]> diplodocus.org Git - nmh/log
nmh
13 years agoSuppress warning from test-manpages that groff can't break a line if
David Levine [Sat, 2 Feb 2013 17:55:23 +0000 (11:55 -0600)]
Suppress warning from test-manpages that groff can't break a line if
the user's prefix is too long.

13 years agoThe Solaris compiler was right about unreachable statements.
David Levine [Sat, 2 Feb 2013 16:54:32 +0000 (10:54 -0600)]
The Solaris compiler was right about unreachable statements.
Fixed all of them in uip/.

13 years agoAdded l modifier to X printf conversions of wchar_t's.
David Levine [Sat, 2 Feb 2013 16:21:14 +0000 (10:21 -0600)]
Added l modifier to X printf conversions of wchar_t's.

13 years agoRemoved unused vec[0] relic from before argsplit conversion.
David Levine [Sat, 2 Feb 2013 16:18:24 +0000 (10:18 -0600)]
Removed unused vec[0] relic from before argsplit conversion.

13 years agoAdded trailing newline to sed input in mhmail when processing
David Levine [Sat, 2 Feb 2013 16:12:57 +0000 (10:12 -0600)]
Added trailing newline to sed input in mhmail when processing
-headerfield argument, to squelch complain from Solaris sed.

13 years agoSet MHTMPDIR to try to keep all test-related files together.
David Levine [Sat, 2 Feb 2013 15:59:28 +0000 (09:59 -0600)]
Set MHTMPDIR to try to keep all test-related files together.

13 years agoMore test-mhmail probing.
David Levine [Sat, 2 Feb 2013 15:58:20 +0000 (09:58 -0600)]
More test-mhmail probing.

13 years agoTry this test-mhmail probe.
David Levine [Sat, 2 Feb 2013 15:23:49 +0000 (09:23 -0600)]
Try this test-mhmail probe.

13 years agoMore test-mhmail probing.
David Levine [Sat, 2 Feb 2013 14:43:32 +0000 (08:43 -0600)]
More test-mhmail probing.

13 years agoModified test-mhmail probe.
David Levine [Sat, 2 Feb 2013 14:26:43 +0000 (08:26 -0600)]
Modified test-mhmail probe.

13 years agoAdded temporary probe to test-mhmail.
David Levine [Sat, 2 Feb 2013 14:15:15 +0000 (08:15 -0600)]
Added temporary probe to test-mhmail.

13 years agoConvert post over to use new argsplit code.
Ken Hornstein [Sat, 2 Feb 2013 06:10:06 +0000 (01:10 -0500)]
Convert post over to use new argsplit code.

13 years agoHandle the case where (on Solaris) the width of U+2019 is 2 (which seems
Ken Hornstein [Sat, 2 Feb 2013 05:48:54 +0000 (00:48 -0500)]
Handle the case where (on Solaris) the width of U+2019 is 2 (which seems
TOTALLY bizarre to me, but there you have it).

13 years agoRework multibyte test so it will calculate the width of all of the
Ken Hornstein [Sat, 2 Feb 2013 05:33:51 +0000 (00:33 -0500)]
Rework multibyte test so it will calculate the width of all of the
Unicode characters we are using.

13 years agoIn setup_test(), set PATH after install.
David Levine [Sat, 2 Feb 2013 05:24:27 +0000 (23:24 -0600)]
In setup_test(), set PATH after install.

13 years agoAdded set -x to test-ali.
David Levine [Sat, 2 Feb 2013 05:12:09 +0000 (23:12 -0600)]
Added set -x to test-ali.

13 years agoMaybe the mkdir in setup_test() needs a -p?
David Levine [Sat, 2 Feb 2013 04:52:00 +0000 (22:52 -0600)]
Maybe the mkdir in setup_test() needs a -p?

13 years agoMore test-ali diagnostics for Solaris.
David Levine [Sat, 2 Feb 2013 04:39:49 +0000 (22:39 -0600)]
More test-ali diagnostics for Solaris.

