Re: [Bug-AUCTeX] auctex build fixes for xemacs

[David Kastrup]

Ralf Angeli <address@hidden> writes:

>> + test -r info/auctex.info
>> + echo info/auctex.info
>> + for i in auctex preview-latex
>> + test -r info/preview-latex.info
>> ghanima:/usr/local/lib/xemacs/xemacs-packages# echo $?
>> + echo 1
>> 1
>> ghanima:/usr/local/lib/xemacs/xemacs-packages# 
>>
>> This is because the return code of "test" becomes the return code of
>> "for". My patch was of course to hackish. Better do:
>>
>> if test -r ...; then
>>   echo ... > pkginfo;
>> fi
>>
>> "if" does not propagate the value of its conditional, in contrast to the
>> "conditional && then" construction.

A myth, unfortunately.

(info "(autoconf) Limitations of Builtins")

`if'

[...]

     There are shells that do not reset the exit status from an `if':

          $ if (exit 42); then true; fi; echo $?
          42

     whereas a proper shell should have printed `0'.  This is especially
     bad in Makefiles since it produces false failures.  This is why
     properly written Makefiles, such as Automake's, have such hairy
     constructs:

          if test -f "$file"; then
            install "$file" "$dest"
          else
            :
          fi

I don't remember off-bat what the test is supposed to do.  Maybe the
right operation would be

test ! -r whatever || do something with whatever

> David can probably say more about why && is used instead of if..then
> in the case at hand.

if..then without else is not always good with regard to exit status.
It may be that the wrong replacement was used here.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum