program down as it is called the same number of times as the old one! (It should even speed up a touch if anything...) Exit level now set from within Binkley code itself. Now runs a file ( '<Binkey$EndBatch>' - defined in !Run) when Binkley quits. This (should, and seems to) stop the bugs in the Obey module in RISC OS III. To use EndBatch, remove all lines from it and move everything _AFTER_ 'Run <Binkley$Dir>.!RunImage ....' in !Binkley.!Run into it. Make sure however, that you must remove the line 'Set Binkley$Return%0 <Sys$ReturnCode>' from it otherwise ARCbbs will get confused and will probably not work! 31 Aug 1992: V2.09Da Magic file requests fixed properly now! New magic file generator: @@BBSbase. This is a variation on the @@ArcBBS magic file generator. The latter always generates a filename with a base of <ArcBBS$FileData> and a step value of 50. This is suitable (of course) for use with ArcBBS. The new verb takes the form: @@BBSbase [!password] (base path name) (step value) The password is optional, as normal. The base path name is the path which is to be appended before the '.000000.000000.000000' type filename. The step value is the number of file divisions used by the filebase: ArcBBS uses 50, RiscBBS uses 70. The command: @@BBSbase <ArcBBS$FileData> 50 is directly equivalent to: @@ArcBBS 20 Jul 1992: V2.08Df *R* Whoops! Freq's were never allowed due to a wrong capabilities byte being accessed during setup of the Hello packet header.... 19 Jul 1992: V2.08De Double click in the outbound window now correctly forces a call to the node clicked on - not just the next one which can be called to! Unknown modem responses are now ignored rather than causing a fail. In Autobaud mode the baud rate will follow the connect rate for none ARQ connects, but use the max baud for ARQ connects. (Like it used to!) 06 Jul 1992: V2.08Dd Further additions to the domain functions. Filerequests now check for all inbounds - ready for EMSI support! 29 Jun 1992: V2.08Dc *R* Modem initialisation routine accidentally deleted during mods for V2.08Db.... Now back in place! Events can now have month and day qualifiers added to the start time: Event All 06:00,12,25 06:00 F E1=200 This would create a forced event which does an exit with errorlevel 200 but *only* on the 25th December. You can miss off the day - in which case the event is valid for every day during the specified month, or you can make the month 0 - which indicates every month.. you could have something which happens on every Friday the 13th: Event Fri 06:00,0,13 06:00 F E1=201 Let me know if you can think of any good uses for this - I have it set to do a backup on the first Monday of each month.... (Specify seven separate events - all active on a Monday, but each one for a different day of the month - 1st to the 7th..) 24 Jun 1992: V2.08Db (GIR) Added initial domain support. In config file address & aka now allows @domain to be added. Fixed problem which stopped Quick nodelist from working. 16 Jun 1992: V2.08Da (SRH) Did some looking at sched, found odd things (I think). Hum. Anyway, sched has been poked a bit. Changed the HISTORY struct to include a point number for the last_called entry. Changed the version number in BinkleyScd to reflect this, as per sched.h. Removed werr call giving "Not yet implemented" error boxes Wrote a nice, big, commented BinkleyEvt file after looking at sched. Added "void show_event(int)", which is called when a new event is started. It updates the event display. Event display now shows "M" and "X" flags. Changes in "BinkleyEvt" now seem to be detected. ??? Started "real" 4D stuff: 4D outbound scanning works a treat! Works like this: Outbound.00fa0074OT <- stuff to 250/116 Outbound.00fe0192OT <- stuff to 254/405 Outbound.00fa0074PN.0002OT <- stuff going STRAIGHT to 250/116.2 Wrote in 4D nodeproc. 4D is supported for QuickBBS and ArchyList 1.01 Added nodeproc test for index/data consistency. It finds the address in the index and checks that the index address and data address match. This is if you get a new nodelist and forget to re-load the index (using *LoadNodes or *RMLoad NodeServer) Manual poll icon now 4D. To poll a point, you must supply the zone, i.e. "2:250/116.2" is OK, but "250/116.2" is read as "250/116". Did some stamping out of references to "<Binkley$Path>" (particularly in b_bbs.c). Did some work on ExtrnMail, it now writes "Mail_x" where x is the port number (like BBS Batch writes "BBS_x"). Added ArchyList format. Requires NodeServer 2.23 to work correctly. mailer.c now recognises the B_never flag, which is available from ArchyList format. Changed announcement ("* NetWork Address z:f/n.p using " etc.), the first line is the announcement of the version and the second gives your address (and A.K.A.s if any) Went through ftsc.c and wzsend.c checking that provision was being made for 4D outbound. Added _determine_aka to ftsc.c Updated product code list to 0xe2. Need a way to keep this externally. 