Mail Archives: djgpp/2003/03/10/15:15:05
Charles Sandmann <sandmann AT clio DOT rice DOT edu> wrote:
>> I'ts said in readme.DJGPP that one needs to have latest update of
>> djdev203.zip.
> Perhaps it would be best if this specifically said you need to have
> updated your djdev203 since August 2002, so that sys/djtypes.h
.. or whatever else is needed to make sure the user, reading that
note, can go and check whether they have to get a fresh djdev203.zip,
and how to make sure the one they get is the one they need. A
"patchlevel" or similar would do --- to be displayed in the
manifest/djdev203.ver or the DSM file, maybe. And/or by go32-v2.exe
as part of its output.
"The latest" is not a useful description of the needed version,
because it's impossible to check what it means without external help.
> But I personally don't think this would fix the problem in many cases,
> since many people don't read documentation.
We really can't do anything about those guys --- nothing short of
switching to a M$ Installer like scheme where the thing simply won't
install at all unless you spent at least 1 minute staring at some
text, and then click "Yeah, OK, I have read it" afterwards. While at,
let's add some tricky test of their command of the English language,
too, so we're sure they *understood* what they read. ;->
But that shouldn't stop us from trying to help those who are actually
R'ing TFM as best we can.
> Only including this file in the GCC 3.x distributions and
> conditionally installing would fix the problem. But we don't have a
> conditional installer.
There is this 'pakke' thing... not bad at all from my limited
experience with it, but currently not an official requirement.
>> Also. I completely agree that it would perhaps be better to name updated
>> version as 2.04 ...
> I'm not convinced this would fix anything. Today, if you go download
> all the components it works. It only breaks if you use an old djdev
> with a new GCC. But anyone who has an old djdev probably has an old
> GCC, and has decided to move to a newer version. They could just as
> easily miss upgrading djdev if it were renumbered.
Not nearly as easily as if it shows no outward symptom of having been
updated in the first place (same name, time stamp might just as well
be garbled as actually new, ...)
> So, how are users deciding to upgrade GCC without doing the full
> upgrade? Should the zip picker, FAQ, readmes warn about GCC 3.x
> compatibility?
I think that may be necessary. The zip picker and readme.1st in
particular.
--
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.
- Raw text -