13 years agoUse test -t 1 instead of test -w in test-pick.
David Levine [Sat, 2 Feb 2013 04:25:51 +0000 (22:25 -0600)]
Use test -t 1 instead of test -w in test-pick.

13 years agoMore setup_test() and test-ali attempts on Solaris.
David Levine [Sat, 2 Feb 2013 04:24:48 +0000 (22:24 -0600)]
More setup_test() and test-ali attempts on Solaris.

13 years agoRemoved the ls of folder from setup_test().
David Levine [Sat, 2 Feb 2013 00:46:02 +0000 (18:46 -0600)]
Removed the ls of folder from setup_test().

13 years agoTry full path to folder in setup_test().
David Levine [Sat, 2 Feb 2013 00:35:13 +0000 (18:35 -0600)]
Try full path to folder in setup_test().

13 years agoMaybe the PATH needs to be exported again in setup_test()?
David Levine [Sat, 2 Feb 2013 00:16:26 +0000 (18:16 -0600)]
Maybe the PATH needs to be exported again in setup_test()?

13 years agoSo folder does exist the first time through setup_test().
David Levine [Fri, 1 Feb 2013 23:37:02 +0000 (17:37 -0600)]
So folder does exist the first time through setup_test().
Let's see its stdout.

13 years agoSuppressed display of info message when test-eom-align isn't
David Levine [Fri, 1 Feb 2013 23:20:37 +0000 (17:20 -0600)]
Suppressed display of info message when test-eom-align isn't
run with stdout to a terminal.

13 years agoTemporarily added ls of folder to setup_test().
David Levine [Fri, 1 Feb 2013 23:19:40 +0000 (17:19 -0600)]
Temporarily added ls of folder to setup_test().

13 years agoIn setup_test(), create the Mail directory if it doesn't
David Levine [Fri, 1 Feb 2013 21:08:32 +0000 (15:08 -0600)]
In setup_test(), create the Mail directory if it doesn't
already exist.

13 years agoIgnore the return status of diff, so we don't exit if diff fails when using
Ken Hornstein [Fri, 1 Feb 2013 20:41:00 +0000 (15:41 -0500)]
Ignore the return status of diff, so we don't exit if diff fails when using
set -e.

13 years agoAdd some diagnostics to determine what is going wrong.
Ken Hornstein [Fri, 1 Feb 2013 20:18:47 +0000 (15:18 -0500)]
Add some diagnostics to determine what is going wrong.

13 years agoIt looks like the first test fails on Solaris because it
David Levine [Fri, 1 Feb 2013 20:08:45 +0000 (14:08 -0600)]
It looks like the first test fails on Solaris because it
can't find folder.  So, added a test/sleep before its first
use, in setup_test().

13 years agoIf the multibyte width test fails, dump the values of wcwidth() for the
Ken Hornstein [Fri, 1 Feb 2013 20:03:16 +0000 (15:03 -0500)]
If the multibyte width test fails, dump the values of wcwidth() for the
complete BMP.

13 years agoAdded diagnostic to setup_test() to try to figure out
David Levine [Fri, 1 Feb 2013 19:38:36 +0000 (13:38 -0600)]
Added diagnostic to setup_test() to try to figure out
why the first test fails on the Solaris buildbot.

13 years agoRestored test-ali to its prior condition. It failed on the
David Levine [Fri, 1 Feb 2013 19:36:23 +0000 (13:36 -0600)]
Restored test-ali to its prior condition.  It failed on the
Solaris buildbot because it was the first test.

13 years agoMake sure we created the test file even if we don't have iconv support;
Ken Hornstein [Fri, 1 Feb 2013 19:06:49 +0000 (14:06 -0500)]
Make sure we created the test file even if we don't have iconv support;
it's needed for two tests.

13 years agoMove ICONV_ENABLED so it always gets initialized even if the header check
Ken Hornstein [Fri, 1 Feb 2013 18:42:51 +0000 (13:42 -0500)]
Move ICONV_ENABLED so it always gets initialized even if the header check
for iconv.h fails.

