Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

[Opinion] How annoyed am I at Microsoft? Let me count the ways...

  • Subject: [Opinion] How annoyed am I at Microsoft? Let me count the ways...
  • From: The Ghost In The Machine <ewill@xxxxxxxxxxxxxxxxxxxxxxx>
  • Date: Thu, 3 Jan 2008 11:28:18 -0800
  • Bytes: 4749
  • Newsgroups: comp.os.linux.advocacy
  • User-agent: slrn/0.9.8.1 (Linux)
  • Xref: ellandroad.demon.co.uk comp.os.linux.advocacy:589879
A rant, indeed a rant.

[1] Microsoft Windows Visual Studio 2008.  I'll give
it points for working, and they have improved the GUI
in spots (the list of libraries can now be edited as an
actual *list*, for starters).  Unfortunately it wants to
convert old format project files and has no backwards
compatibility mode for project files.  MAKE files do
not have this problem.  Not sure regarding GTK/Glade or
Kdevelop but suspect they're better off in this area, too.

[2] VS2008 has the bad habit of specifying paths with
respect to its destination, as opposed to VS2005 which
specified them with respect to the source.  This meant that
I had to manually retype in paths often.  I'll admit to
some curiosity as to how to properly handle the multihost
compilation and cross-compilation problems, but at least
with a Makefile one has a fighting chance.

[3] The compiler for 2005 allowed constructs such as

    for(int i = 0; i < 10; i++)
    { ... }
    printf("%d\n", i);

and

    std::vector<SomeLocalClass *>::iterator it = NULL;

to compile.

The former has since been removed (the variable 'i' is
local to the for loop now); the latter used to work as
the iterator was a pointer; however, now it's a class.

Blech.  Not sure if g++ was much better in this regard,
admittedly.  I think I can get by without fixing this
by reinstalling 2005.  Yeah, there's a great option. :-P

[4] The defragmenter is very weird today; the report says
"no files were unable to defragment" but the display
clearly shows a red band, indicating that there's one
file that has too many fragments.  Also, the defragmenter
defragmented the files but still left the free space rather
fragmented.  I'm going to try to rerun the defragmenter
after creating a 2 GiB file (I'm running it without a paging
file at the moment).  Of course this will tip it over into
the below-15% area, which it doesn't like.

At this point, the best I can do is create 256Mi chunks and
let it play with those; at least that way I can guarantee
the chunks are clear.

A plus: it will create the paging file without a reboot.

A minus: it's in a weird spot: Control Panel, Change a
Setting, Performance and Maintenance (OK so far, actually),
System (as if I'm setting things for someone else's
system??), Advanced, Performance (what, again?), Advanced
(again again??), Virtual Memory at the bottom.  One can
specify Custom Size (I would recommend a fixed size),
system managed size (oh, hello pagefile fragmentation),
or no paging file (which the system apparently does honor
now -- but one needs to reboot).

Not sure if Linux has -- or needs -- an analogue to this
weirdness.  Those "in the know" can do

    dd if=/dev/zero of=/var/tmp/scratch_space bs=1048576 count=size
    mkswap /var/tmp/scratch_space
    swapon /var/tmp/scratch_space

where size is in megabytes.  To remove this,

    swapoff /var/tmp/scratch_space
    rm /var/tmp/scratch_space

Note that scratch_space must be on a local drive.

Those not in the know probably should just let the
installer create the default -- about 2x one's RAM --
as opposed to allowing for auto-extension, fragmentation,
system malfunctions, and other such stupidity which is a
little too common on Windows.

[5] VS2005 and VS2008 like to copy files into a temp area,
then extract them to the final location.  This tends to use
more space than it really needs to.  To be fair, the other
method of direct extraction into the destination probably
gives Windows some interesting headaches.  (Linux doesn't
care too much as long as the package is guaranteed new.)

Yep, Windows is indeed state of the art.  Can we shoot it
now or wait until the end of the election?  :-P

-- 
#191, ewill3@xxxxxxxxxxxxx -- insert random irritations here
Useless C/C++ Programming Idea #992381111:
while(bit&BITMASK) ;

-- 
Posted via a free Usenet account from http://www.teranews.com


[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index