14 Jun 1992: V2.06ßm (Steven Haslam, or SRH) Well, folks, I'm not going to get your domain support in fast, I'm afraid.. Added four dimensional multi-address functionality. :-). The way that you configure your address has changed, see example config file for details. Bink complains is you attempt to use the old-style system. Binkley still has 3d-only things in various strategic places, however. For example, it's 3d-only nodelist procedures preclude a 4D outbound for a while. However, it works today, I've polled all over the world (well, to zones 1 & 3 anyway) and I can connect. Plus the outbound scanner is only 3D at the moment. Added debugging log entries; in the future, versions of Bink will have debugging entries in them, but at LogLevel 6. These entries (type ">") are therefore only available by using the "DebugLog" keyword. So, if you get a repeatable problem, send a loglevel 6 log (please!) Translates 'ß' to 'á' (ANSI beta) in banners, so you can have beta characters in the version, and get them displayed as Latin1 on the desktop and as betas for callers! Despite a hitch that went if (*c='ß') *c='á'... hehe, you see what I mean. For those in the dark, the side-effect in the conditional is to set ALL characters to CHR$223, and then to the á; the upshot of which was lots of beta characters! Added modem report string configuration, not everyone's modem EXACTLY returns "CONNECT x/ARQ/REL" you know! Plus, mine returns "No Dialtone" when "No Dial Tone" was looked for... grrr... 8-O Added ECDelay for modems that report the connect speed and then, later, report the ARQ type (if any). "ECDelay <x>" where <x> is a number giving the number of seconds to wait for an ARQ report. Too high and you'll start getting Intro: lines as ARQ reports! If an ARQ connect is established, it is shown as a ":"-status entry, so if you get a "too many errors" entry, you know if it was an ARQ connection. ARQ reports (in mdm_reliable[]) have spaces translated to dots, so a report of "Connect 2400 V42 Bis" will be reported in BBS_x as "SPAWNBBS 2400 0 V42.Bis". This version was not publicly released, it was just handed out to a few people to test. 10 Apr 1992: V2.06l *R* If SmallIcon is not set to 2; pressing the right mouse button crashed the program, when the non-existant floating window is attempted to be closed... Initial zone and domain support included - not yet functional, but the support routines are appearing. 01 Apr 1992: V2.06k Added separate 'outbound' window: first step to replacing the large status window. Display:Inbound/Outbound/Nodelist now looks at the provided config for the filer to open - rather than assuming that they are in <Binkley$Dir>. Left mouse button now opens all windows, whilst the right mouse button now closes all windows. When Binkley is quitting the small floating window or message under the icon now reads "Quitting". During a poll, the node being polled is now highlighted in the outbound window. An immediate poll to a node listed in the outbound window may be initiated by double clicking on the entry. 27 Mar 1992: V2.06j Fixed spelling of 'dialling'. Added command line option and config file option: SmallIcon - this restricts the icon width to 7 characters; this means that the node doesn't all fit on at connect time, but it uses less space. In the config file only, this option may be followed by a number: 0 = normal wide icon with messages beneath. 1 = narrow icon with messages beneath. 2 = narrow icon with 'floating' message window. 13 Mar 1992: V2.06i Taken VERSION numbering out of header file and placed into a separate module - doesn't necessarily need to recompile the whole lot for a small change any more. 29 Feb 1992: V2.06h Removed all foreground monotasking routines. Terminal mode no longer available. 19 Jan 1992: V2.06g Only try to generate ArcBBS filenames if the reqested file is all numeric. Fixed manual exit from Binkley to return 1 as the exit code - it had started to return 0 by mistake! Now got our own FTSC product code: 0xCA is Arc-Binkley. Product code identification code changed to identify more of the existing products without having a long list of case statements. 13 Jan 1992: V2.06f *R* Why can't I learn to read? I went through the whole freq processing routine, and realised that the line read from the OKfile is placed in 's' not 'request'! 12 Jan 1992: V2.06e Realisation at last - the compare for @@ArcBBS should be using the counted compare routine strnicmp (). 01 Jan 1992: V2.06d I hope this is not how 1992 is going to continue - I'd managed to miss out an '== 0' which stopped all freq's except for ArcBBS file requests from working. (But in 2.06b they didn't work either...[sigh]). 30 Dec 1991: V2.06c *R* Whoops... ArcBBS file requests were well out of flunter! Should be nearer the mark now... 30 Dec 1991: v2.06c (Hugo Fiennes) Zmodem assembler slotted into place - it seems to work. Should means less CPU load/faster transfers maybe. 23 Dec 1991: V2.06b Added new command option to open Nodelist directory. Solved the modem reinit problem - it was the timer routine working in signed int's instead of unsigned int's...! Added a command to rescan the Outbound immediately. 