Re: The LyX licence
Dear LyX developers,
from the text of Angus's message I'd say you were sending these messages
before. I moved last July and my old address @cs.wm.edu is not
that useful any more. Fortunately, Angus has found my other address.
I do grant permission to LyX to use any of my (small) contributions
under BSD or GPL license. Whichever you decide suits you better.
Best regards,
Zvezdan Petkovic
Re: Online bibliography support
On Mon, Apr 05, 2004 at 12:52:45PM +0200, Andre Poenitz wrote: > ... or pickes the right distribution. [And no, I am not too big a fan of > SuSE as they don't include LaTeX + Co in the personal edition anymore...] Some people can't pick. We were talking about academic non-technical users. They get introduced to Linux most of the time in their university department, where somebody else picks up for them. In USA, many universities picked Red Hat long time ago. Some of these users may decide to buy their own Linux machine after a while. So they go online to Dell, IBM, whatever, and they get a machine with preinstalled -- surprise, surprise -- Red Hat. And most of such users don't bother learning about distributions. For them Linux is Linux, plus they prefer to have the same environment as at work. Another trend I noticed among both technically unsavvy and savvy academic users in USA is that they are buying Apple computers with Mac OS X more and more. They will not hurt Microsoft any time soon, but I would say that is a progressive trend. We may want to support that Mac OS X port with greater care. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Online bibliography support
On Mon, Apr 05, 2004 at 12:15:08AM +0200, Janus Sandsgaard wrote: > Huh? I am pretty sure that LyX came with my SuSE ditribution together with a > lot of LaTeX stuff. As far as I remember the same goes for Mandrake. Red Hat doesn't distribute LyX, which is sad considering how many people, at least in USA use it. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Online bibliography support
On Sun, Apr 04, 2004 at 01:09:03PM +0100, Nachev, Parashkev C wrote: > ... I spoke of academia in general because I meant to include those > for whom a LaTex-based system is not *essential*, and those who have > no interest in computing. These people will use Word, and persist in > struggling with its inadequacies, because they find the alternatives > too complex to set up and use without technical knowledge. I agree. There's one more problem. Linux distributions do not include LyX or pybliographer yet. Red Hat comes to mind. That means that people who are not technically skilled are expected to install these applications for themselves, even on Linux, unless they have a LyX-friendly admin. I can explain that previously they refused to do that on account of XForms-0.8x licence, but now when Qt backend is available there's no excuse for not including it. It seems that everybody is forcing OpenOffice on us these days. That thing is so huge and clunky that I wouldn't touch it with a stick. KOffice is getting better, but still not there. Since I do not use them anyway, I'd like to see LyX included in Red Hat (Fedora) with some bibliography tool. > Now you may of course feel that these people deserve their misery, but > surely the whole point of Lyx is to make the power of LaTex easily > accessible to those who have neither the facility nor the wish to > understand its technicalities. Agreed. A little help from Linux distros wouldn't hurt either. :-) As an aside note, for people who use BibTeX only databases, tkbibtex is still unbeatable in terms of ease of installation. Just drop it in your $HOME/bin and add exec permission to it. And it works on Windows too, since Tcl/Tk is available for all platforms. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Online bibliography support
On Sun, Apr 04, 2004 at 02:09:04AM +0100, John Levon wrote: > I didn't mean to offend. Sorry. It's all right. I know you didn't. > > All these organisations prefer LaTeX. All of them are the most > > important places to publish if you are a Computer Scientist, Engineer or > > Mathematician. > > That's fine. But you should, and cannot, discount the reality of the > huge number of potential LyX users who are forced into the realities of > Word. Oh I definitely don't. I limited myself to scientific community in computer science/engineering and math. In fact, a few days ago I was asked to resubmit a CV for my job application in Word. PDF wasn't good enough. :-) BTW, anyone has a job offer? -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Online bibliography support
On Sun, Apr 04, 2004 at 01:15:52AM +0100, John Levon wrote: > On Sat, Apr 03, 2004 at 03:32:49PM -0500, Zvezdan Petkovic wrote: > > > You mentioned scientific community before mentioning Word. > > I do not know which scientific community you belong to, > > but I know for sure that all computer scientists, mathematicians, and > > physicists I know use LaTeX. It's _THE_ format in their journals and > > This is (almost hopelessly) naive. A LOT of places REQUIRE Word format. > This includes physicists, and for sociology etc. I understand the > situation is even worse. John, you really have a penchant to step on my toes for no particular reason. :-) Hopelessly naive implies that I'm an idiot who doesn't know what's he talking about. Yet I'm not. I have clearly indicated that I'm talking about the scientific community I belong to, _NOT_ sociology, etc. I know too well how ubiquitous Word is, and I do not need a reminder. Again, I was talking about my profession, and again about scientific community only. Because in business, Word and Microsoft rule. I grant you that some Physics journals require Word. I have a few friends who are physicists; they use LaTeX, and that's why I mentioned physicists above, too. You've probably noticed that I said above "and physicists _I KNOW_". Let's take them out of the argument since I do not know enough about the physics journal requirements. However, all important computer science and math journals/conferences welcome, and in fact prefer LaTeX. IEEE has LaTeX style for their Transactions on XXX and conferences -- that's the biggest world organisation for Electrical, Electronics and Computer Engineers. ACM has LaTeX style for their Journals and Transactions -- extremely important organisation for all Computer Scientists/Engineers. As mentioned in previous post their huge Digital Library has all the documents back to the middle '80s in PDF format and the citations for the documents in BibTeX format _ONLY_. USENIX has both LaTeX and troff styles for their conferences. Very important conferences to publish at, if you are a serious researcher. AMS -- the world biggest Math society is the major promoter of (La)TeX. SIAM -- second largest math society has LaTeX style MAA -- another large math society enables some features _ONLY_ to authors who used LaTeX All these organisations prefer LaTeX. All of them are the most important places to publish if you are a Computer Scientist, Engineer or Mathematician. Most of the publishers in Computer Science and Mathematics accept LaTeX typeset books. Too many to list them here. I rest my case. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Online bibliography support
On Sat, Apr 03, 2004 at 11:32:25AM +0100, Nachev, Parashkev C wrote: > It seems to me the main thing that is stopping the whole of the > scientific community switching to Lyx is the lack of an easy way of > integrating it with bibliographic management software. None of the open > source bibliographic software has the functionality of Endnote, in > particular the ability to search, retreive, and insert citations to > online references. I think you never tried pybliographer then. Sixpack can also work with remote databases. Sixpack can convert from other formats to BibTeX, and from BibTeX to other formats. It uses excellent bp Perl package by Dana Johnson for conversions. Both Pybliographer and Sixpack can pass the selected citation into LyX document being written. Just yesterday I converted a rather big BibTeX file to refer, and that is exactly what Endnote uses. You give Endnote as a paragon of functionality, yet it uses the same format as good old troff (refer + a few extra fields they added), which is available on every UNIX machine since beginning. :-) > If there is any way of incorporating a plug-in that will do this I > think Lyx may even kill off Word, in academia at least. You mentioned scientific community before mentioning Word. I do not know which scientific community you belong to, but I know for sure that all computer scientists, mathematicians, and physicists I know use LaTeX. It's _THE_ format in their journals and societies. However, it's hard to persuade a long time LaTeX user to switch to LyX or any other GUI. That's more of a problem than Word. Also, regarding the online references, that again depends on the community one belongs to. The huge Digital Library of the Association for Computing Machinery, offers BibTeX citation for each publication in the library. No other formats. Why? Because computer scientists use LaTeX or even troff, not Word. Since BibTeX is richer format than refer, that makes sense. troff users can convert BibTeX to refer as already mentioned. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: how to get smooth fonts in pdf
On Thu, Mar 18, 2004 at 04:54:44PM +0100, Juergen Spitzmueller wrote:
> No. The difference is that palatino.sty loads helvetica and does not scale it
> correctly. And mathpazo supports not only the palladio fonts (palatino
> clones) which are included in the base35, but also true palatino fonts with
> true smallcaps and old style nums (of course you have to buy those. I did,
> and I didn't regret the EUR25 ever since. The smallcaps and the oldstyle nums
> are just beautiful). Also, IIRC, the metrics in mathpazo.sty are better than
> in palatino.sty.
> http://home.vr-web.de/was/pplx.html
So, that's the explanation for this piece of code in mathpazo:
[EMAIL PROTECTED]
\renewcommand{\rmdefault}{pplj}
\renewcommand{\oldstylenums}[1]{%
{\fontfamily{pplj}\selectfont #1}}
[EMAIL PROTECTED]
\renewcommand{\rmdefault}{pplx}
\renewcommand{\oldstylenums}[1]{%
{\fontfamily{pplj}\selectfont #1}}
\else
\renewcommand{\rmdefault}{ppl}
\fi\fi
But it seems that if you do not have the commercial fonts, you still end
up with the good ol' {ppl}. :-)
> > Notice that you could scale Helvetica by 0.9 to 0.92 after loading
> > palatino and euler similar to the suggestion in the article.
>
> Sure. But no one does it. IMO LyX should scale correctly.
Agreed.
> > A person who doesn't care about math can _much_ easier recognise the
> > name palatino than mathpazo. If you write only non-math texts, why
> > would you care whether it's palatino or mathpazo? My wife writes her
> > letters in LyX. She really doesn't care either about math or Helvetica
> > scaling. Palatino works great for her. It's far from obsolete.
>
> Does not matter if you use LyX. The name in the combo will still be palatino.
> Just LyX loads the mathpazo package. Notice that Walter Schmidt, the
> maintainer of the psfnss bundle (the type-1 support for LaTeX) declared
> palatino.sty obsolete. It's likely to be removed in favor of mathpazo.
You are right. I've just tried this in a test TeX file:
\usepackage{mathpazo,courier,euler}
\usepackage[scaled=.95]{helvet}
and it worked just fine. So even if one doesn't want mathpazo for math,
it will still work fine, as long as one loads the replacement math font
after mathpazo.
But regarding the removal of palatino, wouldn't it be better to simply
change it from:
\renewcommand{\rmdefault}{ppl}
\renewcommand{\sfdefault}{phv}
\renewcommand{\ttdefault}{pcr}
to
\RequirePackage{mathpazo}
\RequirePackage[scaled=.95]{helvet}
\RequirePackage{courier}
rather than simply remove it?
> > > And instead of ae, I'd recommend latin modern, a much better type-1
> > > version of cm (though still beta and not yet perfect).
> >
> > Which package do you load for this?
>
> \usepackage{lmodern}. See
> http://www.tex.ac.uk/cgi-bin/texfaq2html?label=uselmfonts
Well, it's still not a part of teTeX distribution on Red Hat Linux 9,
and OpenBSD 3.4 that are available to me.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: how to get smooth fonts in pdf
On Thu, Mar 18, 2004 at 09:48:39AM +0100, Juergen Spitzmueller wrote:
> Zvezdan Petkovic wrote:
> > If you choose Layout->Document and in the dialogue that opens set Fonts
> > button to "palatino" (or some other non-default), there are no problems
> > with PDF whatsoever. Well, almost. Math is still in Computer Modern
> > and comes out fuzzy in PDF.
