Linux-Development-Apps Digest #412, Volume #6 Wed, 8 Mar 00 03:13:13 EST
Contents:
Learning Curses (Binyomin Kaplan)
Re: I can't stand this X anymore! (Donovan Rebbechi)
Re: I can't stand this X anymore! (Christopher Wong)
Re: removing a module as a user (Olaf Klischat)
Re: Convert GB to Unicode (Takeyasu Wakabayashi)
Re: I can't stand this X anymore! (John Hasler)
Re: Convert GB to Unicode (Takeyasu Wakabayashi)
Re: I can't stand this X anymore! (Donovan Rebbechi)
having trouble w/ GCC native install (Neil J.)
sched_setpriority lockups ("Joe N.")
sched_setpriority lockups ("Joe N.")
Re: Struct size and allocate problem! need help. (Charles Bryant)
KDE problem ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: Binyomin Kaplan <[EMAIL PROTECTED]>
Subject: Learning Curses
Date: Tue, 07 Mar 2000 21:07:05 -0600
Deos anyone know of any resources for learning about
curses or ncurses besides the man pages and the sample
programs in /usr/doc?
Thanks in advance,
Binyomin
------------------------------
From: [EMAIL PROTECTED] (Donovan Rebbechi)
Crossposted-To: comp.os.linux.x,comp.os.linux.advocacy
Subject: Re: I can't stand this X anymore!
Date: 8 Mar 2000 03:51:54 GMT
On 8 Mar 2000 00:58:42 GMT, David T. Blake wrote:
>Donovan Rebbechi <[EMAIL PROTECTED]> wrote:
>> On 7 Mar 2000 22:35:34 GMT, David T. Blake wrote:
>
>> >Even though we're unable to tell for now whether FreeType
>> >violates them, this could mean that the free use of the library
>> >could be illegal in the US or other countries like Japan, be it
>> >in commercial or free projects.
>>
>> Wrong. It would mean that freetype would have to license the patent.
>
>Well, it is unclear whether this could ever happen. Apple
>would be unlikely to
Speculation. You said that Apple were planning to kill TrueType
on Linux, and haven't offered any supporting evidence.
>What is more likely is that Apple would propose a licensing
>plan that is untenable for the Freetype developers,
Again, pure speculation. And what would apple do by proposing that
anyway ?
>forget, metafont rendering PRECEDES TrueType by some time, and
>is actually a bit nicer.
Metafont is actually a vastly superior technology to TrueType *or*
Type1.
>The whole point of their patent was the same reason for
>developing any patent - protection of innovation for a fixed
>length of time through a legal granted monopoly. They enjoy
>this protection with TrueType fonts and make a lot of money
>from it.
Yes, IOW when someone makes money off it, they want some of it.
>any form of contract. LZW has certainly not been licensed on a
>royalty basis - but on a per use basis.
Actually, I believe LZW is free if your software is. This is why programs
like GIMP can legally use gif format. The patent onlygets expensive
if you want to make money.
The main problem is that you're supposed to pay if you use the GIFs on
a commrecial website.
>I find it highly likely.
Again, purely speculative/subjective.
>their products. IT people like Apple do this especially through
>patent protection. Their mentality
Again, speculative.
>Freetype people are NOT rich developers or even well funded. The
>site will shut down while they use a work around technique
>like metafont.
The site is still up.
>> Patent holders do not benefit from bad publicity. And they
>> don't stand to gain financially from setting their lawyers onto
>> the little guys. And there's no evidence they plan to do that.
>
>This is not something tremendously widespread in Unixland. Most
>people don't even understand that Freetype exists, or even
The entire free software community knows, and the trade press are
always open to a good story from the free software community. Apple
have attempted to woo the OSS community by getting their blessing on
their new OS. The OSS community are starting to carry a fair degree of
weight, and I think you are mistaken to dismiss the ability of OSS
spokespeople to rattle some chains where appropriate.
--
Donovan
------------------------------
From: [EMAIL PROTECTED] (Christopher Wong)
Crossposted-To: comp.os.linux.x,comp.os.linux.advocacy
Subject: Re: I can't stand this X anymore!
Reply-To: [EMAIL PROTECTED]
Date: Wed, 08 Mar 2000 04:53:27 GMT
On 8 Mar 2000 02:22:57 GMT, Donovan Rebbechi <[EMAIL PROTECTED]> wrote:
>Fair enough. My main point is that outside these fonts, you won't notice
>much difference. Even with the sans fonts ( Arial, Verdana ), the difference
>is quite small.
On my part, what I am trying to say is this: TrueType is the only way to
get screen fonts comparable to Mac/Windows/OS/2. The only way, that is,
short of writing your own Type 1 rasterizer. It does not really bother
me that the selection of good TT fonts is limited. Frankly, I really
only need one good serif font for web browsing. The saying goes that you
judge a mountain range by its highest peaks. The same goes here for me:
I only are about the BEST fonts I can get, because those are the ones I
would use. I am not contesting your assertion that Type 1 fonts are
comparable or better on average. The thing is, I won't settle for
average: I want only the best fonts, and that means (cough) Microsoft's.
>>In practice: if I "load up a gazillion fonts", there are two
>>scenarios. If I go with TT, I get excellent fonts most of the time and
>>ugly fonts some of the time. With Type 1, I get ugly fonts all of the
>>time. I think it does matter which one I would go with.
>
>This is misrepresenting the situation. If you "go with" TT, you get
>*very few* excellent fonts. In particular, outside the special webfonts,
>you will notice next to no difference.
And if I set up my system correctly, I need never be "outside the
special webfonts". There are 10 of them from MS, after all. That is why
I said I would get excellent fonts most of the time. It's not like I
write my letters with a gazillion fonts at once.
>>I would need more than a decent Type 1 font. I need a decent Type 1
>>rasterizer: the one that comes with XFree86 is crap.
>
>Like I said, an apples to apples comparison, where you convert TrueType
>to Type1 shows that there is very little difference. Perhaps I can post
>some screenshots.
Please do. I am genuinely interested. Are you saying that if I convert
one of MS' hinted TTF fonts (say, Georgia) to Type 1, I would hardly see
the difference? This is hard for me to imagine without seeing, unless
you regularly set your browser font to 32points. TrueType hints do not
survive conversion, from what I understand, and X's Type 1 rasterizer
does not respect hints anyway. I have tried your font RPM (thanks, by
the way, for offering it) and saw the usual disparity between
well-hinted TT fonts and generic Type 1 fonts. There really is no
comparison: for Linux/X screen fonts, TrueType's best leaves Type 1's
best in the dust.
I also stand by my assertion that X's Type 1 rasterizer is crap. That
URL I posted was not just a comparison between Type 1 and TTF. It also
compared Type 1 rasterizers, and the one in X11R6 was pretty bad
compared to Adobe's. Hopefully, as I think you said, the Freetype folks
will come up with a Type 1 rasterizer comparable to the 10 year old
ATM.
Chris
------------------------------
From: Olaf Klischat <[EMAIL PROTECTED]>
Subject: Re: removing a module as a user
Date: 08 Mar 2000 06:06:59 +0100
[EMAIL PROTECTED] (kai-martin) writes:
> Moin,
>
> I'd like to have a program in userspace remove a module from
> the kernel, but I just can't figure out how.
>
> This is, what I tried:
>
> A little program that calls a shellscript. The script in turn
> executes the command rmmod.
This doesn't work because /bin/sh executes scripts only with its own
real UID and not with its effective UID (in programs that have the
s-bit set, the EUID will be set to the the UID of the owner of the
executable, while the real UID will remain that of the calling
user). Consequently, your script will only run with the UID of the
user who called the wrapper program (the UIDs are handed over from the
wrapper to /bin/sh).
This behaviour is built into /bin/sh for security reasons.
> Owner and group of the program is root and the s-bit is set.
> ---------------8<----------
> #include <stdlib.h>
> #include <stdio.h>
>
> int main()
> {
> execlp("/bin/sh", "-e", "/local/pci-mio-e/restart_comedi");
Hmm.. if I got the info pages right, this has to be:
execlp("/bin/sh", "/bin/sh", "-e", "/local/pci-mio-e/restart_comedi",NULL);
The 2nd argument to execlp is argv[0], which, by convention, has to be
the path to the executable. The `real' cmdline parameters start from
the third argument to execlp.
> }
> -------------->8---------
> The program seems to be allowed to execute the root-only shell script,
> but removing the module fails ---> "rmmod: comedi: permission denied"
> Calling the script directly as root works fine.
>
>
> I also tried to run rmmod directly via execlp .
> This produces the error:
> " The insmod/rmmod combined binary is having an identity crisis.
:))
> Please give it a proper name."
You might have made the same mistake as above..
try:
execlp("/sbin/rmmod","/sbin/rmmod","modname",NULL);
--
Olaf Klischat Berlin Univ. of Technology
mail: [EMAIL PROTECTED]
------------------------------
From: Takeyasu Wakabayashi <[EMAIL PROTECTED]>
Subject: Re: Convert GB to Unicode
Date: 08 Mar 2000 14:23:44 +0900
Robert Brady <[EMAIL PROTECTED]> writes:
>
> Takeyasu Wakabayashi <[EMAIL PROTECTED]> wrote:
> >Current glibc doesn't support multibyte chars at all.
