Re: [Lzip-bug] Request for help with Windows version of Plzip.

[Antonio Diaz Diaz]

wrotycz wrote:
> If it really is working as they say it would be reasonable to
> differentiate between hacked for mingw version of pread()/pwrite() and
> one shipped with Cygwin in this environment/platform. As far as I
> recollect there were people here who used plzip under cygwin.

As far as I can remember, plzip compiles on Cygwin without problems (using 
the pread/pwrite functions provided by Cygwin). So I'm not sure the change 
you propose is needed.


> /// not sure if cygwin defines __MSVCRT__

More reason to leave plzip unchanged until a real problem is found and a 
well documented fix is developed. What if I change the code and then it 
stops working on some systems?


> 3) This is just loosely related but I won't make new thread.
> Description of preadblock() in `$plzip/decompress.cc' states
> // Returns the number of bytes really read.
> // If (returned value < size) and (errno == 0), means EOF was reached.
>
> which is not accurate. Man for read()(https://linux.die.net/man/3/read)
> and pread()(https://linux.die.net/man/3/pread) is clear about:

In what sense the references you provide prove that the comment about 
plzip's preadblock is not accurate? 'preadblock' is not 'pread'.


> non-negative value returned indicates the number of bytes actually read,
> so it's not an error per se, you should read again or check errno if
> there was some interruption.

Which is exactly what preadblock does.


> User-Agent: GWP-Draft
> [...]
> X-Draft-Contenttype: text/html

Please, reconfigure your mail program (Mail-User-Agent) to send messages to 
this list as 'text/plain'. I find it annoying when I receive a multipart 
message whose 'text/plain' part is just the 'text/html' part without 
newlines but with html character codes (like "that&#39;s" instead of 
"that's"). Thanks.

Antonio.