>
> Actually, \usepackage{palatino} is completely outdated as almost the
> whole LyX font stuff:
> ftp://ftp.gwdg.de/pub/dante/info/l2tabu/english/l2tabuen.pdf
>
> You'd better
> \usepackage{mathpazo}
> then you'll also get matching math fonts.
First of all, thanks for pointing me to this text, but I must tell you
that after reading it I realised that I do _not_ commit _any_ of those
"deadly sins". Did you notice that I used Euler for math with palatino?
Namely, if you load palatino + euler you get:
1) Palatino for Serif (roman)
2) Helvetica for Sans
3) Courier for Monospace
4) Euler for math
If you load mathpazo + helvetica + courier as the article above suggests
you get:
1) Palatino for Serif (roman)
2) Helvetica for Sans
3) Courier for Monospace
4) Italic Palatino for math
As you can see the difference is only in the math font. And that
difference is deliberate. I _want_ Euler for my math.
Hack, Euler matches _much_ better with Palatino then with the Concrete
fonts used for text in Knuths "Concrete Math". Nobody accused Knuth of
committing sin for using different fonts for text and math in that book.
:-)
Notice that you could scale Helvetica by 0.9 to 0.92 after loading
palatino and euler similar to the suggestion in the article.
A person who doesn't care about math can _much_ easier recognise the
name palatino than mathpazo. If you write only non-math texts, why
would you care whether it's palatino or mathpazo? My wife writes her
letters in LyX. She really doesn't care either about math or Helvetica
scaling. Palatino works great for her. It's far from obsolete.
On the other hand, if you really care about math, you are right. One
needs to use either mathpazo or other matching math font. Now, I knew
about old mathppl package, I didn't like it, and I always used Euler
with Palatino. That is the reason I never bothered to see if there is
something new, like mathpazo. Mathpazo simply loads italicised Palatino
for math. Frankly, I still prefer Euler for my math.
If you read French, see this nice article about math fonts in TeX,
and how to make math match your Multiple Master fonts:
http://www.gutenberg.eu.org/pub/GUTenberg/publicationsPDF/25-bouche.pdf
The footnote number 4 on the page 3 prompted me to start using
Palatino+Euler combination, and I'm happy with it ever since.
Of course, this is my personal decision, and there's no accounting for
taste. You are absolutely right that people who simply want to use
Palatino, and have any math in their document should use mathpazo.
>
> And instead of ae, I'd recommend latin modern, a much better type-1
> version of cm (though still beta and not yet perfect).
Which package do you load for this?
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: how to get smooth fonts in pdf
On Wed, Mar 17, 2004 at 04:18:12PM +, Andreas wrote:
> Some of you might be familiar with the poor quality of the fonts when
> viewing a pdf file created from a lyx document. The fonts look
> pixelised and it is not possible to mark text.
Only if you use the default Computer Modern fonts.
Now, with all due respect to Knuth as a great scientist,
he's not a great designer, and I prefer to use fonts designed by
Hermann Zapf -- art is an art (Palatino is used in IEEE Transactions,
anyway).
If you choose Layout->Document and in the dialogue that opens set Fonts
button to "palatino" (or some other non-default), there are no problems
with PDF whatsoever. Well, almost. Math is still in Computer Modern
and comes out fuzzy in PDF.
Solution: add to the preamble
\usepackage{euler}
Euler fonts were designed by Zapf (with Knuths programming help) for AMS
(American Mathematical Society) to look, as Knuth puts it, like a math
written on a blackboard by a mathematician with perfect handwriting.
Knuth used Euler in his book "Concrete Math".
Euler and Palatino match amazingly well (Zapf did design both, but
still, the match is too good for the fonts created so far apart in
time) for math and text.
Now, if you still prefer Computer Modern (CM from now on) for math, and
find that you mostly produce Postscript and PDF, you can enforce the use
of Type1 fonts for CM. Find (depending on distribution)
/usr/share/texmf/dvips/config/updmap
or
/usr/local/share/texmf/dvips/config/updmap
Change the lines:
type1_default=false
# type1_default=true
to
# type1_default=false
type1_default=true
and run updmap to update maps (all this as superuser).
Then, the CM math will be in Type1 with your Postscript fonts (Palatino,
Times, whatever).
> The reason for this is the
> \usepackage[T1]{fontenc}
> statement that lyx automatically adds to the document header.
If you use Postscript fonts, this is not an obstacle at all.
With CM fonts, this is true, but there's a simple solution.
> Commenting it out would do the trick but we don't want to lose the European
> special characters, so an additional package can be used to fix this:
>
> \usepackage{ae}
> should always be included when fontenc is included.
> According to
> http://www.uni-koeln.de/rrzk/kurse/unterlagen/latex/ergaenzungen/pdftex.pdf
> (if you read German), this controls the use of AE-fonts (whatever they
> are) that are needed to support the use of Type1 Fonts in pdf.
You just need to select Layout->Document and in the dialogue that opens
set Fonts button to "ae". That's simple enough.
> For the moment, I have helped myself by putting this into my preamble
> but please check if it would make sense to always include it.
No, it wouldn't. I do not use CM, and that package is not needed for
me, and many other people who prefer Postscript fonts (Palatino, Times,
etc.) to CM.
> It would save others from the hours of debugging me and my sysadmin at
> uni have spent on this problem.
This "problem" is fairly well documented on Internet.
And it is documented in LyX: Help->Extended Features
Section "Exporting Other Formats".
> PS: Please CC, as I'm not a subscriber to this list.
Well if you expect reply you should subscribe at least for a short time.
This guy explains well why this is not a proper netiquette:
http://www.trumpetpower.com/Rants/Netiquette/
Section: Don't do drive-by postings
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: pthread problems
On Wed, Mar 17, 2004 at 04:05:45PM +0900, Rob Lahaye wrote: > I remember similar things happened here when I compiled with Qt. > I believe it was caused by the Qt libs which were compiled with > threads by default from the ports. > > However, threads are not needed for LyX and thus LyX/Qt developers > will say: "get/create Qt libs without threads". At least that how > I remember this discussion ended. > > Keep in mind that I myself am not at all a threads expert! Hmm. I've built a lyx-1.3.4 port for OpenBSD and it uses qt3-mt (multi-threaded by default since that is the version installed for KDE) $ pkg_info | grep qt qt3-base-1.2 include and docs for qt3 qt3-mt-1.2 C++ X11 GUI toolkit lyx-1.3.4-qt graphical frontend for LaTeX (nearly WYSIWYG) $ ldd /usr/local/bin/lyx | egrep "(qt|thread)" 05242000 25367000 rlib 1 /usr/local/lib/libqt-mt.so.3.12 07ae5000 27aef000 rlib 1 /usr/lib/libpthread.so.2.1 Perhaps you can find something useful in my recent posts about changes I had to do to compile it for OpenBSD. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdanl
Re: fixes for lyx-1.3.4 on OpenBSD
On Mon, Mar 01, 2004 at 05:24:49PM +0100, Jean-Marc Lasgouttes wrote: > >>>>> "Zvezdan" == Zvezdan Petkovic <[EMAIL PROTECTED]> writes: > > Zvezdan> Hello everybody, I've built a port of lyx-1.3.4 for OpenBSD. > Zvezdan> Here are the fixes I had to apply. > > Hi Zvezdan, > > Following the discussion of your patches, I have applied the qt.m4 patch > to HEAD and BRANCH_1_3_X, and the boost patch to BRANCH_1_3_X only. > Yes. Thank you Jean-Marc. That will mean that lyx could compile in the future releases without any patches or environment variable tricks on OpenBSD. That's always OK. :-) > Is that OK with you? > > Thanks. > > JMarc Thank you. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
Angus,
On Wed, Feb 25, 2004 at 09:29:42AM +, Angus Leeming wrote:
> I remember a chat I had with Paul Smith on the gnu-make list.
> Basically, he suggested that if I *had* to use a syntax common to all
> makes then the only way was
>
> .C:.ui
> $${HEADER}=... \ (can't remember the syntax ;-)
> $(UIC) $(UICFLAGS) $< -o $${HEADER} \
> $(UIC) $(UICFLAGS) -impl $${HEADER} $< -o $@
>
> Ie, generate the header file as a by-product of generating the .C
> file. Could I get you to investigate this some more? That would leave
> only the "%_moc.C: $(srcdir)/../%.h" rule to think about more deeply.
That suffix rule line should be:
.ui.C: # I remember it as .FROM.TO:
if it's going to be portable.
However, unless you want the portability because of other systems,
I think we can leave it as it is.
When I posted the port to OpenBSD ports mailing list I asked about their
opinion on switching to Qt as the default flavour, and dependence on
gmake introduced by that. I also mentioned that gmake dependence is
negligible compared to teTeX dependence, but still attached proposed BSD
make equivalents (notice that they were _not_ used in port, I simply
made it use gmake)
One of the OpenBSD ports committers replied today regarding the make
stuff. Quote:
"You're going overboard here. Just depend on gmake."