10 Nov 1991: V2.06a If a file is inbound with a filename longer than 8 characters before the extention then it would cause an error, and the file would be aborted. The arcify routine now spots this. (The PC version of Binkley still renames the file to BadWazoo.??? - so that's another one we're up on!) 08 Oct 1991: V2.06 Extra options added off the Display menu: Outbound Dir & Inbound Dir. These open the filer windows for the main outbound directory and the inbound directory. Text position under the BT icon on the icon bar increased to accomodate a full address at connect time. 07 Oct 1991: V2.06 Code to scan ArcBBS filebases added for file requests. If a filename of six digits is requested, and '@@ArcBBS' is in the OKFile, then an ArcBBS filebase name is generated. 24 Sep 1991: V2.05 Removed all the overlays. (Hugo's pretty sure they cause more hassle than they're worth!) Long filenames overwriting stack fixed - any filename longer than about 30 characters in a FLO file would corrupt the stack in the zsend function. NodeList.xxx now renamed to NListxxx and NodeDiff.xxx now renamed to NDiffxxx. Allow modem response strings up to 60 chars instead of 20! 11 Sep 1991: V2.05 (Geoff Riley, or GIR) Added alias code for file attaches. Format is (within FLO files): [<token>]<Archimedes filename> [<Alias>] The side effect of this is that you can no longer put comments at the end of a filename line. 30 May 1990 Fixed a problem in the file-request logic where it was searching for "\" between directories. The file-request logic is gonna need a lot more work though - wildcard processing is still rather PC specific. Also need to make 2 attempts - one with the file name given as-is, and the other after arcify()ing it. But simple file requests where an exact name (<8 chars) comes in and is matched by an OKFile entry ending in ".*" seems to be OK. Matching magic names <8 chars without a "." also seems to work. (re)"Released" 2.03 source for Hugo. 29 May 1990 Did a quick-fix for the hogging that Bink does when sending fast. The send entries to Com_ do a time_release after 0.35 second has elapsed without one. See h.donthog, c.com, c.wimp. Not a very good solution - Bink is using far more CPU than needed. Each byte goes through about 3 function calls to get sent! 28 May 1990 Did 1/2 a job on SpawnEvents. Get's a bit more complicated because of assumptions that errl_exit() doesn't return, and that the re-start of Bink will rescan the holding area etc. Rather do this when there's more time! 27 May 1990 Remember to rmkill the nodeserver after recompiling the nodelist - forget and strange things happen! 26 May 1990 Did a nodeserver module - simply loads the index into memory from the QNL_IdxBBS file in the directory given on the RMLOAD line and returns its address (and # of records) on call NodeServer_Op with R0==0. Hugo may want to fancy this up with a "find node" call etc, but I didn't need that. With the server WimpSlot goes back to 256k but I expect Hugo was hoping for less.... Still terminal and sb_ code to zap if it is really a problem! 20 May 1990 Finished the WIMP interface. New file TextWindow based on my Little-Smalltalk stuff. "full-screen" mode is still in there if you want it. Changes in c.wimp too. And a Template file. Did a Q&D for scrolling the pending window. Would be nice to fake a scroll-bar. Can't easily use a pane 'cos it can't be "behind" the edge of the window. WimpSlot has gone back up to 320k. Oops. "background" is now the default. You can now use "foreground" on the command line and Bink will take the screen. Increased time-out after connect to 18 seconds to help non-MNP callers out. Did the extra SelectPort()s. Added the extra time_release()s. Went around trying to spot the other places where Bink hogs the system! 19 May 1990 DTR is left on at exit, though we pulse it to try and hang up. This is because leaving it down causes other programs problems. Fixed a problem with the new nodelist - the internal copy of the index went >64k and the get_size() routine had an unsigned int type for the size. Changed unsigned ints to native_ints. 12 May 1990 To do: Fix bug with deleting outbound files before they have been closed. Move nodelist access stuff info a module. Hugo wanted some other changes? Phone him. Encourage using the "Answer" statement 'cos that helps if the wimp is suspended when a call comes in. If the modem auto- answers but Bink isn't getting any CPU the other guy is going to think we are down. Rather he gets no answer. Then drop the switch-off of DTR on exit. Only causes other programs problems. 16 May 1990 To do from 'phone converstation with Hugo: € Remember to reselect the serial port after every possible Wimp_Poll. Especially event handlers called via event_process. Also remember to select it after all serialop(0)'s due to a bug in the firmware. € Add time_release to the send/receive with wait stuff in c.com. x Drop the Binkley terminal emulator. Keep tabs on the "standard" for handing over serial port control that Acorn is supposed to come up with. x Think about a non-exit way to handle exit-events. € Have a go at moving the nodelist handling into a module. € Hugo's gonna send the latest !Spark. 