13 years agoIf we don't have iconv, skip the multibyte test that requires character
Ken Hornstein [Fri, 1 Feb 2013 18:25:17 +0000 (13:25 -0500)]
If we don't have iconv, skip the multibyte test that requires character
set conversion.

13 years agoFinally switch over from vfork() to fork() completely.
Ken Hornstein [Fri, 1 Feb 2013 17:30:45 +0000 (12:30 -0500)]
Finally switch over from vfork() to fork() completely.

13 years agoMore code converted to use argsplit().
Ken Hornstein [Fri, 1 Feb 2013 17:11:59 +0000 (12:11 -0500)]
More code converted to use argsplit().

13 years agoFix configure tests for -Wfoo option support.
Lyndon Nerenberg [Fri, 1 Feb 2013 15:25:44 +0000 (07:25 -0800)]
Fix configure tests for -Wfoo option support.

Drooped support for -Wunused-macros and -Wno-unused-macros.  The
warnings serve no useful purpose, and causing subsequent configure
checks for -Wfoo flags to get messed up.

Split each test of -Wfoo into a separate check. Include -Werror in
the individual flag tests to ensure the test compile fails if the
other flag isn't supported.  Without this, clang issues a diagnostic
message about the flag being ignored, but otherwise carries on, fooling
configure into thinking the flag is in fact supported.

13 years agoMerge branch 'master' into savannah
Lyndon Nerenberg [Fri, 1 Feb 2013 12:44:29 +0000 (04:44 -0800)]
Merge branch 'master' into savannah

13 years agoUpdate manpages to use .TP for tagged paragraphs (part I).
Lyndon Nerenberg [Fri, 1 Feb 2013 08:20:44 +0000 (00:20 -0800)]
Update manpages to use .TP for tagged paragraphs (part I).

This replaces .fc/.nf and friends with the man(7) .TP macro.

This commit covers about half of the man pages.  The rest will
follow in a couple of days.

13 years agoUpdate manpages to use .TP for tagged paragraphs (part I).
Lyndon Nerenberg [Fri, 1 Feb 2013 08:20:44 +0000 (00:20 -0800)]
Update manpages to use .TP for tagged paragraphs (part I).

This replaces .fc/.nf and friends with the man(7) .TP macro.

This commit covers about half of the man pages.  The rest will
follow in a couple of days.

13 years agoUpdate the manpage template documentation.
Lyndon Nerenberg [Fri, 1 Feb 2013 05:28:51 +0000 (21:28 -0800)]
Update the manpage template documentation.

Update the section ordering to follow the UNIX order.  This
pulls nmh-specific sections into the body of the manpage, and
pushes FILES, SEE ALSO, and BUGS to the end.

Clarifies a few other minor points.

13 years agoMore probing of test-ali on Solaris: exit before the
David Levine [Fri, 1 Feb 2013 04:32:26 +0000 (22:32 -0600)]
More probing of test-ali on Solaris:  exit before the
setup_test call, which has the folder invocation.

13 years agoOK, test-pick should be good on Solaris now.
David Levine [Fri, 1 Feb 2013 04:30:36 +0000 (22:30 -0600)]
OK, test-pick should be good on Solaris now.

13 years agoAdded exit to test-ali, temporarily, to isolate failure on
David Levine [Fri, 1 Feb 2013 03:27:46 +0000 (21:27 -0600)]
Added exit to test-ali, temporarily, to isolate failure on
Solaris.

13 years agoMore homing in on the test-pick warning. Also, removed
David Levine [Fri, 1 Feb 2013 03:24:53 +0000 (21:24 -0600)]
More homing in on the test-pick warning.  Also, removed
warning to user that the tty test is being skipped, it's
not that important.

13 years agoRemoved an used sed invocation from test-mhmail. I don't
David Levine [Fri, 1 Feb 2013 02:10:59 +0000 (20:10 -0600)]
Removed an used sed invocation from test-mhmail.  I don't
know if it was causing the warnings on Solaris, we'll see.

13 years agoFix some warnings for Solaris.
Ken Hornstein [Fri, 1 Feb 2013 02:04:43 +0000 (21:04 -0500)]
Fix some warnings for Solaris.