I think that settles it. :-)
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
On Tue, Feb 24, 2004 at 06:13:42PM +, John Levon wrote: > No you don't. qt-config would default to the "system" Qt, as I showed > you in the fake ls output. In the case above, qt3 with mt. If a program > depends on qt2 or non-mt, it specifically runs with --no-mt or qt-config > -v 2 or whatever. > > Simple. Right. But LyX config didn't do that. It bailed out because it couldn't find moc or moc2 despite the fact that my environment variables told it nicely that it's called moc3-mt. > > The ports infrastructure > > is a distribution-specific feature, and no use whatsoever to us > generally. Right again. Nor did it affect LyX config in any way. I didn't want to argue. I just want the LyX configure to do a better job. If you can make it find the right Qt with qt-config or whatever, everything is fine and dandy. However, if you can't (which is currently the case) then I think this "distribution feature" can be really helpful under condition that configure really respects environment variables. It does respect LIBS, I do not see any reason not to respect MOC or UIC. Just my 2 cents. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
On Tue, Feb 24, 2004 at 05:22:17PM +, John Levon wrote: > with --moc, --uic options etc. *That's* sane. I suspect the problem is when a user has KDE with qt3-mt dependency, and perhaps a single application which either doesn't support threading and depends on qt3 or even qt2. Then you have both installed. If your configure program picks up moc2 or qt-config finds qt2 first, than you build LyX with inferior version of Qt, just because some old Qt2 application is lurking around your desktop. The ports infrastructure ensures that you get what you want. You ask for -mt explicitly with MODQT_MT=Yes, you get environment filled properly, you pass it to the building process. The application gets built with the most suitable version of Qt. (Similar stands for Gtk applications and Gtk-1.x and Gtk-2.x) You as a port builder control your port through the environment variables used in the build process. The end user then simply does: cd /usr/ports/print/lyx sudo make install and goes to sleep while all the dependencies (teTeX, Qt, ispell, ...) and LyX build. Or (s)he simply installs the prebuilt packages. :-) > Anyway, we don't have that, so I suppose your patch is OK. It's clearly > preferable to hardcoding "moc3-mt" as another name to try anyway. I absolutely agree. Because they may decide tomorrow that it should be called only moc-mt or moc4-mt. :-) Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
On Tue, Feb 24, 2004 at 12:01:43PM +0100, Jean-Marc Lasgouttes wrote: > > What about asking on the automake list whether there is some provision > to do what we want to do? > > And what is the main problem in this rule? The fact that append _moc > to the name? Do we really need to do that? No. If you have such a rule: %.h: %.ui $(UIC) $(UICFLAGS) $< -o $@ it can be easily replaced (and that works in every make) with: .SUFFIXES: .h .ui .ui.h: $(UIC) $(UICFLAGS) $< -o $@ But the rule like this %_moc.C: $(srcdir)/../%.h $(MOC) -o $@ $< cannot. Not because of _moc.C. You can put _moc.C in the SUFFIXES rule. The problem is the path in front of .h that can't go in the SUFFIXES rule ($(srcdir)/../). Even worse is the third case we have there: %.C: %.h %.ui $(UIC) $(UICFLAGS) -impl $^ -o $@ This is basically, telling that you make X.C file using X.h file, but they also depend on X.ui. You can't do that with suffix rules. .h.C: X.ui is a wrong syntax. Plus, X has to be different every time. BSD make saves the day with the .for loop. Is there something like that in SysV make. I don't really remember. Last time I used SysV make was back in 1995. :-) -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
On Tue, Feb 24, 2004 at 11:32:14AM +, Angus Leeming wrote: > Lars Gullik Bjønnes wrote: > > What about the ".for" stuff and LOOPLIST? > > > > (.for might be part of "standard" make, but I have never seen it > > used before.) > > I've never seen it before either, but I'll try it out. I have a > machine here running DEC/Compaq/HP Alpha with a 1.3.x cvs tree. I've > applied Zvezdan's patch to qt2/moc/Makefile.am, qt2/ui/Makefile.am > and to qt2/ui/moc/Makefile.am. > > I'm starting the build process from scratch (having just run make > maintainer-clean) and am using DEC/Compaq/HP make to build it. This > flavour of make does not understand the gnu-isms. Let's see if it > understands Zvezdan's alternative. > > No joy :-( Because it's System V Release 4 UNIX, I suppose. And that make is quite another beast. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
On Tue, Feb 24, 2004 at 11:23:47AM +0100, Lars Gullik Bjønnes wrote:
> What about the ".for" stuff and LOOPLIST?
>
> (.for might be part of "standard" make, but I have never seen it
> used before.)
It's BSD make specific.
As I mentioned in the previous mail, the good thing about both GNU make
and BSD make is that if they see a file called GNUmakefile and
BSDmakefile, respectively, they will use that file first, rather than
Makefile.
So, autoconf, I suppose, could be made to produce two files:
Makefile, with GNU pattern rules, and BSDmakefile, with BSD .for loop.
That way, both makes will be happy, and you cover rather big user base
(Linux + {Free,Net,Open}BSD).
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
On Tue, Feb 24, 2004 at 10:36:08AM +0100, Lars Gullik Bjønnes wrote: > Angus Leeming <[EMAIL PROTECTED]> writes: > | Excellent! I tried to remove the gnu make dependency myself but these > | three had me stumped. > > Except that I do not think we can use it as is... the am__object_X > stuff might AFAIK change with different versions of automake. am__object_X is a list of *.lo files in that Makefile. Whenever that list name changes, this name changes in the same manner, right? So that's not an issue. > > (and seeing how much simpler the GNU syntax is I do not really > understand why BSD make does not support this.) > Because the chicken is older than an egg. Or was it the other way around? :-) I guess, because, as my equivalents show, it's not really necessary. It's just a shorthand for .for loop. And every added piece of syntax complicates the parser, which introduces more bugs, blah, blah... This is just a wild guess, I do not know really. I just take it as a fact of life. :-) -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: fixes for lyx-1.3.4 on OpenBSD
John,
On Tue, Feb 24, 2004 at 02:11:02AM +, John Levon wrote:
> On Mon, Feb 23, 2004 at 07:45:15PM -0500, Zvezdan Petkovic wrote:
>
> >automagically defined. The trouble is that LyX's configure doesn't
> >respect MOC and UIC environment variables and overwrites them.
> >
> >The patch mentioned in the comment above that would fix this is
> >attached below.
>
> I can't say I like it. Why can't you just set up PATH correctly ? I'm a
> bit sick of every OS/distro doing Qt versioning in completely different
> ways. Trolltech don't care because they want you to use qmake.
I don't think you quite understood the details.
PATH is set up correctly and has nothing to do with this.
I didn't want to go to all the details in the first email, but it seems
I have to.
LyX configure script looks for moc2 and moc only.
On OpenBSD it happens to be moc3-mt. And it never gets found.
If a user on OpenBSD installs KDE, (s)he will get qt3-mt package
installed as a dependence, as well as qt3-base. If (s)he installs an
older KDE package or a package that doesn't work with thread, that
package will perhaps depend on the plane qt3 or even qt2.
The only way to have all of them in the same time is to call them
differently (moc2, moc3, moc3-qt) and store the libraries in different
places (/usr/local/lib/{qt3,qt2}).
Now, as a port builder you mostly do not worry about this. The OpenBSD
port infrastructure takes care of the gory details in a very simple way.
You just add to your port's Makefile this line:
MODULES+= qt
and if you're going to use threaded version
MODQT_MT= Yes
That is all. A side effect of that is that the important environment
variables for the build process get extra values.
CONFIGURE_ENV, MAKE_ENV, and MAKE_FLAGS all get these values added if
MODQT_MT=Yes (otherwise -mt suffix is not there):
MOC=${LOCALBASE}/bin/moc3-mt
UIC=${LOCALBASE}/bin/uic3-mt
MODQT_INCDIR=${LOCALBASE}/lib/qt3
MODQT_LIBDIR=${LOCALBASE}/include/X11/qt3
Besides, the variable LIB_DEPENDS, which determines port (package)
dependencies is added to appropriately (lib/qt3/qt-mt.3::x11/qt3,mt),
and the important variable is set:
MODQT_CONFIGURE_ARGS= --with-qt-includes=${MODQT_INCDIR} \
--with-qt-libraries=${MODQT_LIBDIR}
Notice that as a port builder you do _not_ worry about this. This is
all performed by a ports infrastructure on your behalf because of the
line MODULES+=qt.
So, what I did in the port is after MODULES+=qt and MODQT_MT=Yes lines:
CONFIGURE_ARGS= --with-frontend=qt ${MODQT_CONFIGURE_ARGS}
which adds the correct Qt include and library directory paths using
configure options --with-qt-{includes,libraries}.
Since CONFIGURE_ENV _already_ has MOC=${LOCALBASE}/bin/moc3-mt and
UIC=${LOCALBASE}/bin/uic3-mt I'd expect a configure script to respect
these environment variables. It doesn't. Because it looks for moc2 and
moc without even checking whether these variables are set.
That's why I had to trick them that ac_moc1 and ac_uic are already set
by passing this:
CONFIGURE_ENV+= ac_moc1=${MODQT_MOC} ac_uic=${MODQT_UIC}
However, it's sort of a hack, rather than a solution.
Respecting $MOC and $UIC would be a solution. That's what the qt.m4
patch was all about.
>
> > 2. Another problem with configure is that it tries to compile a simple
> >Qt application testing different variants of Qt libraries, but
> >doesn't link thread library with qt-mt. Since OpenBSD installs qt-mt
> >by default (for KDE) that was the version I had. Configure couldn't
> >pass because the libraries wouldn't link without -pthread.
> >
> >Fix was to pass LIBS=-pthread to configure
> > ->8-
> > CONFIGURE_ENV= LIBS=-pthread
> > ->8-
> >This line was used before the += line shown in the previous point.
> >
> >I don't know if this is platform specific or not. Keeping this line
> >in the port is not big deal if it's OpenBSD specific.
>
> This could easily go wrong ... I do not know what the correct/best thing
> is here.
As I said, having the above in the OpenBSD port Makefile only is fine
for me. I just wanted you to know, that if the library is -mt it
wouldn't link even the simple configure test program without the
appropriate thread library in the environment. You, as developers may
want to investigate this issue.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
fixes for lyx-1.3.4 on OpenBSD
Hello everybody,
I've built a port of lyx-1.3.4 for OpenBSD.
Here are the fixes I had to apply.
1. OpenBSD uses special conventions for storing qt files to be able to
have multiple versions installed at the same time. When building a
port it's enough to add
MODULES+= qt
to the port Makefile and the appropriate environment variables get
automagically defined. The trouble is that LyX's configure doesn't
respect MOC and UIC environment variables and overwrites them.
I had to apply this trick to get it accept MOC and UIC
->8-
# A hack until LyX developers apply the patch to qt.m4 to respect $MOC and $UIC
CONFIGURE_ENV+= ac_moc1=${MODQT_MOC} ac_uic=${MODQT_UIC}
->8-
MODQT_{MOC,UIC} have the same content as the variables MOC and UIC
which are added to CONFIGURE_ENV automatically by OpenBSD qt port
module.
The patch mentioned in the comment above that would fix this is
attached below.
2. Another problem with configure is that it tries to compile a simple
Qt application testing different variants of Qt libraries, but
doesn't link thread library with qt-mt. Since OpenBSD installs qt-mt
by default (for KDE) that was the version I had. Configure couldn't
pass because the libraries wouldn't link without -pthread.
Fix was to pass LIBS=-pthread to configure
->8-
CONFIGURE_ENV= LIBS=-pthread
->8-
This line was used before the += line shown in the previous point.
I don't know if this is platform specific or not. Keeping this line
in the port is not big deal if it's OpenBSD specific.
3. I had to patch boost platform specific file. It makes a wrong
assumption that all BSDs have macro based . Patch is
attached below.
4. I had to add a compile option:
->8-
CXXFLAGS+= -ftemplate-depth-30
->8-
I'm not sure if this is platform specific, compiler specific
(gcc-2.95.3), or just my machine specific. I have found the
suggestion to do this at
http://www.boost.org/libs/graph/doc/trouble_shooting.html
and it has obviously worked.
5. I have noticed that autoconf checks the make include style by testing
first whether the make file with such a line
include a-file-name
works. If it does it concludes (wrongly) that include style is GNU.
The next test for BSD style make (.include "a-file-name") is not even
attempted. On OpenBSD make accepts the plain include form to keep
compatibility with SysV make. It doesn't mean it prefers that style.
:-)
6. I had to depend on GNU make anyway because of only three files in
src/frontends/qt2. My port has this line to do that.
->8-
USE_GMAKE= Yes
->8-
but I was seriously considering of making the patches for these files
(moc/Makefile, ui/Makefile, ui/moc/Makefile) and use BSD make as
before. It's a shame that such a big program as LyX breaks on BSD
make only because of the few lines in only three Makefiles.
Below is attached a BSD equivalent to GNU make pattern rules.
You may want consider storing those in a file called BSDmakefile.