> >To use UNIX standard `iconv' facility to convert multibyte
> >chars, you need CVS version of glibc 2.2.0.
>
> This is false. glibc 2.1 iconv converts multibyte characters
> just fine.
>
Even without LC_TYPE locale files for multibyte character locales?
The localedef program for glibc 2.1 can't create locale files for
zh_CN.GB, ja_JP.eucJP or such. I don't know for UNICODE locales.
To be honest, I've never used iconv and don't know whether it uses
LC_TYPE file, but I doubt it can...
--
Takeyasu Wakabayashi
Faculty of Economics, Toyama University
[EMAIL PROTECTED]
------------------------------
From: John Hasler <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.x,comp.os.linux.advocacy
Subject: Re: I can't stand this X anymore!
Date: Wed, 8 Mar 2000 03:04:41 GMT
David T. Blake writes:
> Apple would be unlikely to grant a free ride for True Type fonts on a
> free platform while it accumulates royalties from other platforms - it
> doesn't make sense.
How would it harm them to permit free use of True Type on free platforms
when there is little chance that they could collect any royalties for such
use anyway? While I doubt that they will do it, it certainly could make
sense.
> They would have to freely license the technology to all.
They would have to do no such thing.
--
John Hasler
[EMAIL PROTECTED]
Dancing Horse Hill
Elmwood, Wisconsin
------------------------------
From: Takeyasu Wakabayashi <[EMAIL PROTECTED]>
Subject: Re: Convert GB to Unicode
Date: 08 Mar 2000 14:26:32 +0900
Takeyasu Wakabayashi <[EMAIL PROTECTED]> writes:
> Even without LC_TYPE locale files for multibyte character locales?
> The localedef program for glibc 2.1 can't create locale files for
> zh_CN.GB, ja_JP.eucJP or such. I don't know for UNICODE locales.
>
> To be honest, I've never used iconv and don't know whether it uses
> LC_TYPE file, but I doubt it can...
Ooops, typo! Read LC_CTYPE instead of LC_TYPE.
--
Takeyasu Wakabayashi
Faculty of Economics, Toyama University
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Donovan Rebbechi)
Crossposted-To: comp.os.linux.x,comp.os.linux.advocacy
Subject: Re: I can't stand this X anymore!
Date: 8 Mar 2000 06:37:04 GMT
On Wed, 08 Mar 2000 04:53:27 GMT, Christopher Wong wrote:
>On 8 Mar 2000 02:22:57 GMT, Donovan Rebbechi <[EMAIL PROTECTED]> wrote:
>And if I set up my system correctly, I need never be "outside the
>special webfonts".
Well that's fine if you never do anything besides browse the web.
Personally, I don't want to use those same fonts in every document
I create. Especially since I usually care more how my documents look
on paper than on screen.
>I said I would get excellent fonts most of the time. It's not like I
>write my letters with a gazillion fonts at once.
It's not like the web fonts are always the best for writing letters,
either.
>>Like I said, an apples to apples comparison, where you convert TrueType
>>to Type1 shows that there is very little difference. Perhaps I can post
>>some screenshots.
>
>Please do. I am genuinely interested. Are you saying that if I convert
>one of MS' hinted TTF fonts (say, Georgia) to Type 1, I would hardly see
>the difference?
Two points:
(1)
You'll only get visible degradation at small point sizes. The URL you post --
uses point sizes between 10 and 20, *at 72 DPI*. If you have a 100dpi
display, this is analagous to adjusting the range to 7-14 points.
Note further that the degradation doesn't kick in until the second or
third font down -- so you won't really see degradation until you're
looking at 10 point fonts or so. Personally, I magnify everything to
at least 14 point on screen ( even if I write a word processor document
in 10 point font, I magnify the screen view 150% )
(2)
Like I said before, *most TrueType fonts are not hinted* Not a little bit.
in fact, not at all. So with the exceptions of special "web fonts", you
will not see any difference.
> This is hard for me to imagine without seeing, unless
>you regularly set your browser font to 32points.
Completely misleading.
>I also stand by my assertion that X's Type 1 rasterizer is crap. That
I'd agree that this is a big part of the problem.
>URL I posted was not just a comparison between Type 1 and TTF. It also
>compared Type 1 rasterizers, and the one in X11R6 was pretty bad
I'd agree ... but also I'd add that unless you're using fairly small fonts,
it doesn't matter that much.
