David Pilgram wrote:

> Dear J�rgen,
> 
> Line 397, according to nedit is  "else".  To put it in context, here are
> the lines before.
> 
> 


So in this case you were not using the current sources.... and from your 
example I see you were not.

In the current sources (one CVS) you will find a comment in line 396 and a 
waitpid() call in 397. So if you don't have the latest release (see the 
beginning of the file to find the release number, here I have 2.16) you 
won't find my changes (or even any changes).

However, after getting the latest file your should give it a try before you 
change the code as told here because I think this will already do it.

Btw. don't check out CVS tag 2_0 because there won't be any changes and the 
CVS tarball is created in the morning (around 9 am CET) so maybe you got an 
old one if you tried this. The the "cvs snap" web page on the xcmail 
homepage for the correct CVS syntax for getting the latest release. "cvs 
co" won't update sources if you already checked it out. You must use "cvs 
update".

J�rgen


> 
> int io_gzip::close(io_data *f)
> {
>     io_gdata *r=(io_gdata*)f;
>     int ret=0;
>     if (r->mode=='r' || r->mode=='n')
>     {
>         r->error=pclose(((io_gdata*)r)->file)||r->error;
>         ret=r->error;
>     }
>     else if (r->mode=='w')
>     {
>         r->error=fclose(((io_gdata*)r)->file)||r->error;
>         ret=r->error;
> 
>         struct stat st;
>         stat(r->filename,&st);
>         if (r->fsize!=st.st_size || r->fdate!=st.st_mtime)
>         { // changed again
>             ret=-1;
>             remove(r->filename_new);
>         }
>         else  <<<<<<<<<<<<LINE 397
>         {
> 
>         .... 
> 
> it strikes me as not being the right line to remove!
> 
> Sorry, my c coding is pretty basic, so need more advice.
> 
> By the way, gzip -9 on my received folder shrinks it to 3.6MB, so it looks
> as if the whole folder is being gzipped, but there is something going
> wrong in the process as you suggest.
> 
> Regards,
> 
> David Pilgram.
> 
> 
> 
> 


Reply via email to