That would work with BSD make, while Makefile would work with gmake.
I believe that would be all. Thanks, and keep the good work.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
$OpenBSD$
--- config/qt.m4.orig 2002-12-19 18:41:18.0 -0500
+++ config/qt.m42004-02-22 16:29:24.0 -0500
@@ -199,11 +199,15 @@ AC_DEFUN(QT_DO_IT_ALL,
AC_SUBST(QT_INCLUDES)
AC_SUBST(QT_LDFLAGS)
- QT_FIND_MOC
- MOC=$ac_moc
+ if test -z "$MOC"; then
+ QT_FIND_MOC
+ MOC=$ac_moc
+ fi
AC_SUBST(MOC)
- QT_FIND_UIC
- UIC=$ac_uic
+ if test -z "$UIC"; then
+ QT_FIND_UIC
+ UIC=$ac_uic
+ fi
AC_SUBST(UIC)
QT_CHECK_COMPILE
$OpenBSD$
--- boost/boost/config/platform/bsd.hpp.orig2002-05-21 19:39:55.0 -0400
+++ boost/boost/config/platform/bsd.hpp 2004-02-22 02:24:05.0 -0500
@@ -44,7 +44,9 @@
//
// The BSD has macros only, no functions:
//
-#define BOOST_NO_CTYPE_FUNCTIONS
+#if !defined(__OpenBSD__)
+# define BOOST_NO_CTYPE_FUNCTIONS
+#endif
//
// thread API's not auto detected:
-8<-
LOOPLIST = ${am__objects_1:.lo=.C}
.for file in ${LOOPLIST}
${file}: ${srcdir}/../${file:_moc.C=.h}
${MOC} -o $@ $?
.endfor
# The above is a BSD make equivalent of the following GNU make rule
#%_moc.C: $(srcdir)/../%.h
# $(MOC) -o $@ $<
-8<-
LOOPLIST = ${am__objects_1:.lo=.C}
.for file in ${LOOPLIST}
${file}: ../${file:_moc.C=.h}
${MOC} -o $@ $?
.endfor
# The above is a BSD make equivalent of the following GNU make rule
#%_moc.C: ../%.h
#
Re: patch for lyx.spec
On Fri, Feb 21, 2003 at 10:23:06AM +0100, Jean-Marc Lasgouttes wrote: > Kayvan, another thing to think about: how difficult would it be to > create > lyx-common.noarch.rpm > lyx-xforms.i386.rpm > lyx-qt.i386.rpm > where the last two would probably only contain the rpm, I guess. It's fairly easy to produce multiple rpm binary packages from a single source rpm. However, having two different executables would require two compilations during the building of rpms. Also, the executables would have to be called differently. One of them can be called lyx, the other perhaps lyx-qt. -- Zvezdan Petkovic
xforms.spec file
Please find attached the xforms.spec file. The changes I did are
described in its changelog.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
%define libname libforms
%define rel release
%define xdirX11R6
Summary: XForms library
Name: xforms
Version: 1.0
Release: %{rel}
License: LGPL
Group: System/Libraries
Source0: ftp://ncmir.ucsd.edu/pub/xforms/OpenSource/%{name}-%{version}-%{rel}.tgz
Patch: %{name}-%{version}.patch
URL: http://world.std.com/~xforms/
BuildRequires: xpm
BuildRequires: xpm-devel
BuildRoot: %{_tmppath}/%{name}-%{version}-root
Prefix: %{_prefix}
%description
XForms is a GUI toolkit based on Xlib for X Window Systems. It features
a rich set of objects, such as buttons, sliders, menus, etc., integrated
into an easy and efficient object/event callback execution model that
allows fast and easy construction of X-applications. In addition, the
library is extensible and new objects can easily be created and added to
the library.
%package -n %{libname}
Summary: XForms Library
Group: System/Libraries
Provides: %{name}
%description -n %{libname}
XForms libraries.
%package -n %{libname}-devel
Summary: XForms development header files
Group: System/Libraries
Requires: %{libname}
Provides: %{name}
%description -n %{libname}-devel
XForms header files for development.
%prep
%setup -q -n %{name}-%{version}-%{rel}
%patch -p0
%build
xmkmf -a
make
%install
rm -rf $RPM_BUILD_ROOT
#%makeinstall_std DESTDIR=$RPM_BUILD_ROOT OWNER=$UID GROUP=$GROUPS
make install DESTDIR=$RPM_BUILD_ROOT
make install.man DESTDIR=$RPM_BUILD_ROOT
%post -n %{libname}
/sbin/ldconfig
%postun -n %{libname}
/sbin/ldconfig
%clean
rm -rf $RPM_BUILD_ROOT
%files -n %{libname}
%defattr(-,root,root)
%doc COPYING Copyright Changes 00README README.OS2
%attr(755,root,root) %{_prefix}/%{xdir}/bin/*
%attr(755,root,root) %{_prefix}/%{xdir}/lib/*.so.*
%files -n %{libname}-devel
%defattr(-,root,root)
%attr(755,root,root) %{_prefix}/%{xdir}/lib/*.a
%attr(644,root,root) %{_prefix}/%{xdir}/include/X11/*
%changelog
* Mon Feb 10 2003 Zvezdan Petkovic <[EMAIL PROTECTED]>
- Commented out makeinstall_std because of use of fg without job control error.
- Fixed the permissions for include files to 644 instead of 755.
- Added Requires to devel package to make it depend on libforms.
- Added xdir macro so that package can be moved easier.
- Used macros consistently.
* Sat Dec 21 2002 Kayvan A. Sylvan <[EMAIL PROTECTED]> 1.0-release
- Fixed up patch from 1.0-RC5.
* Sun Oct 6 2002 Kayvan A. Sylvan <[EMAIL PROTECTED]> 1.0-RC5
- Changed /usr/local to /usr/X11R6 and set BuildGL to NO.
- Added %post and %postun section for /sbin/ldconfig.
* Sun Jul 14 2002 Greg Hosler <[EMAIL PROTECTED]> 1.0-RC4
- Pass DESTDIR to makeinstall_std.
* Thu Jul 11 2002 Peter Galbraith <[EMAIL PROTECTED]> 1.0-RC4
- Move from libxforms to libforms to match other distros.
* Tue Jul 8 2002 Chris Freeze <[EMAIL PROTECTED]> 1.0-RC4
- First stab at spec file.
patch for lyx.spec
Please, find attached the patch for lyx.spec file used for building
RPMs. I have introduced two new macros at the top which enable the
creation of new RPM with a different frontend by simply changing their
definition. For example, to build xforms frontend (since the original
spec file used Qt frontend), one just needs to redefine:
%define frontend xforms
%define frontdep libforms >= 1.0
The rest of the spec file stays untouched and everything works great!
I've built xforms and qt rpms without a problem.
I had to enforce the consistent use of macros to achieve this though.
I urge you to accept the patch -- it will make rpm builder's life
much easier.
One thing I have in my personal lyx.spec file that I didn't put here is
the creation of lyx.desktop file, which creates a menu entry for both
Gnome and KDE. It resides in %{_sysconfdir}/X11/applnk/Applications.
However, I didn't know whether other Linux distributions use the same
place to store .desktop files (on Red Hat it is /etc/X11/applnk), so I
left it out. If you think it can be useful I can send the patch for
that too.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
--- lyx.spec.orig Fri Feb 7 03:58:10 2003
+++ lyx.specWed Feb 19 17:51:37 2003
@@ -1,16 +1,19 @@
+%define frontend qt
+%define frontdep qt >= 2.2.1
+
Summary: A WYSIWYM (What You See Is What You Mean) frontend to LaTeX
Name: lyx
Version: 1.3.0
-Release: 1qt
+Release: 1_%{frontend}
License: see COPYING file
-Group: X11/Editors
-Url: http://www.lyx.org/
+Group: Applications/Publishing
+URL: http://www.lyx.org/
Packager: Kayvan A. Sylvan <[EMAIL PROTECTED]>
-Source: ftp://ftp.lyx.org/pub/lyx/stable/lyx-%{PACKAGE_VERSION}.tar.gz
-BuildRoot: %{_tmppath}/%{name}-root
-Icon: lyx.xpm
-Prefix: /usr
-Requires: qt >= 2.2.1, pspell, tetex-xdvi, tetex, tetex-latex
+Source: ftp://ftp.lyx.org/pub/lyx/stable/%{name}-%{version}.tar.gz
+BuildRoot: %{_tmppath}/%{name}-%{version}-root
+Icon: %{name}.xpm
+Prefix: %{_prefix}
+Requires: %{frontdep}, pspell, tetex-xdvi, tetex, tetex-latex
Obsoletes: tetex-lyx
%description
@@ -30,14 +33,15 @@
With LyX, the author can concentrate on the contents of his writing,
and let the computer take care of the rest.
-This is LyX built with the Qt frontend.
+This is LyX built with the %{frontend} frontend.
+
%prep
%setup
%build
unset LINGUAS || true
-./configure --with-pspell --with-frontend=qt --prefix=%{_prefix} --mandir=%{_mandir} \
- --bindir=%{_bindir} --datadir=%{_datadir} \
+./configure --with-pspell --with-frontend=%{frontend} --prefix=%{_prefix} \
+ --mandir=%{_mandir} --bindir=%{_bindir} --datadir=%{_datadir} \
--without-warnings --disable-debug --enable-optimization=-O2
make
@@ -53,13 +57,13 @@
#
TEXMF=%{_datadir}/texmf
mkdir -p ${RPM_BUILD_ROOT}${TEXMF}/tex/latex
-mv ${RPM_BUILD_ROOT}%{_datadir}/lyx/tex \
- ${RPM_BUILD_ROOT}/${TEXMF}/tex/latex/lyx
+mv ${RPM_BUILD_ROOT}%{_datadir}/%{name}/tex \
+ ${RPM_BUILD_ROOT}/${TEXMF}/tex/latex/%{name}
#
# Miscellaneous files
#
-cp -a lib/images/lyx.xpm ${RPM_BUILD_ROOT}%{_datadir}/lyx/images/
+cp -a lib/images/%{name}.xpm ${RPM_BUILD_ROOT}%{_datadir}/%{name}/images/
cp lib/reLyX/README README.reLyX
%clean
@@ -78,14 +82,13 @@
# Now configure LyX
#
echo "Configuring LyX for your system..."
-cd %{_datadir}/lyx
+cd %{_datadir}/%{name}
./configure --srcdir
# Fix reLyX perl program if the prefix is non-standard
if [ "%{_prefix}" != "/usr" ]
then
-perl -pi -e "s!/usr/share/lyx!%{_datadir}/lyx!" \
-%{_bindir}/reLyX
+perl -pi -e "s!/usr/share/%{name}!%{_datadir}/%{name}!" %{_bindir}/reLyX
fi
%postun
Re: kmap quote character doesn't work
I repeat my question from the last week since nobody bothered to reply.
Why the quote character doesn't get remaped properly?