--
Donovan
------------------------------
From: Neil J. <[EMAIL PROTECTED]>
Subject: having trouble w/ GCC native install
Date: Wed, 08 Mar 2000 06:30:21 GMT
Having spent several hours toying with it last night, I really hope someone
can help me (a newbie) out. I downloaded, unzipped, and untarred GCC
2.95.2. GCC recommended that I configure it from a different directory
than where it was installed (the new one is /root/gcc). From there, when I
run configure from the command line, it correctly identifies my machine as
an i686-pc-linux-gnuoldld, but then outputs that command "cc" is
unrecognized (evidently it tests "cc" functionality in the config scripts).
The GCC help site said that for installation as a native compiler, I would
have to set cc in my environment variables. I thought this was done with
"env [filename]" at the command line. Am I wrong? I have tried
practically every file (except *.c and *.h) as the target for this. Is it
a different file, or is the technique altogether different?
Thank you for your help -- switching from Windows to UNIX/Linux has been
time-consuming, but I know it will pay off eventually.
Neil (Sorry about the e-mail address - I'd rather look at Britney than ads,
though.)
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: "Joe N." <[EMAIL PROTECTED]>
Subject: sched_setpriority lockups
Date: Tue, 7 Mar 2000 14:50:58 -0700
I wrote a simple TCP/IP client/server program to do some timing test on my
network. It uses blocking sockets and works great with the default Linux
scheduling policy. When I try to set it to a real-time policy (SCHED_RR) at
any priority (I tried max and min) it locks up the system after the
connection is established. Isn't the scheduler supposed to give other
processes CPU time when my process blocks on a send or recv? I've also
tried it with non-blocking sockets and a usleep call. This also locks up
the system. Does anyone have any comments or suggestions? Is anyone having
similar problems? Thanks.
I'm using Red Hat 6.1 (Kernel 2.2.12-20).
Joe
[EMAIL PROTECTED]
------------------------------
From: "Joe N." <[EMAIL PROTECTED]>
Subject: sched_setpriority lockups
Date: Tue, 7 Mar 2000 14:51:18 -0700
I wrote a simple TCP/IP client/server program to do some timing test on my
network. It uses blocking sockets and works great with the default Linux
scheduling policy. When I try to set it to a real-time policy (SCHED_RR) at
any priority (I tried max and min) it locks up the system after the
connection is established. Isn't the scheduler supposed to give other
processes CPU time when my process blocks on a send or recv? I've also
tried it with non-blocking sockets and a usleep call. This also locks up
the system. Does anyone have any comments or suggestions? Is anyone having
similar problems? Thanks.
I'm using Red Hat 6.1 (Kernel 2.2.12-20).
Joe
------------------------------
From: Charles Bryant <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux,comp.os.linux.development.system,comp.unix.sco.misc,comp.unix.sco.programmer,comp.unix.unixware.misc,tw.bbs.comp.linux
Subject: Re: Struct size and allocate problem! need help.
Date: 8 Mar 2000 02:09:10 -0000
In article <[EMAIL PROTECTED]>,
P.G.Hamer <[EMAIL PROTECTED]> wrote:
>Charles Bryant wrote:
>
>> In article <[EMAIL PROTECTED]>,
>> P.G.Hamer <[EMAIL PROTECTED]> wrote:
>> >My dated 2nd edition of K&R states that sizeof() gives the size in /bytes/.
>>
>> Where?
>
>p204 under A7.4.8 Sizeof Operator 1st sentence.
Ok. I have checked that this means what it ought to, and it's a usage
of the term 'byte' that seems to be less common nowadays. In this
sense a byte is the smallest addressable unit of storage, which can
be any number of bits. Since so many processors address units of
eight bits, many people now use the term 'byte' to mean eight bits.
When referring to data communications this meaning is clear enough,
since addressability is maningless for data in transit, but when
referring to CPUs accessing memory it can be confusing.
That's what this reference is saying:
>Looks like this is true, although I'm uncertain of its date
>http://www.lysator.liu.se/c/rat/c3.html#3-3-3-4
>says:
>It is fundamental to the correct usage of functions such as malloc and
>fread that sizeof (char) be exactly one. In practice, this means that a byte
>in C terms is the smallest unit of storage, even if this unit is 36 bits wide;
>and all objects are comprised of an integral number of these smallest units.
but it doesn't explain that 'byte' originally meant 'storage unit'.
--
Eppur si muove
------------------------------
From: [EMAIL PROTECTED]
Subject: KDE problem
Date: Wed, 08 Mar 2000 07:42:19 GMT
Hi 2 everyone,
After hard days of precompiling my linux kernel :), I build a brand new
one (without audio support). But now when I try to run my KDE X-server
tells me that cannot start kaudio. How can I tell him not to do such
thing. Thank U 4 your help.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.apps) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-Apps Digest
******************************
