• smb: Header field contains control characters issue

    From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 10:28:18 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5590

    I doubt the control characters were in the header field(s) as a part of import. But if they were, that'd be a bug.

    Instead, more likely, the control characters are a result of some other corruption to the message base that occurred *after* the message was imported. Running chksmb with the '-e' option and reporting the results would help me to know.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 11:06:53 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5591

    As per the request "Can we have fixsmb strip out invalid control characters from the header" - it could, but that's likely only a symptom of the problem. Looking at the header fields of the message(s) in question would be a good idea to determine if the header is even recoverable. Simply stripping control character isn't likely what we actually want to do in this case.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 11:28:50 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5592

    As requested

    ```
    Checking dove-syncops Headers

    93% #14922 (3DA220) SYS64738 Header field contains control characters

    message base dove-syncops
    To Digital Man
    Subject Re: Mail Server Issue
    Sender SYS64738
    SenderNetAddr REPLY: <66A92984.55864.SYNC_SYS@VERT.SYNCHRO.NET>
    @TZ: C168
    IF IT'S EXACTLY THE SAME PROBLEM, THEN YOUR PASSWORD IS STILL WRONG. I SUSPECT IT IS NOT THE "SAME PROBLEM", SO POST THE NEW LOG
    SenderNetType Internet
    Message-ID <66A9C774.764.998internetdiscussi@bbs.tangonine.net> X-FTN-MSGID 55877.sync_sys@1:103/705 2b102c17
    X-FTN-PID Synchronet 3.20a-Linux master/d3c114993 Feb 1 2024 GCC 12.2.0 X-FTN-CHRS CP437 2
    OtherHeader WhenImported: 20240731041114-0700 c1e0
    OtherHeader WhenExported: 20240731041558-0700 c1e0
    OtherHeader ExportedFrom: VERT sync_sys 55877
    OtherHeader @REPLY: <66A92984.55864.SYNC_SYS@VERT.SYNCHRO.NET>
    OtherHeader @TZ: C168
    OtherHeader > IF IT'S EXACTLY THE SAME PROBLEM, THEN YOUR PASSWORD IS STILL WRONG. I
    OtherHeader > SUSPECT IT IS NOT THE "SAME PROBLEM", SO POST THE NEW LOG OtherHeader WhenImported: 20240731141116+0900 121c
    OtherHeader WhenExported: 20240731201115+0900 121c
    OtherHeader ExportedFrom: TANGO99 998internetdiscussi 764
    OtherHeader @REPLY: <66A92984.55864.SYNC_SYS@VERT.SYNCHRO.NET>
    OtherHeader @TZ: C168
    OtherHeader > IF IT'S EXACTLY THE SAME PROBLEM, THEN YOUR PASSWORD IS STILL WRONG. I
    OtherHeader > SUSPECT IT IS NOT THE "SAME PROBLEM", SO POST THE NEW LOG SenderNetAddr REPLY: <66A92984.55864.SYNC_SYS@VERT.SYNCHRO.NET>
    @TZ: C168
    IF IT'S EXACTLY THE SAME PROBLEM, THEN YOUR PASSWORD IS STILL WRONG. I SUSPECT IT IS NOT THE "SAME PROBLEM", SO POST THE NEW LOG
    SenderNetType QWKnet
    when_written 66A983A8 0000 Tue Jul 30 19:22:00 2024 UTC
    when_imported 66AA1CFB C168 Wed Jul 31 06:16:11 2024 CDT
    type 0000h
    version 0300h
    attr 0000h ()
    auxattr 00000000h ()
    netattr 00000000h ()
    header_offset 3DA220h
    header_fields 25
    header_length 1200 (calc: 1200)
    number 14922
    thread_id 14922
    data_offset 48B600h
    data_field[0] TEXT_BODY, offset 0, length 67

    100%

    Checking dove-syncops Index

    100%

    Checking dove-syncops Hashes

    100%

    Status Total (=): 5165
    Total Indexes (=): 5165
    Active Indexes (=): 5165
    Active Headers (=): 5165
    Active Header Blocks ( ): 15514 3,971,584 bytes used Active Data Blocks ( ): 18174 4,652,544 bytes used Header Records ( ): 5636
    Deleted Indexes ( ): 0
    Deleted Headers ( ): 471
    Deleted Header Blocks ( ): 1446 370,176 bytes used Deleted Data Blocks ( ): 1755 449,280 bytes used Oldest Message (import) ( ): 965 days (0 max)
    Largest Message (data) ( ): 72125 bytes (#13716)
    Control Characters in Header Fields (!): 1

    dove-syncops Message Base has Errors!

    Total Deleted Messages : 471 819,456 bytes used
    ```
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 11:32:30 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5593

    That message is well corrupt

    ![image](/uploads/ce7cf4e9b3562a1aeddad2bcc3f4152e/image.png){width=918 height=709}
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 11:45:07 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5594

    I recommend deleting the message and then let me know if this happens again (is reproducible).
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 14:19:17 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5599

    Now it seems even more screwed up.