See the explanation below. This has worked in 1.1.6. It doesn't in
1.2.3 and 1.3.0 (neither xforms nor qt)
On Thu, Feb 13, 2003 at 11:02:52PM -0500, Zvezdan Petkovic wrote:
> \kmap @ \"
> \kmap \" "\\'{C}"
>
> On pressing @, I get " and I should be getting the quotation sign <<
> and >>, and on pressing " I get quotation signs << >>, instead of
> capital C' (C acute).
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: Bibliography
On Sun, Feb 16, 2003 at 11:57:16PM +0100, Jade wrote:
> I have found a package called "jurabib", that can help doing this, but I
> wondered if there was a way to include the package automatically using Lyx.
Did you try putting \usepackege{jurabib} in your LaTeX preamble?
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
kmap quote character doesn't work
Hello,
I haven't been on the list for a long time, so sorry if the same
question has been already asked. (I couldn't find anything on
bugzilla).
A few years ago I've made kmap file for serbocroatian.
Last time I used it in 1.1.6fix4 it worked fine.
Today I have installed 1.3.0 (and then just to check 1.2.3) and the
result was the same. The quote key mappings do not work (see
/usr/share/lyx/kbd/serbocroatian.kmap):
\kmap @ \"
\kmap \" "\\'{C}"
On pressing @, I get " and I should be getting the quotation sign <<
and >>, and on pressing " I get quotation signs << >>, instead of
capital C' (C acute).
Where should I start looking for the bug (kbmap.C wasn't too
promising)? This was obviously introduced between 1.1.6 and 1.2.x.
Regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: Feedback from www.lyx.org
On Wed, Jan 23, 2002 at 02:56:29PM +0100, Jean-Marc Lasgouttes wrote: > >>>>> "Zvezdan" == Zvezdan Petkovic <[EMAIL PROTECTED]> writes: > > Zvezdan> I also propose a patch to spec file that is attached to > Zvezdan> enforce use of macros everywhere. > > Could you send a proper version against lyx.spec.in from branch > BRANCH_1_1_6 of cvs, with a ChangeLog entry? > > JMarc Yes. In the next few days. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Feedback from www.lyx.org
On Sun, Jan 20, 2002 at 12:24:10PM +0200, Dekel Tsur wrote: > On Sun, Jan 20, 2002 at 04:50:05AM -0500, Zvezdan Petkovic wrote: > > On Sun, Jan 20, 2002 at 11:24:17AM +0200, Dekel Tsur wrote: > > > On Sat, Jan 19, 2002 at 03:53:44PM -0800, Kayvan A. Sylvan wrote: > > > > It may be a good idea to set up a matrix of libc-versus-xforms versions > > > > for this so that our users can choose the one most likely to work on > > > > their system. > > > > > > How about supplying a statically linked binary ? > > > > IMO it beats the purpose of RPM. One can run > > > > rpm --rebuild lyx-1.1.6fix4-1.src.rpm > > > > if binary is not all right for their platform. > > What I meant is putting a statically linked binary in a .tar.gz. > It has the advantage of working in ANY linux distribution. I see your point. People who want to use it could do that. Probably quite fine for home use. Statically linked binaries have many disadvantages for big networked installations though. For example, a recent bug/security bunch of fixes for glibc in RH7.2/7.1 would have no benefit on a statically linked binary (not to mention a security vulnerability lurking somewhere in it). I'd rather recompile. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Feedback from www.lyx.org
On Sun, Jan 20, 2002 at 11:24:17AM +0200, Dekel Tsur wrote: > On Sat, Jan 19, 2002 at 03:53:44PM -0800, Kayvan A. Sylvan wrote: > > It may be a good idea to set up a matrix of libc-versus-xforms versions > > for this so that our users can choose the one most likely to work on > > their system. > > How about supplying a statically linked binary ? IMO it beats the purpose of RPM. One can run rpm --rebuild lyx-1.1.6fix4-1.src.rpm if binary is not all right for their platform. But we do have to supply binary for the _latest_ Red Hat (7.2 or 7.1) instead of four releases old distribution (6.1). Such a binary should install without trouble on Suse, Mandrake, etc. What bothers me more is that OpenBSD port of LyX is marked broken in OpenBSD 3.0. Reason given: "does not build in any reproducible way". I plan to contact the port maintainer to see if fix4 changes things in any way. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Feedback from www.lyx.org
On Sat, Jan 19, 2002 at 10:59:32PM +0100, Jean-Marc Lasgouttes wrote:
>
> This is becquse they have been produced on rh6.1. Not sure what the
> proper fix is. The 0.88 vs 0.89 thing cannot be solved easily, since the
> two versions are binary incompatible (note that you should not use any
> xforms library provided by redhat, they are buggy; use the ones from
> ftp.lyx.org/pub/lyx/contrib instead).
>
Anyone who wants to use 0.89 xforms must rebuild from source RPM.
Regarding a build on RH6.1 it's really way too old. Do you need a
rebuild for RH7.2? I can do that, but I guess so can anybody else with a
newer version of RH on his machine.
I also propose a patch to spec file that is attached to enforce use of
macros everywhere.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
--- lyx-1.1.6.spec Fri Jan 11 19:31:16 2002
+++ lyx-1.1.6zp.specMon Jan 14 20:37:04 2002
@@ -3,13 +3,13 @@
Version: 1.1.6fix4
Release: 1
Copyright: see COPYING file
-Group: X11/Editors
+Group: Applications/Publishing
Url: http://www.lyx.org/
Packager: Kayvan A. Sylvan <[EMAIL PROTECTED]>
-Source: ftp://ftp.lyx.org/pub/lyx/stable/lyx-%{PACKAGE_VERSION}.tar.gz
-BuildRoot: %{_tmppath}/%{name}-root
-Icon: lyx.xpm
-Prefix: /usr
+Source: ftp://ftp.lyx.org/pub/lyx/stable/%{name}-%{version}.tar.gz
+BuildRoot: %{_tmppath}/%{name}-%{version}-root
+Icon: %{name}.xpm
+Prefix: %{_prefix}
Requires: xforms >= 0.88, tetex-xdvi, tetex, tetex-latex
Obsoletes: tetex-lyx
@@ -53,13 +53,13 @@
#
TEXMF=%{_datadir}/texmf
mkdir -p ${RPM_BUILD_ROOT}${TEXMF}/tex/latex
-mv ${RPM_BUILD_ROOT}%{_datadir}/lyx/tex \
- ${RPM_BUILD_ROOT}/${TEXMF}/tex/latex/lyx
+mv ${RPM_BUILD_ROOT}%{_datadir}/%{name}/tex \
+ ${RPM_BUILD_ROOT}${TEXMF}/tex/latex/%{name}
#
# Miscellaneous files
#
-cp -a lib/images/lyx.xpm ${RPM_BUILD_ROOT}%{_datadir}/lyx/images/
+cp -a lib/images/%{name}.xpm ${RPM_BUILD_ROOT}%{_datadir}/%{name}/images/
cp lib/reLyX/README README.reLyX
%clean
@@ -78,14 +78,13 @@
# Now configure LyX
#
echo "Configuring LyX for your system..."
-cd %{_datadir}/lyx
+cd %{_datadir}/%{name}
./configure --srcdir
# Fix reLyX perl program if the prefix is non-standard
if [ "%{_prefix}" != "/usr" ]
then
-perl -pi -e "s!/usr/share/lyx!%{_datadir}/lyx!" \
-%{_bindir}/reLyX
+perl -pi -e "s!/usr/share/%{name}!%{_datadir}/%{name}!" %{_bindir}/reLyX
fi
%postun
Re: Enough is enough! (lyx 1.1.6fix4)
On Fri, Jan 11, 2002 at 01:04:02PM -0800, Kayvan A. Sylvan wrote: > > Hi Zvezdan, > > I moved the xforms RPMs to old/ and instead put a note in xforms-README > to point people in the right direction. Hopefully xforms will have a > true src.rpm soon, instead of this chicanery. > > Is this okay? > It's OK. Having them on the official ftp site soon would be better though. :-) > -- > > IMPORTANT NOTE ABOUT LYX AND XFORMS > > > The correct xforms binary must be installed on your system for LyX > to work. > > You can get the latest RPMs for glibc-2.1 systems (most Redhat 6.X and > Redhat 7.X systems) here: > > http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/xforms-0.88-4.src.rpm > http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/xforms-0.88-4.i386.rpm > > If you wish to get the latest xforms (which *does* work with LyX, but > has not been officially endorsed), then you can get it here: > > http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/xforms-0.89-1.src.rpm > http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/xforms-0.89-1.i386.rpm > > These are not real source RPMs. They install a binary tarball which must > be compatible with your architecture and libc combination. If LyX will > not start up (or starts up and core dumps), then it is a fair bet that you > have installed the wrong version of XForms for your machine. > > -- -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Enough is enough! (lyx 1.1.6fix4)
On Fri, Jan 11, 2002 at 06:25:04PM +0100, Jean-Marc Lasgouttes wrote: > > SO this time I decided not to warn that I was about to prepare > 1.1.6fix4, in order to avoid people checking in last minute changes > and sending files to me. You can beg, you can cry, it is too late: I > tagged it under the funky name lyx-1_1_6fix4. > > Kayvan, could you prepare rpms? I probably will do the upload and > announcement on monday. > And please make sure that xforms rpms are correct too. The ones that were on LyX ftp site before install xforms even if one just compiles source rpm because they did not use RPM_BUILD_ROOT. You can pick up the correct spec file (or rpms for that matter) at: http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/ [The lyx.spec changes from the above page have already been included in LyX rpms] Regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Roman text inside math; How?
On Sun, Nov 11, 2001 at 09:17:42AM +0900, R. Lahaye wrote:
>
> with "V" in the usual math-slanted style, but with "t"
> as subscribted roman-straight style. For the roman style
> I used "\rm{...}" in TeX mode, however, that seems not to
> be possible in CVS version anymore.
>
The proper way in LaTeX to enter roman text in math is \mathrm{...}.
That works in LyX math mode too. Simply start typing \ and continue with
the rest. I do not know about CVS version though. Give it a shot...
Cheers,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: BUGREPORT: lyxpipe deletion
On Thu, Oct 18, 2001 at 06:24:28PM +0200, Moritz Moeller-Herrmann wrote:
> On Thu, Oct 18, 2001 at 03:34:27PM +0200, Andre Poenitz wrote:
> >On Thu, Oct 18, 2001 at 03:21:02PM +0200, Moritz Moeller-Herrmann wrote:
> >> I have to support some less clueful windows user who switched to lkinux
> >> because of me and stuff like this is a nightmare. ("Simple, just type ALT
> >> + F2 konsole [Enter] then rm ~/.lyx/lyxpipe.* After this restart lyx and
> >> pybliographer, aren't you glad you left windows?"]
> >
> >put them
> >
> >alias lyx='rm -f ~/.lyx/lyxpipe.* ; lyx'
> >
> >into .bashrc...
> >
> >
> >But you certainly have a point here.