13 years agoit looks like the write to /dev/tty is causing the warning from
David Levine [Fri, 1 Feb 2013 02:04:50 +0000 (20:04 -0600)]
it looks like the write to /dev/tty is causing the warning from
test-pick on Solaris, here's an attempt to squelch that.

13 years agoRemoved sizes from mhlist output. Apparently tar files are
David Levine [Fri, 1 Feb 2013 01:44:47 +0000 (19:44 -0600)]
Removed sizes from mhlist output.  Apparently tar files are
so much smaller on Solaris that they cause the check of its
output to break.  The test also compares the contents of
the tar file, so it's still adequate.

13 years agoRedirected stderr to /dev/null in the test for presence of
David Levine [Fri, 1 Feb 2013 01:43:09 +0000 (19:43 -0600)]
Redirected stderr to /dev/null in the test for presence of
/dev/tty.  Solaris warns that it can't create the device.
This may not fix it, but it should help isolate which part of
the test is causing the warning.

13 years agoSet the locale for one of the "pick" invocations to UTF-8, so we can decode
Ken Hornstein [Fri, 1 Feb 2013 01:01:27 +0000 (20:01 -0500)]
Set the locale for one of the "pick" invocations to UTF-8, so we can decode
the UTF-8 header correctly in the test.

13 years agoUse cmp(1) to compare acheieved and expected test results.
Lyndon Nerenberg [Thu, 31 Jan 2013 11:56:55 +0000 (03:56 -0800)]
Use cmp(1) to compare acheieved and expected test results.

'diff -c' on Solaris 11 generate a spurious 'No differences encountered'
message (on stdout, no less!) when the compared files are identical. This
changes the test logic to compare with 'cmp -s', and then perform a
'diff -c' if the results don't match expectations.

13 years agoWrap 'pick -subject foobar 13' test with set +e/-e.
Lyndon Nerenberg [Thu, 31 Jan 2013 09:44:03 +0000 (01:44 -0800)]
Wrap 'pick -subject foobar 13' test with set +e/-e.

The command was failing silently on FreeBSD, leading to
premature script termination.

13 years agoChange the MD5FMT macro in configure.ac to not emit filename.
Lyndon Nerenberg [Thu, 31 Jan 2013 07:55:26 +0000 (23:55 -0800)]
Change the MD5FMT macro in configure.ac to not emit filename.

13 years agoHow hard can it be to write a portable shell function that does nothing?
Lyndon Nerenberg [Thu, 31 Jan 2013 07:07:17 +0000 (23:07 -0800)]
How hard can it be to write a portable shell function that does nothing?

Pretty hard, apparently :-P

13 years agoHush progress indicator during tests when stdout is not a tty.
Lyndon Nerenberg [Thu, 31 Jan 2013 05:45:16 +0000 (21:45 -0800)]
Hush progress indicator during tests when stdout is not a tty.

This eliminates 100 lines of logfile spam during buildbot runs.

13 years agoMerge branch 'master' of ssh://git.sv.gnu.org/srv/git/nmh into nmh
Lyndon Nerenberg [Thu, 31 Jan 2013 03:58:16 +0000 (19:58 -0800)]
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/nmh into nmh

13 years agoIPPROTO_TCP requires <netinet/in.h>
Lyndon Nerenberg [Thu, 31 Jan 2013 03:57:22 +0000 (19:57 -0800)]
IPPROTO_TCP requires <netinet/in.h>

At least on BSD it does.

13 years agoMore changes to use argsplit() for helper programs.
Ken Hornstein [Wed, 30 Jan 2013 20:47:32 +0000 (15:47 -0500)]
More changes to use argsplit() for helper programs.

13 years agoSwitch the editor invocation over to use argsplit().
Ken Hornstein [Tue, 29 Jan 2013 20:58:34 +0000 (15:58 -0500)]
Switch the editor invocation over to use argsplit().

