Alpha.07 in its basic configuration (e.g. no Kerberos or SSL support) seems
to have been OK on every platform tested, aside from the ever-increasing
assortment of picky-compiler warnings. So after some polishing-up and a
major version-number promotion, we have the first Beta test.
C-Kermit 10.0 Beta.01 changes since Alpha.07
Aside from the new version number, the changes since Alpha.07 are
modest.
|> New features
* Serial-port speed selections for 1000000, 2000000, 2500000,
3000000, 3500000, and 4000000 bps, suggested by Pali Roh�r. The new
high speeds have not been tested because I no longer have a
computer with serial port.
* New VDIRECTORY command with 1-letter synonym V, a convenient alias
for the DIRECTORY command (and a familiar one to old DEC-20 hands).
* New HELP COMPACT-SUBSTRING command to show the syntax for the
compact substring notation that was added in C-Kermit 9.0.300.
* HELP text improved for GREP, TRANSLATE, TOUCH /MODTIME, FREAD
/TRIM, FUNCTION DIRECTORIES, FUNCTION FILEINFO.
* New -DNODEPRECATED parameter for makefile KFLAGS to disinclude FTP,
Telnet, and RLOGIN clients plus WTMP logging, all "deprecated" now.
This saves only about 200K in the executable size.
* New Makefile targets: New "list" target to list all targets and
"count" target to show how many targets there are.
|> Bugs fixed
* ckufio.c: Bug fixed affecting -DNOUUCP builds (such as Mac OS) when
receiving a file in which C-Kermit created a bogus empty backup
file (e.g. bug.txt.~1~). Reported and diagnosed by MacPorts port
maintainer Lin Dongyu.
* ckcfns.c, ckcfn2.c, ckcfn3.c, ckcpro.w: SET SEND PACKET-LENGTH
applied to all packets when it should have applied only to Data
packets, now it does [28]details here.
* ckuus3.c: Bug fixed in S-Expression parser; "if (xx = 0)" should
have been "if (xx == 0)" (did you know that C-Kermit had a
[29]mini-LISP iterpreter built in?... [30]What day is Easter in
2033?)
* ckcfns.c: The 'debug(F110,"sipkt rp","",rp);' statement in sipkt()
had a bad argument list all these years.
* Extraneous line in "help function tablelook" removed.
* SET HOST command switch /TIMEOUT: was removed because the
supporting code was never written.
* Top-level DEBUG command that was added in C-Kermit 9.0 was removed
because it never did anything and there was no help text for it.
|> Compiler warnings hushed
... most of them issued only by the very picky Mac OS Clang compiler:
ckcftp.c: changed a long if-else-if.. (dangling else) sequence to a
switch
in hopes of silencing some "Pointers are not assignment-compatible"
and "Argument #1 is not the correct type" warnings on a platform I
don't have access to (HP-UX).
ckcpro.w: fixed 1 dangling else
ckuusr.c: fixed 7 dangling elses
ckuus2.c: fixed 1 dangling else
ckuus3.c: fixed 10 dangling elses
ckuus4.c: fixed 67 dangling elses
ckuus5.c: fixed 1 dangling else
ckuus6.c: fixed 1 dangling else
ckuus7.c: fixed 1 dangling else
ckuusx.c: Fixed several dummy ck_curpos() declarations to suppress
warnings.
ckuusx.c: fixed 2 extraneous parentheses and 1 dangling else
ckcfn2.c: fixed 1 dangling else
ckutio.c: fixed 1 dangling else
makefile: added -Wdeprecated-declarations to macos target (for
logwtmp)
"Dangling else" is (or was) perfectly legal when used correctly; the C
language was originally designed for use by people who knew what they
were doing and was thoroughly documented by its authors, Kernighan and
Plauger, in a book (two really: original C and ANSI C). It makes no
sense to keep changing this or any other programming language out from
under all the millions of programs that have already been written in
it.
For greater detail see https://www.kermitproject.org/ckupdates.html .
C-Kermit 10.0 Beta.01 16 May 2022
Alpha.07 in its basic configuration (e.g. no Kerberos or SSL support) seems to have been OK on every platform tested, aside from the ever-increasing assortment of picky-compiler warnings. So after some polishing-up and a major version-number promotion, we have the first Beta test.
C-Kermit 10.0 Beta.01 changes since Alpha.07
Aside from the new version number, the changes since Alpha.07 are modest.
|> New features
|> Bugs fixed
|> Compiler warnings hushed
... most of them issued only by the very picky Mac OS Clang compiler:
ckcftp.c: changed a long if-else-if.. (dangling else) sequence to a switch
ckctel.c: changed: tn_ssbopt(TELOPT_NEWENVIRON,TELQUAL_SEND,request, strlen(request));
tn_ssbopt(TELOPT_NEWENVIRON,TELQUAL_SEND,(CHAR )request, strlen((CHAR )request)); / 2022-01-27 /
in hopes of silencing some "Pointers are not assignment-compatible" and "Argument #1 is not the correct type" warnings on a platform I don't have access to (HP-UX).
ckcpro.w: fixed 1 dangling else
ckuusr.c: fixed 7 dangling elses
ckuus2.c: fixed 1 dangling else
ckuus3.c: fixed 10 dangling elses
ckuus4.c: fixed 67 dangling elses
ckuus5.c: fixed 1 dangling else
ckuus6.c: fixed 1 dangling else
ckuus7.c: fixed 1 dangling else
ckuusx.c: Fixed several dummy ck_curpos() declarations to suppress warnings.
ckuusx.c: fixed 2 extraneous parentheses and 1 dangling else
ckcfn2.c: fixed 1 dangling else
ckutio.c: fixed 1 dangling else
makefile: added -Wdeprecated-declarations to macos target (for logwtmp)
"Dangling else" is (or was) perfectly legal when used correctly; the C language was originally designed for use by people who knew what they were doing and was thoroughly documented by its authors, Kernighan and Plauger, in a book (two really: original C and ANSI C). It makes no sense to keep changing this or any other programming language out from under all the millions of programs that have already been written in it.