>
> This as you point out, does NOT work for KDE desktops. And the people I
> transferred from Windows to KDE have no intention to use xterms :-)
>
Then put in their .profile (or .bash_profile)
[ -p ~/.lyx/lyxpipe.in ] && rm -f ~/.lyx/lyxpipe.in
[ -p ~/.lyx/lyxpipe.out ] && rm -f ~/.lyx/lyxpipe.out
or if they are using (t)csh then in their .login put
if ( -p ~/.lyx/lyxpipe.in ) rm -rf ~/.lyx/lyxpipe.in
if ( -p ~/.lyx/lyxpipe.out ) rm -rf ~/.lyx/lyxpipe.out
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: trivial latex-y question (printing print time) & doc. question
On Mon, Sep 24, 2001 at 04:02:38PM -0400, [EMAIL PROTECTED] wrote:
> herbert held,
>
>
> > On Mon, 24 Sep 2001 [EMAIL PROTECTED] wrote:
>
> > > herbert harumphed,
> > > > "Richard E. Hawkins" wrote:
>
> > > Thanks, but this gives the error, File `time.sty' not found.
>
> > package time available at CTAN
>
> OK, now for the sillier part: I don't see anything in the preferences
> or document specification to tell lyx where to look for these. I know
> I did osmething similar with my dissertation, but it doesn't seem to be
> on this machine for some reason.
>
It's in teTeX documentation. Basically there are several approaches. The
quickest is to
setenv TEXINPUTS directory/where/your/styles/are//:
in your .login, or if you use sh .profile (export TEXINPUTS=...)
Colon ':' could be at the beginning also, but it must be somewhere to
inherit the existing TEXINPUTS path. Double slash is needed to include
all the subdirectories too.
Then run your LyX Edit->Reconfigure menu option.
Trouble begins when you want to add new BibTeX files, new fonts, etc.
and you begin using BIBINPUTS, ...
The most general solution (that I use) is to create a directory in your
home directory called texmf. The subdirectories should be arranged in
the same way as the system texmf tree (see teTeX documentation again --
$TEXMF/doc/index.html). Of course you define only those you need (e.g.
texmf/tex/latex/misc/time.sty or texmf/fonts/type1/adobe/Cheq/). Then in
your .login (adjust for .profile if you use sh) put only this:
setenv TEXMF '{~/texmf,\!\!$TEXMFLOCAL:\!\!$TEXMFMAIN}'
and it will see all your texmf subtree additions including fonts,
bibtex styles, latex styles, ...
All this is described in the above mentioned teTeX documentation (see
links to FAQ and TeX directory structure.
Hope this puts you on the right path. :-)
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Thu, Sep 20, 2001 at 06:12:00PM +0200, Lars Gullik Bjønnes wrote: > Zvezdan Petkovic <[EMAIL PROTECTED]> writes: > | Does that mean that DESTDIR takes care of all that through autoconf? > > Yes. The DESTDIR is the autoconf (automake really) way of doing > staging directories/installs. > > -- > Lgb Thanks Lars. It's neat because it saves a lot of space on make install line. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Thu, Sep 20, 2001 at 02:04:56AM -0700, Kayvan A. Sylvan wrote:
> > IMHO, this patch is broken:
> >
> > Try the steps below and you will probably understand, what I have been
> > talking about (_bindir and _datadir not being treated correctly):
>
> Allright, then. Here is the corrected patch.
>
> Someone please apply this one...
> ...
> -./configure --prefix=/usr --without-warnings --with-included-string
> +./configure --prefix=%{_prefix} --mandir=%{_mandir} \
> + --bindir=%{_bindir} --datadir=%{_datadir} \
> + --without-warnings --with-included-string
> ...
> +make DESTDIR=${RPM_BUILD_ROOT} install
Kayvan, I have a small question.
Is the make line above enough after you did changes to configure?
I remember that we had before: mandir=${RPM_BUILD_ROOT}%{_mandir}
Now I do not see neither that nor something as:
bindir=${RPM_BUILD_ROOT}%{_bindir}.
Does that mean that DESTDIR takes care of all that through autoconf?
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Wed, Sep 19, 2001 at 06:06:41PM +0200, Jean-Marc Lasgouttes wrote:
> >>>>> "Zvezdan" == Zvezdan Petkovic <[EMAIL PROTECTED]> writes:
>
> >> If using %{_bindir} and %{_datadir} in %files sections and not
> >> passing %{_bindir} and %{_datadir} to configure, users will have to
> >> edit lyx.spec to a larger extent than otherwise if not wanting to
> >> use the defaults provided by rpm.
>
> Zvezdan> Excuse me? To what extent they would have to edit it if
> Zvezdan> instead of %{_bindir} we had /usr/bin or /usr/local/bin
> Zvezdan> there?
>
> Just a point while reading this thread: the $datadir variable is
> used as a #define for main.C and lyx_main.C, which means that whatever
> different flags should be passed to 'make all' (or whatever you use),
> and not only to 'make install'.
>
> JMarc
I think configure takes care of make all.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Wed, Sep 19, 2001 at 07:33:21AM +0200, Ralf Corsepius wrote:
> On Mit, 2001-09-19 at 05:23, Zvezdan Petkovic wrote:
> >
> > IMHO, you are exaggerating.
> Well, of cause, I have to disagree :)
>
> Primarily, it's just convenience for users/installers not wanting to use
> the defaults.
>
The point is that this is a developer's RPM, not user's/installer's.
What I'm afraid of is that we would need to supply many RPMs for various
distros + *BSD packages in that case. :-)
> If using %{_bindir} and %{_datadir} in %files sections and not passing
> %{_bindir} and %{_datadir} to configure, users will have to edit
> lyx.spec to a larger extent than otherwise if not wanting to use the
> defaults provided by rpm.
Excuse me? To what extent they would have to edit it if instead of
%{_bindir} we had /usr/bin or /usr/local/bin there?
>
> I.e. if configure --bindir=%{_bindir} --datadir=%{_datadir} .. was used,
> users/packagers could simply
> %define _bindir whatever1
> %define _datadir whatever2
> at the top of lyx.spec and the rest of lyx.spec would still remain
> valid, if not wanting to use rpm-defaults
>
Besides the install section must be changed too:
make ... bindir=${RPM_BUILD_ROOT}%{_bindir} ... install
And that is all they need to change. It would work just fine for them.
I must admit it is easy to put two more flags on the configure and
install line. What other people think about this?
The question is only do we want to have unnecessary flags in the default
RPM simply because of a few people who do not want to use the defaults,
and who would need to edit source spec file anyway to redefine the
paths?
I suppose this is something to think about and have a majority vote.
It's certainly not too much of a burden in the spec file.
> > Please, give me an example of a single version of UNIX that doesn't put
> > its executables in $prefix/bin and its shared data in $prefix/share!
> > I know of none.
>
> I know several, SuSE's lyx.spec is one of these.
>
> > And that is true for $prefix = /usr or $prefix = /usr/local
> >
> > We want to install LyX files to the usual place where that distro puts
> > executables. And that is either /usr/bin or /usr/local/bin.
> SuSE uses --prefix=/usr --bindir=/usr/X11R6/bin for lyx.
>
Well, even prefix=/usr/X11R6 could have worked with bin and share
subdirectories, but some people simply have to do it differently. :-)
I would simply make a separate RPM for SuSE users. As we conversed
above, they have to edit source spec anyway to add defines at the top.
Some people do not know how to do that. It's just simpler for us to have
somebody of developers to apply the above mentioned changes and roll an
RPM for SuSE only.
> > More precisely, we want to use the same prefix as other programs, and
> > that is again /usr or /usr/local
> Well, but users might have diffent needs, eg. if installing multiple
> versions of lyx in parallel.
>
Yes, and some users may not have root privileges. My point is, again, do
we need to provide for the whole world? Such users should compile from
source in my opinion. Non-root users would have great difficulties to
apply RPM anyway. It's possible -- I'm testing RPMs I build as a
non-root user first -- but guys who know how to setup that, know how to
change a few lines in the spec file too.
> > On a properly installed system rpm will use one of the above as a
> > %{_prefix}. Take a look at the defaults from /usr/lib/rpm/macros:
> I know them ;)
> [..]
> > Only if there's a _radical_ change from the default autoconf or rpm
> > installation these things would be different. And a final question:
> > Who in the right mind would change $prefix/bin and $prefix/share to
> > something else for any $prefix?
> Cf. above. SuSE prefers installing X11-based programs to /usr/X11R6/bin
> but installs there support files to /usr/*. You might consider this to
> be dumb, but I am not the one to blame for (I am not affiliated to SuSE
> at all).
>
I don't think it's dumb. I think it's arrogant to change the traditional
things even when not necessary. Traditionally, only the programs
necessary for X and showing the capabilities of X (e.g. xedit) go into
/usr/X11R6. But there are many non-traditional things in Linux distros
for better or for worse.
> Another scenario, where this occasionally happens, is if actually
> sharing architecture-independent files (mandir, infodir, datadir,
> docdir) across architectures in heterogenious networks. However this
> topic is beyond the scope of this thread, IMHO.
>
That was my point all the time. Should we provide for all possible
scenarios in the developer's RPM. In such cases, this is not left to a
regular user but to a good
Re: rpm lyx.spec.in
On Wed, Sep 19, 2001 at 04:46:40AM +0200, Ralf Corsepius wrote:
> On Die, 2001-09-18 at 20:32, Kayvan A. Sylvan wrote:
> > You are very correct.
> >
> > I was overtaken with the moment!
> >
> > Forgive me. I sent a corrected one to Lars. It is also attached here.
> >
>
> Now you're using %{_bindir} and %{_datadir}, too.
>
> IMO, therefore you also might want to pass --bindir=%{_bindir} and
> --datadir=%{_datadir} to configure to guarantee consistency between the
> values used by the autotools and the ones provided by rpm.
>
> Otherwise installers rsp. rpm could use different values for
> %{_ defined values.
>
> Ralf
>
IMHO, you are exaggerating.
Please, give me an example of a single version of UNIX that doesn't put
its executables in $prefix/bin and its shared data in $prefix/share!
I know of none.
And that is true for $prefix = /usr or $prefix = /usr/local
We want to install LyX files to the usual place where that distro puts
executables. And that is either /usr/bin or /usr/local/bin.
More precisely, we want to use the same prefix as other programs, and
that is again /usr or /usr/local
On a properly installed system rpm will use one of the above as a
%{_prefix}. Take a look at the defaults from /usr/lib/rpm/macros:
#==
# configure macros.
# Macro(s) slavishly copied from autoconf's config.status.
#
%_prefix/usr
%_exec_prefix %{_prefix}
%_bindir%{_exec_prefix}/bin
%_sbindir %{_exec_prefix}/sbin
%_libexecdir%{_exec_prefix}/libexec
%_datadir %{_prefix}/share
%_sysconfdir%{_prefix}/etc
%_sharedstatedir%{_prefix}/com
%_localstatedir %{_prefix}/var
%_lib lib
%_libdir%{_exec_prefix}/%{_lib}
%_includedir%{_prefix}/include
%_oldincludedir /usr/include
%_infodir %{_prefix}/info
%_mandir%{_prefix}/man
Notice that comment about copying from autoconf defaults.