13 years agoThe start of the use of argsplit() to process command strings.
Ken Hornstein [Tue, 29 Jan 2013 20:30:36 +0000 (15:30 -0500)]
The start of the use of argsplit() to process command strings.

13 years agoChanged m_getfld() buffer size to 4096 from 8192, to benefit
David Levine [Tue, 29 Jan 2013 03:38:39 +0000 (21:38 -0600)]
Changed m_getfld() buffer size to 4096 from 8192, to benefit
scan(1).  We could consider using a larger buffer size when
readying message bodies.  But I expect any benefit would be
overshadowed by the way parse_mime() currently reads MIME
messages:  it reopens the file for each part.

13 years agoAdded m_getfld_track_filepos() for callers to indicate that they
David Levine [Tue, 29 Jan 2013 03:20:22 +0000 (21:20 -0600)]
Added m_getfld_track_filepos() for callers to indicate that they
may interleave fseek*/ftell* calls with m_getfld() calls.  This
eliminates the need for m_getfld() to call fseek() when called
by inc(1) and scan(1).

13 years agoAdded scan_reset_m_getfld_state(), for msh only. It needs to
David Levine [Tue, 29 Jan 2013 01:12:17 +0000 (19:12 -0600)]
Added scan_reset_m_getfld_state(), for msh only.  It needs to
call that so that it can support more than one operation on its
mailbox.

13 years agoMerge branch 'm_getfld'
David Levine [Sun, 27 Jan 2013 16:37:54 +0000 (10:37 -0600)]
Merge branch 'm_getfld'

13 years agoIn the sortm warning message, say "file modification time"
David Levine [Sun, 27 Jan 2013 16:11:10 +0000 (10:11 -0600)]
In the sortm warning message, say "file modification time"
instead of "file mtime", per Paul F.'s suggestion.

13 years agoChanged sortm warning message from an amonish() to and advise(),
David Levine [Sun, 27 Jan 2013 15:34:58 +0000 (09:34 -0600)]
Changed sortm warning message from an amonish() to and advise(),
and included the remedy of using the file mtime, per suggestions
of Norm and Ralph.

13 years agoFixed a couple of calls to m_getfld(): they were in while loops
David Levine [Sun, 27 Jan 2013 14:37:31 +0000 (08:37 -0600)]
Fixed a couple of calls to m_getfld():  they were in while loops
and the loop bodies were missing brackets.

13 years agoDeprecated msh(1).
David Levine [Sat, 26 Jan 2013 15:42:25 +0000 (09:42 -0600)]
Deprecated msh(1).

13 years agoIn build_nmh's output from the test suite: retain
David Levine [Sat, 26 Jan 2013 15:00:48 +0000 (09:00 -0600)]
In build_nmh's output from the test suite:  retain
colorization and list any test failures.

13 years agoEnabled colorized test output (PASS/FAIL/summary) by default
David Levine [Sat, 26 Jan 2013 14:52:50 +0000 (08:52 -0600)]
Enabled colorized test output (PASS/FAIL/summary) by default
via "color-tests" automake option.  To disable at runtime,
either set your AM_COLOR_TESTS environment variable to "no",
or provide that on the make check command line.

13 years agoReworked the new documentation in m_getfld.c.
David Levine [Sat, 26 Jan 2013 04:21:53 +0000 (22:21 -0600)]
Reworked the new documentation in m_getfld.c.

13 years agoHandle the case in the test suite where wcwidth() can return different
Ken Hornstein [Sat, 26 Jan 2013 03:22:46 +0000 (22:22 -0500)]
Handle the case in the test suite where wcwidth() can return different
values for combining Unicode codepoints on different operating systems.

13 years agoFix the format engine so it properly keeps track of multibyte characters
Ken Hornstein [Thu, 24 Jan 2013 21:05:36 +0000 (16:05 -0500)]
Fix the format engine so it properly keeps track of multibyte characters
and column positions when processing components and other functions.

When working on this code I realized we have nothing that checks to make
sure right justification of format function output works properly, so a
test for right justification is also included.

