Hallo Martin,
Martin Neitzel wrote in <201812012042.wb1kg4dz025...@oker.escape.de>:
|Steffen Nurpmeso wrote:
|>|If I make a mistake, the prompt turns into 'ERROR# ?'
|>
|> This is actually a configurable compile-time option. Because this
|> MUA is a command line based application error messages that show
|> up on standard error fly by unseen fast, and pretty often.
|
|I never understood the need for this feature.
|
|At work, I handle 250+ new messages daily with heirloom-mailx as MUA.
This is a familiar number, about that here, too.
|I certainly cause a lot of errors but I am not aware of any error
|flying by unseen, ever. I make my error, get the message, done.
|
|What am I doing wrong?
I mean, nothing, but there are errors which are generated when you
look at mails, for example. All you want is that the message is
displayed, but the errors are recognized and will show up on
standard error, and you will not notice because the PAGER takes
the screen, or however you do it, anyway, the message is gone.
Also, for example, at the beginnig is saw a lot of errors when
opening old MBOXes, because of non-conforming From_ lines.
Short: i want to know what happens, one of the first programs
i have ever written was a s[imple]p[ager], and it had a help
buffer like less(1) has, but also a message buffer:
"SP COMMAND "
"SUMMARY\n"
"\n"
"[Caret (^): press CONTROL and X (or x) simultanously]\n"
"\n"
"h, H\t\tDisplay this help\n"
"m, M\t\tDisplay messages and errors (compile time switch)\n"
"q, Q\t\tQuit (help and error buffer: back to normal)\n"
It got all the errors and normal messages. You may think of this
as noisy, but i switched to the cwm(1) window manager last month,
and it was terrible that the binary package is silent even on
failing program invocations, you will not notice that the ssh
feature is not capable of dealing with all entries from
known_hosts etc.: you need to read the source code for that.
(From looking at that i seem to recall it can be compiled with
debugging, which then does some logging.)
|No, seriously: I'd like to hear of settings/scenarios where
|you'd miss an s-nail error, "pretty often" or not.
Most of the ones i see now are generated by running macro hooks.
But: as above.
| Martin
|
|[Sent from a system where I use and coarsely track s-nail versions.
Friendly, friendly. Of course the newest has lesser bugs than the
elder ones, but i am not a Linux kernel developer and therefore
i am not in the position to say "users must upgrade".
But a pity it is :)
|I use s-nail for about 12 mails per day there, with OPT_ERRORS=no.]
Good. It will always be an option, and maybe in a not too distant
future such functionality cannot only be a compile time option,
but also exist in a special .TEXT section of a compiled program,
with a strip(1) option where such things can be removed. That
would be cool, especially if one could do something like
'if(elf_has_text("name"))' to check availability at runtime.
Just so that normal programmers have a nice outcome of the myriads
of object file sections which bloom everywhere.
--End of <201812012042.wb1kg4dz025...@oker.escape.de>
--steffen
|
|Der Kragenbaer,The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)