Since autoconf defaults are also
bindir = $prefix/bin
datadir = $prefix/share
we are safe.
Notice also that architecture specific file
/usr/lib/rpm/i386-linux/macros later redefines:
%_infodir %{_prefix}/share/info
%_mandir%{_prefix}/share/man
That is the only difference between autoconf and rpm. That is why
%{_mandir} only has to be passed to configure.
Only if there's a _radical_ change from the default autoconf or rpm
installation these things would be different. And a final question:
Who in the right mind would change $prefix/bin and $prefix/share to
something else for any $prefix?
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Tue, Sep 18, 2001 at 10:52:22AM -0700, Kayvan A. Sylvan wrote:
>
> This should make Zvezdan happy, I hope. ;-)
>
Certainly. :-)
> +%{_prefix}
It's rather unfortunate that I pressed `r' instead of `L' and my reply
went to Lars only instead of whole list but here's a copy that explains
why this is _THE BAD_ thing to do.
==
For goodness sake _NO_! By using just %{_prefix_} you declare yourself
as an owner of the whole /usr subtree! Is that _really_ what you want?
Imagine just a small bug in RPM: on uninstall or update of LyX you would
remove whole /usr subtree. Please, read Maximum RPM on www.rpm.org about
the dangers of bad declarations in the %files section.
==
In addition, even on a non-buggy RPM if you do rpm -evvv lyx you'll see
that it schedules /usr for removal, but then (fortunately) finds out
that fileutils owns /usr too and doesn't remove it. There are a few KDE
packages that are badly built and claim that they own /usr. We don't
want to be like them, do we?
Kayvan, please stick with the %files section from my spec file. It's
well defined.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Tue, Sep 18, 2001 at 10:36:04AM +0200, Lars Gullik Bjønnes wrote:
> Zvezdan Petkovic <[EMAIL PROTECTED]> writes:
>
> | No, %{_mandir} is just fine since /usr/lib/rpm/i386-linux/macros define
> | _mandir as /usr/share/man just as /etc/man.conf does.
>
> No single manpath is defined in man.conf[ig]...
>
True, but the most important one in /etc/man.config (man.conf on *BSD) is:
MANPATH /usr/share/man
which corresponds to /usr/lib/rpm/i386-linux/macros
%_mandir%{_prefix}/share/man
The old path /usr/man is still there because of the old packages that
want to install there. That's where LyX man pages were going before I
made that RPM.
> | It seems to be a standard now since my old computer runs OpenBSD 2.9 and
> | it has man pages in /usr/share/man too.
>
> but configure --mandir=%{_mandir} has to be used, it does not work
> without.
>
> --
> Lgb
Quite true. The reason is that Linux distros have completely mixed up
operating system and applications. On a traditional UNIX system, only OS
binaries go into /usr/bin and their man pages in /usr/{share/}man, while
applications go into /usr/local/bin and their man pages into
/usr/local/man. Hence, autoconf default is ${prefix}/bin and
${prefix}/man. That's how things work on a say *BSD and as far as I know
Slackware.
On a Linux distro (e.g. Red Hat) autoconf defaults are still the same,
but /usr/local is empty and thus we have to instruct configure that
mandir position had changed. Rather unfortunate.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Tue, Sep 18, 2001 at 06:39:50AM +0200, Ralf Corsepius wrote:
> On Die, 2001-09-18 at 02:55, Zvezdan Petkovic wrote:
> > On Tue, Sep 18, 2001 at 01:15:07AM +0200, Lars Gullik Bjønnes wrote:
> > Exactly, but on a properly configured system RPM follows /etc/man.conf.
> This is an implementation detail and actually is completely irrelevant
> wrt. this topic.
>
It is _very_ relevant. You might say:
%define _mandir /usr/local/share/man
to define your own place for man pages as a packager, and you are right
in that regard. Packager is the one who decides where they go.
However, I usually do _not_ want to put the man pages in some strange
place I want them where the other man pages go. On a properly configured
system /etc/man.config has this:
MANPATH /usr/share/man
and /usr/lib/rpm/macros has
%_mandir%{_prefix}/man
where _prefix == /usr. So, that wouldn't be right for RH7.1, but it
quite right for the older systems or different architectures (e.g.
sparc)
Fortunately, /usr/lib/rpm/i386-linux/macros has this:
%_mandir%{_prefix}/share/man
which is right for i386 RH7.1.
Hence, a packager can use %{_mandir} and know that the man pages will go
in the system wide directory. Yes, I decided as a packager where to put
them, but I simply used what rpm has already provided for me. So, the
configuration is _really_ relevant for the proper use of RPM.
Please, notice also that it would be relevant for an architecture where
man pages go in /usr/man because i386-linux subdirectory wouldn't have
been used.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Tue, Sep 18, 2001 at 01:24:15AM +0200, Lars Gullik Bjønnes wrote:
> [EMAIL PROTECTED] (Lars Gullik Bjønnes) writes:
>
> | | However, similar references should be written as:
> | |
> | | ${RPM_BUILD_ROOT}%{_mandir}/man?/*
> | |
> | | because on some systems it's /usr/man and on some /usr/share/man
> |
> | but it is not rpm that is deciding this, it is the autotools (or
> | manual configuration).
>
> btw. it seems that on RH 7.1 %{_mandir} is /usr/share/man ... but
> "make install" installes it to /usr/man so it fails...
>
> I guess that %{_prefix}/man must be used in the spec file...
>
> --
> Lgb
No, %{_mandir} is just fine since /usr/lib/rpm/i386-linux/macros define
_mandir as /usr/share/man just as /etc/man.conf does.
It seems to be a standard now since my old computer runs OpenBSD 2.9 and
it has man pages in /usr/share/man too.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Tue, Sep 18, 2001 at 01:08:37AM +0200, Lars Gullik Bjønnes wrote:
> [EMAIL PROTECTED] (Lars Gullik Bjønnes) writes:
> ...
> -%attr(-,root,root) /usr/bin/*
> -%attr(-,root,root) /usr/man/*
> -%attr(-,root,root) /usr/share/locale/*/LC_MESSAGES/*
> -%attr(-,root,root) /usr/share/lyx
> -%attr(-,root,root) /usr/share/texmf/tex/latex/lyx
> +%attr(-,root,root) %{_bindir}/*
> +%attr(-,root,root) %{_mandir}/*
> +%attr(-,root,root) %{_datadir}/locale/*/LC_MESSAGES/*
> +%attr(-,root,root) %{_datadir}/lyx
> +%attr(-,root,root) %{_datadir}/texmf/tex/latex/lyx
>
> I am running with this now, and it seems to work well.
>
> --
> Lgb
All this is already done in my spec file and I bet Kayvan is putting
changes in already. I'm running LyX created by RPM from that spec file
for a long time already. Besides, that attr stuff shouldn't be repeated
over and over. This is an excerpt from my spec file
(http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/lyx.spec):
%files
%defattr(-,root,root)
%doc ABOUT-NLS ANNOUNCE COPYING
%doc README UPGRADING ChangeLog NEWS
%doc lib/CREDITS README.reLyX
%config(missingok) %{_sysconfdir}/X11/applnk/Applications/lyx.desktop
%{_bindir}/*
%{_mandir}/man?/*
%{_datadir}/locale/*/LC_MESSAGES/*
%{_datadir}/%{name}
%{_datadir}/texmf/tex/latex/%{name}
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Tue, Sep 18, 2001 at 01:15:07AM +0200, Lars Gullik Bjønnes wrote:
> Zvezdan Petkovic <[EMAIL PROTECTED]> writes:
>
> | However, similar references should be written as:
> |
> | ${RPM_BUILD_ROOT}%{_mandir}/man?/*
> |
> | because on some systems it's /usr/man and on some /usr/share/man
>
> but it is not rpm that is deciding this, it is the autotools (or
> manual configuration).
>
> --
> Lgb
Exactly, but on a properly configured system RPM follows /etc/man.conf.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Mon, Sep 17, 2001 at 07:03:38PM -0400, Zvezdan Petkovic wrote: > > I'm ranting about this for the third time already, but obviously nobody > wants to take my spec files and make this as it should be. > I'm taking this back. I've just seen in another message that Kayvan is working on this. Cheers, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: rpm lyx.spec.in
On Mon, Sep 17, 2001 at 11:53:02PM +0200, Lars Gullik Bjønnes wrote:
>
> I also think that the:
>
> "gzip -f9 ${RPM_BUILD_ROOT}/usr/man/man?/*"
>
> should be done by the "make install".
>
First of all, this shouldn't be done at all. RPM does this automatically
for man pages.
However, similar references should be written as:
${RPM_BUILD_ROOT}%{_mandir}/man?/*
because on some systems it's /usr/man and on some /usr/share/man
I'm ranting about this for the third time already, but obviously nobody
wants to take my spec files and make this as it should be.