13 years agoAdded functions for inc, msh, rcvtty, and scan to access
David Levine [Thu, 24 Jan 2013 02:24:21 +0000 (20:24 -0600)]
Added functions for inc, msh, rcvtty, and scan to access
scansbr's gstate instance so it can be hidden, except for
msh's uses of m_getfld().

13 years agoFix a bug leading to message corruption with inc when incorporation more
Ken Hornstein [Mon, 21 Jan 2013 15:32:37 +0000 (10:32 -0500)]
Fix a bug leading to message corruption with inc when incorporation more
than one message at a time.

When I reworked the scan() code to remove the use of stdio internals,
two code paths were created.  The inc code path consumed another buffer
from the nxtbuf list for processing the message body.  At the end of
scan() there was code to push back all of the consumed buffers to the
nxtbuf list.  The extra buffer consumed in the inc code path wasn't
being pushed back, and as a result the nxtbuf list would end up with
duplicate pointers in it.  This would result in two components sharing
the same text buffer, which caused the message corruption.

When I looked at this closer, I realized that there was no reason to
push those buffers back to the list; nxtbuf gets reset to compbuffers
at the start of every call to scan(), and nxtbuf isn't used after the
buffers are pushed back.  Some researched revealed that this was the
same back in the original MH code.  I decided it was simplest to simply
delete the buffer recycling code.
(cherry picked from commit 22d2acba60374851c291260ff6aa716e9bf44a6d)

13 years agoBrought over changes to test-pop from master to m_getfld branch.
David Levine [Thu, 24 Jan 2013 01:42:56 +0000 (19:42 -0600)]
Brought over changes to test-pop from master to m_getfld branch.

13 years agoAdded scan and inc from mbox with multiple messages.
David Levine [Sun, 20 Jan 2013 14:38:18 +0000 (08:38 -0600)]
Added scan and inc from mbox with multiple messages.
(cherry picked from commit c4c53777458bfcfd288d45b7708e52eb47c56405)

13 years agoAdded test-msg-buffer-boundaries.
David Levine [Fri, 18 Jan 2013 02:58:25 +0000 (20:58 -0600)]
Added test-msg-buffer-boundaries.
(cherry picked from commit dbd0fa0a56a222bd118fea627eba53e5d9811d01)

13 years agoBrought over changes to test-pop from m_getfld branch.
David Levine [Thu, 24 Jan 2013 01:20:40 +0000 (19:20 -0600)]
Brought over changes to test-pop from m_getfld branch.

13 years agoChanged m_getfld() to automatically allocate a new state
David Levine [Wed, 23 Jan 2013 03:02:49 +0000 (21:02 -0600)]
Changed m_getfld() to automatically allocate a new state
instance as needed.  Callers no longer use
m_getfld_state_init().

13 years agoRemoved unnecessary m_getfld_state_destroy() from scan.c.
David Levine [Tue, 22 Jan 2013 23:08:21 +0000 (17:08 -0600)]
Removed unnecessary m_getfld_state_destroy() from scan.c.

13 years agoReworked enter_getfld() and Peek().
David Levine [Tue, 22 Jan 2013 23:07:47 +0000 (17:07 -0600)]
Reworked enter_getfld() and Peek().

13 years agoFix a bug leading to message corruption with inc when incorporation more
Ken Hornstein [Mon, 21 Jan 2013 15:32:37 +0000 (10:32 -0500)]
Fix a bug leading to message corruption with inc when incorporation more
than one message at a time.

When I reworked the scan() code to remove the use of stdio internals,
two code paths were created.  The inc code path consumed another buffer
from the nxtbuf list for processing the message body.  At the end of
scan() there was code to push back all of the consumed buffers to the
nxtbuf list.  The extra buffer consumed in the inc code path wasn't
being pushed back, and as a result the nxtbuf list would end up with
duplicate pointers in it.  This would result in two components sharing
the same text buffer, which caused the message corruption.

When I looked at this closer, I realized that there was no reason to
push those buffers back to the list; nxtbuf gets reset to compbuffers
at the start of every call to scan(), and nxtbuf isn't used after the
buffers are pushed back.  Some researched revealed that this was the
same back in the original MH code.  I decided it was simplest to simply
delete the buffer recycling code.

