home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
System Booster
/
System Booster.iso
/
Archives
/
ForCLI
/
0Utils13.lha
/
0Utils
/
.KnownBugs
< prev
next >
Wrap
Text File
|
1995-04-10
|
3KB
|
87 lines
Some write Operations are not checked for success;
(most of them are directed to Output())
Shall this be fixed?
(20-03-95) Open returnes not only FileAddress, so Close, WriteLn
*FIXED* and ReadLn acted wrong. (eg. the examples did not work)
(20-03-95) TackOn can crash the machine if the resulting string
*FIXED* is more than 256 characters wide.
(20-03-95) TackOn always created RETURN_WARN
*FIXED*
(20-03-95) WriteLn had never worked
*FIXED*
(20-03-95) ReadLn always skipped its alst line and always produced
*FIXED* "Error: -1" warning
(20-03-95) VolName gets problems w/ "VOLUME/S,DEVICE/S" if the
Filename can not be locked (e.g. not existing name)
it then falls back to simple mode. this applies also
to relative names w/o volume or device flags set
(shall this be fixed?)
(20-03-95) GetWord produces wrong IoErrs
when executing inside scripts
S> FailAt 11
S> GetWord 0 a b
>> GetWord: bad number
S> if warn
S> why
>> The last command did not set a return code
S> endif
S> FailAt 10
it works however on commandline
C> GetWord 0 a b
>> GetWord: bad number
C> why
>> Last command failed because : bad number
can't fix this, 'cause I can't explain this behaviour.
(SEEMS TO BE A cbm SHELL PROBLEM!!!!!!!!!!!!!!)
(09-04-95) SRun strippes whitespaces and quotes in its commandline
*FIXED* ReadArgs does a nice job on /M arguments:
it just walks through all arguments up to EoL and concatenates
the found parts together ... (cbm ... no comment)
(09-04-95) IconXX breaks on scripts with a 'NOT FOUND' message
*FIXED* To fix this Problem, I have added a "SRunSX", a version of SRun,
which examines the commandfile, if it is enclosed w/ brackets
and explicitely calls Execute, if it is a Scriptfile.
the new version of Make_IconXX now explaces "Execute" w/ "SRunSX"
(... at least in v37-40 ...)
THIS WAS NOT THE FAULT OF IconXX! It seems, that CBM shell
gets confused, if a implicitly executed script is put into
quotes ... This can be proved very easily:
C> echo > t:xxx "echo hallo"
C> protect t:xxx +s
C> "t:xxx" ; beware of the quotes!
>> EXECUTE: Can't open t:xxx
>> object not found
In fact EXECUTE searches for 't:xxx ', not for 't:xxx'; so we
can see once more really nice behaviour of CBM shell. We can
show it clearer:
C> echo >"t:xxx " "echo cbm shell error - this is no script"
C> "t:xxx" ; beware of the quotes!
>> cbm shell error - this is no script
Can anybody tell me where that damned space is injected into the
search string (since the protection is checked with the right name)
(shall we perhaps examine the 1st argument and add an 'execute'?)
(10-04-95) SRun[SX] does not really do argument parsing,
so the following construct breaks:
> SRun COMMAND echo hallo
COMMAND: Unknown command
COMMAND failed returncode 10
SRun: object not found
This behaviour will probably NOT be fixed.
(since it is a fix for the problem, that things like
> SRun "Ram Disk:command"
are broken otherwise due to ReadArgs()...)