16 May 1990 To do from 'phone converstation with Hugo: € Remember to reselect the serial port after every possible Wimp_Poll. Especially event handlers called via event_process. Also remember to select it after all serialop(0)'s due to a bug in the firmware. € Add time_release to the send/receive with wait stuff in c.com. x Drop the Binkley terminal emulator. Keep tabs on the "standard" for handing over serial port control that Acorn is supposed to come up with. x Think about a non-exit way to handle exit-events. € Have a go at moving the nodelist handling into a module. € Hugo's gonna send the latest !Spark. 12 May 1990 To do: Fix bug with deleting outbound files before they have been closed. Move nodelist access stuff info a module. Hugo wanted some other changes? Phone him. Encourage using the "Answer" statement 'cos that helps if the wimp is suspended when a call comes in. If the modem auto- answers but Bink isn't getting any CPU the other guy is going to think we are down. Rather he gets no answer. Then drop the switch-off of DTR on exit. Only causes other programs problems. 06 May 1990 Turned out I had been using Binkley$Path where I should have had Binkley$Dir. Fixed up. Also affects BinkleyCfg etc. That must either be changed or both variables should be set in !Run. Added a text label under the Binkley icon saying "Waiting", "Dialing", "Hangup" etc. When connected it shows the node number of the other guy. Added a menu for the icon with an Info box. € Don't forget to take the version stuff out of h.xfer and just include version.h 29 April 1990 Added a port argument on the Binkley command line. Use this to select which port as an alternative to the "Port" statement in BinkleyCfg. You get an error if you try and use both. Note the "MaxPort" command if you have more than one serial port. Port numbers are 1 based. The internal is 1, add-ons are 2 and up. Running several Binkley's on the same outbound/inbound areas is probably a bad idea. Either make sure only one can make outgoing calls (using BinkleyEvt) or give each its very own outbound area. Note the "config" command-line parameter to point different binks at different configs. You can have common settings in a file you "Include". Updated the version to 2.02. (hopefully there was no PC-Bink 2.02!) 29 April 1990 Seemed a shame to waste the space at the bottom of the screen. The "Recent Activity" window is now 7 lines bigger. Changes to hard-coded constants all over the show! 29 April 1990 Removed all calls to the RISC-OS library. Especially in c.wimp where calls to the wimp.h stuff was causing all sorts of other modules to be linked in. We call the kernel directly now. WimpSlot can now be 256k (that's now 64k smaller!) 28 April 1990 dosify converts NodList and NodDiff back to the DOS-style. Maybe its time to write a sort of pattern matcher so that users can configure the conversions to be done (rather than hard-coding into arcify/dosify). 28 April 1990 Added a "Background" command line switch to tell Binkley to go straight to the background when it starts up. This minimizes the time that other tasks are held up, particular when Binkley takes an errorlevel exit. Without this parameter Binkley comes to the foreground when it restarts after the exit and stops all the other tasks! NB: This goes on the command line, NOT in BinkleyCfg! Also tidyed the Obey files so you can change bink's switches in one place in !Run. 28 April 1990 Fixed a bug that caused Binkley to loop in xmit_reset (mailer). The cause was an assumption in my dfind() clone that Binkley wouldn't try to scan the same directory recursively. Turns out that xmit_reset scans the outbound, then within that calls bad_call which scans the same directory looking for call-progress files. This alter the "current position" saved by dfind in a static, resulting in xmit_reset finding the same file repeatedly. Fixed the problem by stashing the "next offset to use" information into the returned dta in the reserved area (which is probably where DOS buts it too!) 28 April 1990 Built an overlayed version of Binkley. WimpSlot was 320k, now fits in 288k (big deal!). Overlayed version in !RunImage et al, non- overlayed is in btbig. 26 April 1990 Added the "LockFossil" parameter to allow you to lock the baud rate of Arc-Binkley's internal fossil driver. Use it as follows: Port 1 Baud 9600 ; or whatever your modems top data-link speed is LockFossil 19200 ; the speed that you will talk to the modem Make sure that your init string enables fixed speed (&B1) and enables hardware flow control (&H1&R2). The serial cable must connect CTS and RTS. The init string I use with the HST looks like: Init AT&FE0X4B0&G2&P1&B1&H1&G2S7=60S0=1S9=18| 20 April 1990 The BinkleyScd file (and friends) that are supposed to maintain Binkley's status during error-level exits were not being written 'cos open() was being called with open-mode WRITE|CREATE. A bug in the Arc "clone" of open resulted in this open failing. 20 April 1990 Converted for C-3. 16 March 1990 arcify() spots nodelist and nodediff files. To preserve the week number at the end NODELIST.A66 becomes NodListA66.