13 years agoPlugged small memory leak in m_unknown().
David Levine [Mon, 21 Jan 2013 14:42:20 +0000 (08:42 -0600)]
Plugged small memory leak in m_unknown().

13 years agoReplaced hard-coded value with MAX_DELIMITER_SIZE in m_getfld.c.
David Levine [Mon, 21 Jan 2013 04:22:17 +0000 (22:22 -0600)]
Replaced hard-coded value with MAX_DELIMITER_SIZE in m_getfld.c.

13 years agoRemoved the static m_getfld() state instance and replaced
David Levine [Mon, 21 Jan 2013 03:27:54 +0000 (21:27 -0600)]
Removed the static m_getfld() state instance and replaced
with functions for callers to create and destroy their own
instances.  This also allows inc, msh, and pick to call
m_getfld() with an initial file stream position other than 0.

There is now a global state instance so that scan can call
m_unknown() on maildrop files.  It'd be nice to get rid of
that.

13 years agoFix memory leak in arglist code
Ken Hornstein [Mon, 21 Jan 2013 03:23:58 +0000 (22:23 -0500)]
Fix memory leak in arglist code

13 years agoAdded scan and inc from mbox with multiple messages.
David Levine [Sun, 20 Jan 2013 14:38:18 +0000 (08:38 -0600)]
Added scan and inc from mbox with multiple messages.

13 years agoUpdated test-inc-scanout on master to match version of
David Levine [Sun, 20 Jan 2013 13:49:58 +0000 (07:49 -0600)]
Updated test-inc-scanout on master to match version of
m_getfld branch.

13 years agoAdded a FILE * to the m_getfld_state struct so that most
David Levine [Sat, 19 Jan 2013 14:33:38 +0000 (08:33 -0600)]
Added a FILE * to the m_getfld_state struct so that most
functions don't need to pass it explicitly.  Note that
m_getfld() must still be passed the FILE * because some
callers, e.g., those in mhparse.c, reopen their input file
and therefore pass different FILE *'s for the same input.

13 years agoMoved all file static data in m_getfld.c into the
David Levine [Sat, 19 Jan 2013 14:00:03 +0000 (08:00 -0600)]
Moved all file static data in m_getfld.c into the
m_getfld_state struct.  Changed MSG_INPUT_SIZE from
hard-coded number to BUFSIZ.

13 years agoAdded check of -pack to test-pop.
David Levine [Sat, 19 Jan 2013 12:54:10 +0000 (06:54 -0600)]
Added check of -pack to test-pop.

13 years agoFixed test-header to work on Cygwin, which has a BUFSIZ of
David Levine [Sat, 19 Jan 2013 01:19:00 +0000 (19:19 -0600)]
Fixed test-header to work on Cygwin, which has a BUFSIZ of
only 1024.  That caused an error message to be truncated.

13 years agoBeginning of implementation of new argsplit() function to handle arguments
Ken Hornstein [Fri, 18 Jan 2013 21:03:26 +0000 (16:03 -0500)]
Beginning of implementation of new argsplit() function to handle arguments
to profile-defined commands.

13 years agoRemoved all traces of m_getfld() FLDEOF and BODYEOF states.
David Levine [Fri, 18 Jan 2013 03:58:50 +0000 (21:58 -0600)]
Removed all traces of m_getfld() FLDEOF and BODYEOF states.
They were never set by m_getfld(), so all caller code that
supported them was completely unexercised.

13 years agoMinimized message files reads in m_getfld() when the caller
David Levine [Fri, 18 Jan 2013 03:46:54 +0000 (21:46 -0600)]
Minimized message files reads in m_getfld() when the caller
interleaves calls to fseek().  Added check of handling of
long header field name without a colon to test-header.

13 years agoAdded test-msg-buffer-boundaries.
David Levine [Fri, 18 Jan 2013 02:58:25 +0000 (20:58 -0600)]
Added test-msg-buffer-boundaries.