bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Emacs 22.0.50.1 crash on make_uninit_multibyte_string


From: Daniel Barrett
Subject: Re: Emacs 22.0.50.1 crash on make_uninit_multibyte_string
Date: Fri, 16 Sep 2005 11:47:03 -0400

On September 16, 2005, Daniel Barrett wrote:
>In case it helps, here's the output of vm-submit-bug-report.  I'll also
>forward this to Kyle (author of vm).

Here are Kyle's remarks.

: From: Kyle Jones <address@hidden>
: To: Daniel Barrett <address@hidden>
: Subject: Emacs 22 crash with vm
: Date: Fri, 16 Sep 2005 08:10:27 -0700
: 
: Daniel Barrett writes:
:  > | >  [...]
:  > | > (provide 'vm-byteopts)
:  > | 
:  > | 
:  > | I wonder if some of these settings may trigger some unusual
:  > | byte-code pattern which we don't normally see in emacs 22.x ???
: 
: Almost certainly.  I doubt if many programs support Emacs versions
: as far back as VM does.  But that file is inaccurate in one
: area.  I don't support v18 Emacs anymore.
: 
: I can explain what the code is doing in the backtrace.  It will
: make more sense if you skip to the bottom and read the comments
: in reverse order from there.
: 
:  > | > (gdb) xbacktrace
:  > | > "vm-left-justify-string"
:  > | > "vm-reencode-mime-encoded-words-in-string"
: 
: In this case RFC 2047-style encoded words must have appeared in
: some header, probably Subject.  The words are extracted, decoded,
: left or right justified, then reencoded and returned, eventually
: to be stored in VM's summary cache.
: 
:  > | > "format"
:  > | > "list"
:  > | > "eval"
: 
: The template formatting is done with Lisp calls.
: 
:  > | > "vm-summary-sprintf"
: 
: Generates the summary information for a message using a printf-like
: format template.
: 
:  > | > "vm-su-summary"
: 
: Returns the summary information for a message as a string.  The
: information may be cached, or it might call vm-summary-sprintf to
: generate it.
: 
:  > | > "vm-stuff-attributes"
: 
: Attribute stuffing means transferring message attributes that
: only exist as setting in Emacs variables into actual data in the
: folder buffer.  This is done so that when auto-save happens
: attribute changes are auto-saved as well.
: 
:  > | > "vm-stuff-virtual-attributes"
: 
: vm-stuff-virtual-attributes checks if a message is a virtual
: message and if so calls vm-stuff-attributes on the underlying
: real message.
: 
:  > | > "vm-set-xxxx-flag"
: 
: vm-set-xxxx-flag is used to set all messaqge attribute flags.
: Besides setting/unsetting flags it adds records to VM's undo
: record list.
: 
:  > | > "vm-set-new-flag"
: 
: Since the message has been exposed, it isn't new anymore so the
: 'new' flag is being set to nil.
: 
:  > | > "vm-show-current-message"
: 
: vm-show-current-message unnarrows the folder buffer and sets the
: proper clipping region to display the message.
:  > | > "vm-scroll-forward"
: 
: SPC runs vm-scroll-forward, which exposed the current message,
: which was previously hidden (preview).
: 
:  > | > "call-interactively"
: 
: Someone pressed a key, probably the space bar.

--
Dan Barrett
address@hidden





reply via email to

[Prev in Thread] Current Thread [Next in Thread]