    ```
    92% #14922 (3DA220) SYS64738 Header field contains control characters
    ```

    So if I use smbutil to try and find 14922

    ```
    4756/#14921 Errol Casey phigan Cannot see all message boards
    4757/#14923 Denn Errol Casey Cannot see all message boards
    4758/#14924 Errol Casey Denn Cannot see all message boards
    4759/#14925 SYS64738
    ```

    It's not there.

    So let's run fixsmb

    ```
    smb_getmsghdr corrupt message header ID (04 00 00 00) at offset 3987232
    92% #14922 (3DA220) SYS64738 Not indexing deleted message
    ```

    Says it's deleted. So let's make sure

    ```
    bbs@bbs:/sbbs/data/subs/dovenet$ smbutil mp1000 dove-syncops.shd

    SMBUTIL v3.20-Linux master/d232ecdc6 SMBLIB 3.00 - Synchronet Message Base Utility

    Opening dove-syncops
    Maintaining dove-syncops
    Maintaining dove-syncops hash file
    Loading index...

    Done.

    Scanning for pre-flagged messages...
    100% (0 pre-flagged for deletion)
    Scanning for read messages to be killed...
    100% (0 flagged for deletion due to read status)
    dove-syncops locked successfully
    Packing dove-syncops
    Analyzing dove-syncops
    820992 less than 1024000 compressible bytes.

    dove-syncops unlocked successfully
    ```

    Nothing flagged for deletions so let's double check with another chksmb

    ```
    Checking dove-syncops Headers

    92% #14922 (3DA220) SYS64738 Header field contains control characters
    ```

    Still there. I don't seem to be able to get rid of it.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 14:31:16 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5600

    Use 'smbutil p' (pack) the message base to remove deleted messages.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 14:31:54 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5601

    Might need to use 'smbutil -a p' to force the pack.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 14:45:05 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5602

    Yes, the -a p did it. Thank you.
    I'll maybe run an smbchk every hour and have it flag me if there's corruption. Try and catch it quick.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 15:12:51 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5603

    I'll make a change to chksmb to *not* check header fields of deleted messages for ctrl chars
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 16:11:23 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5604

    Cool, while you're diddling with smb stuff, could you have smbutil l show the type of message if it's not a netmail, for example and UPVOTE or DOWNVOTE instead of who it's too. The blanks confused me for a bit. Also a date would be good, either date written or date imported.

    I can open a separate issue for that.

    ![image](/uploads/ddce94ebc15a5fbc502e7764c237fbaa/image.png){width=960 height=197}

    ![image](/uploads/d27b68641bb4d13bdce638b0f1e46654/image.png){width=804 height=252}
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Mon Sep 9 18:29:39 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5608

    I think we can close this issue now.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Wed Oct 23 00:17:38 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5785

    chksmb -e is no longer displaying extended information.


    ```
    $ chksmb dove-techtalk -e -p -b

    CHKSMB v3.20-Linux master/d84b8e3d6 SMBLIB 3.00 - Check Synchronet Message/File Base

    Checking dove-techtalk Headers

    35% #86 (02F120) TLDR Header field contains control characters
    100%

    Checking dove-techtalk Index

    100%

    Checking dove-techtalk Hashes

    100%

    Status Total (=): 230
    Total Indexes (=): 230
    Active Indexes (=): 230
    Active Headers (=): 230
    Active Header Blocks ( ): 2104 538,624 bytes used Active Data Blocks ( ): 66765 17,091,840 bytes used Header Records ( ): 230
    Deleted Indexes ( ): 0
    Deleted Headers ( ): 0
    Deleted Header Blocks ( ): 0 0 bytes used Deleted Data Blocks ( ): 0 0 bytes used Oldest Message (import) ( ): 249 days (0 max)
    Largest Message (data) ( ): 6308005 bytes (#88)
    Control Characters in Header Fields (!): 1

    dove-techtalk Message Base has Errors!

    'fixsmb' can be used to repair many message/file base problems.
    ```
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Nigel Reed@1:103/705 to GitLab note in main/sbbs on Wed Oct 23 00:17:58 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5786

    @rswindell prod?
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Wed Oct 23 11:25:10 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5798

    I'm not sure what you're asking here. The '-e' option never had any impact on the output of the detected header fields with control characters. "no longer" does not apply here. Are you asking for chksmb to report more information for those particular detected msg header problems when using the '-e' option?
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Rob Swindell@1:103/705 to GitLab note in main/sbbs on Wed Oct 23 11:55:22 2024
    https://gitlab.synchro.net/main/sbbs/-/issues/786#note_5800

    Okay, I think I see now what you're reporting: any/all problematic message headers are indeed dumped when using the '-e' option (even those with unexpected ctrl chars in them), but you have to specify the '-e' option *before* the message base(s) on the chksmb command-line.
    --- SBBSecho 3.20-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)