From luke at mewburn.net  Mon May 15 12:41:31 2023
From: luke at mewburn.net (Luke Mewburn)
Date: Mon, 15 May 2023 22:41:31 +1000
Subject: [Tcsh] [UCE] tcsh-6.24.08 is now available
In-Reply-To: <CADbbPnGWBS-YK6rkMJcEAyXFtOeo9ogpzyjHM-A-EPyKpHYYFA@mail.gmail.com>
References: <ZDZwJjCTtHQAOivN@mewburn.net>
 <CADbbPnGXKquzBcJoNXP-9ZHz=b=3G85cv8S9wWNhRWzuhV45oQ@mail.gmail.com>
 <CADbbPnFKKq+qjo0Z3ydZV_u4ZyoTbfX-mqhROqnTqSBT_ZrDLg@mail.gmail.com>
 <ZDnoxoyBVMN2bUIu@mewburn.net>
 <CADbbPnFmxD6+5rPwDRkMwNvuV7QLSjLkRZ1dwyrbB_BX6r74_g@mail.gmail.com>
 <CADbbPnGWBS-YK6rkMJcEAyXFtOeo9ogpzyjHM-A-EPyKpHYYFA@mail.gmail.com>
Message-ID: <ZGIoezevtXzaG1B6@mewburn.net>

On 23-04-15 11:54, Kimmo Suominen wrote:
  | On Sat, 15 Apr 2023 at 11:47, Kimmo Suominen <kim at netbsd.org> wrote:
  | > ------------------------
  | > > setsid -w sh -c 'tcsh -f -c "echo #no comment"'
  | > #no comment
  | > > setsid -w sh -c 'tcsh -f -c "echo #no comment"' < /dev/null
  | >
  | > ------------------------
  | >
  | > I understand the "$edit" case, but I don't understand the "#no comment" case.
  | 
  | But I guess it has to do with ed_Init() somehow as well, because
  | making the shell interactive with the "-i" flag makes a difference:
  | 
  | ------------------------
  | : beowulf:~; setsid -w sh -c 'tcsh -f -i -c "echo #no comment"' < /dev/null
  | #no comment
  | ------------------------

I've investigated this further.

The difference in behaviour is to be expected, and turns out is actually documented:
     When the shell's input is not a terminal, the character ?#? is taken to
     begin a comment.  Each ?#? and the rest of the input line on which it ap?
     pears is discarded before further parsing.

The mechanism in the testsuite to detect if stdin is a tty or not doesn't
quite work and always causes the test group to be skipped:
	AT_CHECK([if [ ! -t 0 ]; then exit 77; fi],, [Skipping comment tests])
probably due to m4 parsing/quoting of  if [.

I'll replaced this with something like:
	AS_IF([test -t 0], [
		# tests when stdin is a tty and not reading from file
	], [
		# tests when stdin is not a tty or reading from file
	])

I've fixed this up in pull request:
	https://github.com/tcsh-org/tcsh/pull/75

Then it works if I do:
	make check TESTSUITEFLAGS='-k Comments -k edit'
or
	make check TESTSUITEFLAGS='-k Comments -k edit' </dev/null

regards,
Luke.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://mailman.astron.com/pipermail/tcsh/attachments/20230515/db720d49/attachment.asc>

From luke at mewburn.net  Mon May 15 12:46:06 2023
From: luke at mewburn.net (Luke Mewburn)
Date: Mon, 15 May 2023 22:46:06 +1000
Subject: [Tcsh] [UCE] tcsh-6.24.08 is now available
In-Reply-To: <CADbbPnFmxD6+5rPwDRkMwNvuV7QLSjLkRZ1dwyrbB_BX6r74_g@mail.gmail.com>
References: <ZDZwJjCTtHQAOivN@mewburn.net>
 <CADbbPnGXKquzBcJoNXP-9ZHz=b=3G85cv8S9wWNhRWzuhV45oQ@mail.gmail.com>
 <CADbbPnFKKq+qjo0Z3ydZV_u4ZyoTbfX-mqhROqnTqSBT_ZrDLg@mail.gmail.com>
 <ZDnoxoyBVMN2bUIu@mewburn.net>
 <CADbbPnFmxD6+5rPwDRkMwNvuV7QLSjLkRZ1dwyrbB_BX6r74_g@mail.gmail.com>
Message-ID: <ZGIpjvH4aBWvXZnO@mewburn.net>

On 23-04-15 11:47, Kimmo Suominen wrote:
  | [Resending with "reply to all"]
  | 
  | On Sat, 15 Apr 2023 at 02:59, Luke Mewburn <luke at mewburn.net> wrote:
  | > I'll try and debug further why the original problem exists
  | > that requires those two tests to be skipped.
  | 
  | A proper fix would be nice, indeed. I just wanted to get
  | non-interactive builds working as before, and this was the easy way
  | out. :)

That's fine :) I did work out a proper fix - see other reply.


  | > Possibly related, but I've also noticed other weird edge cases about terminal
  | > setup on startup that may or may not be related, that could be worth debugging.
  | 
  | sh.c from line 1095 onwards:
  | - if (intty = isatty(SHIN)) returns zero, "edit" is unset
  | - intty |= intact; /* "tcsh -i" */
  | 
  | sh.c from line 1378 onwards:
  | - if (intty && !targinp) { ed_Init(); } /* "tcsh -c" sets targinp */
  | 
  | It is ed_Init() that would call GetTermCaps() during shell startup.
  | 
  | If GetTermCaps() has not been done (!GotTermCaps), telltc (and echotc
  | and settc) will also call GetTermCaps().

Thanks for the tip. I'll look into this further.

I'll investigate the issues with always setting up the terminal with
ed_Init(), versus adding another mechanism to do this explicitly that
doesn't require the telltc/echotc as the first command. Being able
to test various "terminal size" related behavior in the testsuite will
be good (e.g., ls-F column behavior).

regards,
Luke.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://mailman.astron.com/pipermail/tcsh/attachments/20230515/1e2c0c06/attachment.asc>

From merijn.brand at procura.nl  Thu May 25 09:49:38 2023
From: merijn.brand at procura.nl (H.Merijn Brand)
Date: Thu, 25 May 2023 09:49:38 -0000
Subject: [Tcsh] Tab used as completion from paste
Message-ID: <20230525101605.69e15b4e@pc09>

tab-completion is extremely useful, but very counterproductive
when a clipboard paste includes a tab.

I notice that when I copy text from a plain-text source (editor,
mail, browser, whatever) and I paste it in xterm/tcsh, any tabs
in the pasted buffer are interpreted as completion tab, causing
a lot of fails!

This is not new, but is there an option to *paste* tabs as-is?
Or translate tabs as spaces

This happens for Shift-Insert and Middle-Mouse-Button-Paste the
like.

-- 
H.Merijn Brand    <merijn.brand at procura.nl>   (088) 770 81 21
Procura  https://procura.nl  https://vrijbrp.nl  KvK 37140650
no HTML/LOGOs/DISCLAIMERS!          https://tux.nl/email.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <https://mailman.astron.com/pipermail/tcsh/attachments/20230525/871f766a/attachment.asc>