Re: Test: zip-2.31 and unzip-5.52

[Charles Wilson <cygwin at cwilson dot fastmail dot fm>]

Charles D. Russell wrote:

> I use zip and gzip for backup files, where a bug is unlikely to be 
> detected before the problem is catastrophic.  Thus I like to stick to 
> old, well-tested versions, and am interested in understanding where 
> problems might arise.  I would have thought that the cygwin executable 
> would be the same as that obtained by taking the standard source and 
> running make.
>

Do you really think that every cygwin package compiles out-of-box with 
no changes?  Not even close to true!  In this case, there are a number 
of changes -- even in the "old, well-tested versions" that you've been 
happily using.  For your perusal, I've attached the patches -- of 
course, you could easily have downloaded the -src packages and extracted 
these yourself.


The changes boil down to three areas: (1) ensuring we do not use 
windows-isms when we should be using cygwin/posix-isms (2) ensuring that 
files are opened in binary, not text, mode (e.g. ensuring we don't use 
posix-isms when we should use windows-isms!), and (3) routine changes to 
the build system (enabling DESTDIR installs, building outside the source 
directory, .exe extensions on applications, etc)

> What is special about cygwin that requires patches?

Notwithstanding the 'use posix instead of windows' ethos, we ARE, 
undeniably, running on windows.  That's special.  Most 
non-autotool-based packages (like zip and unzip) which have been ported 
to windows, do so making assumptions about make/nmake, msvc-cl/gcc, etc. 
 These assumptions are usually wrong for cygwin, as we are a hybrid 
blend with posix and gcc features, but also some windows restrictions.

Why does cygwin require patches, indeed...

--
Chuck

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/