[http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/]
Current rpm files do not work well with the latest Linux distros >= 7.x
as explained on the above web page.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: Changes in LyX
On Thu, Aug 30, 2001 at 02:39:49PM +0300, Sinisa Milivojevic wrote: > No, I did not. My wife installed LyX from SuSE 7.0 and it sent > faxes. How can LyX be installed from SuSE 7.2 or RH 7.1, through a > normal RPM installation. You can find source and binary RPMs compiled for 7.1 at: http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/ I'm not sure whether they will help you with fax sending though :-) Did you try compiling the source RPMs for your fax programs. If the KDE libraries are issue, do you have kde1-compat RPMs installed in your distribution by default. YOu might want to grab them from your CD or KDE ftp site. I'm not implying that you didn't try all this already, just throwing ideas randomly. If Jürgen was a little harsh, that's because you started a little harsh. You can be angry that you favourite feature is missing from you favourite program, but it's a program that you got for free and whatever its developers decide to do with -- it will be done. They won't lose any money, believe me, because of that "bad business decision". Finally, you can't coax people to do something for you by yelling at them or being arrogant. It's so typical of people from our region. It was almost as listening to myself. :-) Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: web page changes
On Thu, Aug 23, 2001 at 07:23:48AM +0200, Andre Poenitz wrote: > On Thu, Aug 23, 2001 at 02:55:54PM +1000, Allan Rae wrote: > > On Thu, 23 Aug 2001, Baruch Even wrote: > > > > > Just to end this thread, the format is now 2001-08-32 > > > > Someone added an extra day to August! > > Now all my calenders are wrong! > > I'll have to reprogram all my clocks and watches! > > Not really. You just have to convert from Baruch's base 9 (which is the > only possible interpretation given the occurence of the digit '8') to > everybody else's base 10. > > Which would mean that Columbus will cross the Atlantic in about 33 years, > of course. > > And that, in turn, means that the whole discussion English vs American > pretty senseless for a while... > I thought this was ISO vs American discussion. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: web page changes
On Tue, Aug 21, 2001 at 12:10:04PM +0200, Herbert Voss wrote: > Zvezdan Petkovic wrote: > > > > On Tue, Aug 21, 2001 at 09:25:39AM +0200, Lars Gullik Bjønnes wrote: > > > Juergen Vigna <[EMAIL PROTECTED]> writes: > > > > > > | However I don't care how this is output on the web-page (we could maybe > > > | have a look at the prefered language set in the browser and have a list > > > | of date-formats and output it in the local (prefered language) mode ;) > > > > > > No, we should use the language of the web page. > > > > > > -- > > > Lgb > > > > And that is _ENGLISH_ -- not "American". > > And in England people write 3rd August 2001. > > i learned in school august, 3rd ... :-) > > but it's some years ago > > Herbert > > -- > http://www.educat.hu-berlin.de/~voss/lyx/ > Well, your English teacher was probably an American. :-) Check out: Practical English Usage by Michael Swan Oxford Univ Press; ISBN: 019431197X In fact it doesn't matter. The only fact is that Lars is being stubborn, first pretending that he doesn't understand 2001-08-03, and than replying to Juergen that the language of the page should be applied. Lars could just plainly say that he doesn't like the look of ISO date format, and that is something I can quite agree with. As the saying goes: `There's no accounting for taste'. I just wanted to show that if we are going to apply false logic here, there is always the way to negate it with even falser logic. And that is why I'm for standard. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: web page changes
On Tue, Aug 21, 2001 at 09:25:39AM +0200, Lars Gullik Bjønnes wrote: > Juergen Vigna <[EMAIL PROTECTED]> writes: > > | However I don't care how this is output on the web-page (we could maybe > | have a look at the prefered language set in the browser and have a list > | of date-formats and output it in the local (prefered language) mode ;) > > No, we should use the language of the web page. > > -- > Lgb And that is _ENGLISH_ -- not "American". And in England people write 3rd August 2001. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: web page changes
On Mon, Aug 20, 2001 at 10:43:24PM +0200, Lars Gullik Bjønnes wrote: > Zvezdan Petkovic <[EMAIL PROTECTED]> writes: > | > | Common Lars. > > Hmrppf... are you belittling me? ( :-) ) > :-) -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: web page changes
On Mon, Aug 20, 2001 at 10:21:08PM +0200, Lars Gullik Bjønnes wrote: > Zvezdan Petkovic <[EMAIL PROTECTED]> writes: > > | On Mon, Aug 20, 2001 at 07:39:14PM +0300, Baruch Even wrote: > | > Output looks like: > | > Page last updated on August 3, 2001. > | > | Why not make it according to international standard: 2001-08-03? > > 8. March?? > > Nah. I like the one with the month spelled out. > > -- > Lgb Common Lars. The standard is the standard and you know well it's -mm-dd. I do like the month spelled too, but as far as I rememeber my 32 years of life in Europe it should be 3 August 2001 (even in UK). And as far as I remember, in Japan it should be 2001/8/3. And, now, when I'm in USA it's 8/3/2001. That's why international standard was introduced. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: web page changes
On Mon, Aug 20, 2001 at 07:39:14PM +0300, Baruch Even wrote: > Output looks like: > Page last updated on August 3, 2001. Why not make it according to international standard: 2001-08-03? -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: RPM files
On Fri, Aug 17, 2001 at 10:07:53AM +0200, Ronny Buchmann wrote: > > is this new with rpm-4.x ? > i've never heard before that it is possible to have more than one Prefix: per spec >file > I don't think it is new. I found it in Maximum RPM book (online version at rpm.org) and the new version covering 4.x is still not out. Still I cannot be absolutely sure. > i will check which macros are available in rpm-3.x used by suse (btw, this is used >by ~70% in germany) Suse and Mandrake are used quite a lot in USA too. My department began using Red Hat long time ago, and it just continues. For me it really doesn't matter. Anything UNIX is fine. Regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: RPM files
On Fri, Aug 17, 2001 at 09:30:24AM +0200, Juergen Vigna wrote: > > On 16-Aug-2001 Zvezdan Petkovic wrote: > > > That way my rpm shell script is executed when I type rpm ... > > It adds my rpmrc file, and it adds my macros file. Hence the topdir and > > tmppath get changed to my directories and I can build as a regular user. > > I did this but then the files in the rpm-package have my user-id and group-id > instead of root:root (or whatever it should have) and then on installing > on systems where I don't have such a user I get warnings about user non > existing substituting with root and I don't like that warnings. What I do > is test-build as user (for the same resons as you) but then build the final > as root. > > Jürgen I think you must have forgotten %files %defattr(-,root,root) Two days ago I've built glimpse rpm and here is the result: -rwxr-xr-x1 root root 362300 Aug 14 17:22 /usr/local/bin/glimpse Although, I also very often rebuild the rpms as root once they build nice under my build tree. Regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: RPM files
On Thu, Aug 16, 2001 at 08:41:23PM +0200, Ronny Buchmann wrote:
> > %define _prefix /usr/local
> you dont need this, use "Prefix: /usr/local" instead
>
Well, I told you I build RPMs quite often. The above is not correct.
You can test and see for yourself.
Prefix: /usr/local
is used to tell rpm that the package is relocatable. Later, when you
install such a package you can say "rpm -i --prefix /opt" and it
would go to /opt instead of /usr/local. If you look in my LyX spec
you'll notice the following:
Prefix: %{_prefix}
Prefix: %{_sysconfdir}
Since I didn't redefine anything, the first one is /usr the second one
is /etc. I need two because KDE desktop entry for LyX goes in
/etc/X11/applnk/Applications/lyx.desktop
Since that path is not in /usr rpm would refuse to build a package
because the first Prefix: tells it that it is relocatable, and it can
see that /etc is not under /usr and hence can't find the way to relocate
it. Only defining two Prefix: entries the package is built as
relocatable. It can be relocated using, for example:
rpm -i --relocate /usr=/opt -relocate /etc=/etc/opt
However, %{_prefix} is a macro used for the place where rpm will store
the files of this package, in the same way autoconf uses prefix so you
can later say ./configure --prefix=whatever ...
In the above example: if I use only Prefix: /usr/local and than use
%{_prefix} macro, rpm will still see %{_prefix}=/usr and will refuse to
build the package since I am claiming that package is relocatable with
the Prefix: /usr/local, but my files have prefix /usr.
Hence, one needs both in case one wants relocatable package. In case of
non-relocatable packages %{_prefix} is enough. If you do not want to
change the default path from /usr you simply use it -- no need to
redefine it.
As an aside, it will work if Prefix: /usr but only because it is equall
to %_prefix in that case.
> > %define _mandir %{_prefix}/man
> > %define _defaultdocdir %{_prefix}/doc
> >
> > This is definitely _THE_ solution to our maintanance problems. With each
> > new version of LyX only the version number must be changed, and it
> > should build and install correctly on any rpm based distro.
> the question is, which *internal* macros are supported
> e.g. with rpm-4.x _datadir is already defined but i think this is introduced with
>rpm-4
> mandrake introduced macros which are (were?) not supported in redhat (_icondir as
>far as i remember)
>
I used only the macros derived from autoconf. I suspect they were
introduced to rpm rather early, although I cannot be quite sure.
> do you where these internal macros are defined? is this done in compile time?
>
> --
> ronny
On a RH system they are in /usr/lib/rpm/macros and
/usr/lib/rpm/%{_target}/macros (where %{_target}=i386-linux, for
example)
The system wide changes can be put in /etc/rpm/macros and user changes
to ~/.rpmmacros
Although I have root privileges I always build my RPMs as a regular
user (in case I screw something up I can't inflict the damage :-). I
have a directory /scratch/zvezdan/src with the following structure:
BUILD/ BUILDROOT/ DB/ macros rpm* rpmrc RPMS/ SOURCES/ SPECS/
SRPMS/
Here's the content of the files macros, rpm, and rpmrc
=== macros ===
#%_dbpath/scratch/zvezdan/src/DB
%_topdir/scratch/zvezdan/src
%_tmppath /scratch/zvezdan/src/BUILDROOT
=== rpmrc
macrofiles:
/usr/lib/rpm/macros:/usr/lib/rpm/%{_target}/macros:/etc/rpm/macros.specspo:/etc/rpm/macros.db1:/etc/rpm/macros:/etc/rpm/%{_target}/macros:~/.rpmmacros:/scratch/zvezdan/src/macros
=== rpm ==
#!/bin/sh
/bin/rpm --rcfile /usr/lib/rpm/rpmrc:/scratch/zvezdan/src/rpmrc "$@"
exit $?
==
Notice that I do not use _dbpath since I need to be able to read the
system's regular RPM database in case a package requires another program
to be built (e.g. AucTeX has RequireBuild: emacs), but the commented
line is still there to remind me that I can have even my private
RPM database.
Whenever I want to build RPMs I open another terminal window and do:
set path = (/scratch/zvezdan/src $path) <-- I use tcsh
umask 022
That way my rpm shell script is executed when I type rpm ...
It adds my rpmrc file, and it adds my macros file. Hence the topdir and
tmppath get changed to my directories and I can build as a regular user.
I picked up the idea from /usr/share/doc/rpm-4.0.2/multiplebuilds (might
be /usr/doc/rpm-xxx/ on another distro).
Hope this helps in your efforts to build RPMs on other distros.
All the best.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: RPM files
On Thu, Aug 16, 2001 at 07:33:55PM +0200, Ronny Buchmann wrote:
> the point is, that its not clear, which macros are supported in which
> rpm version so maybe zvezdan's rpms dont compile on (older) suse imo
> the spec files should be compatible with at least the last 6. versions
> from suse and redhat if one has a testsystem with say suse 6.4 (redhat
> has already rpm-4 in its updates for 6.2) i could make a test
> spec-file to see what macros are working
>
> --
> ronny
That's why I said on that web page:
"Hence, it should be usable on any RPM based Linux distribution (at
least those with rpm-4.x)."
As a sysadmin I regularly apply all Red Hat updates -- hence my 6.2
systems run rpm-4.x. I'm sorry I couldn't test it with the older version
of rpm.
I deal all the time with building RPMs, either for the special software
we install and want to be able to controll it through rpm, or through
fixing Red Hat RPMs (e.g. see Bugzilla #42352), or adapting RH RPMs to
our systems (we need xlock to autologout a student after 30 minutes so
nobody can abuse lab computers).
I always use macros and they work great. For example, I wanted to
install software recently to /usr/local instead of /usr. The only thing
I had to change in my SPEC file is three lines I added at the top:
%define _prefix /usr/local
%define _mandir %{_prefix}/man
%define _defaultdocdir %{_prefix}/doc
This is definitely _THE_ solution to our maintanance problems. With each
new version of LyX only the version number must be changed, and it
should build and install correctly on any rpm based distro.
Regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: RPM files
On Thu, Aug 16, 2001 at 07:23:07PM +0200, Michael Schmitt wrote: > It is very nice that you provide "better" rpms on your web page. But maybe > Kayvan and you should merge experience. I think the link on the lyx > page directly points to Kayvan's release. > > Michael Well, Kayvan said that he'll incorporate the changes in the RPMs, but I don't see that happening yet. Perhaps this discussion will prompt him to do it faster. :-) Regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: RPM files
On Thu, Aug 16, 2001 at 06:50:37PM +0200, Michael Schmitt wrote:
> On Thu, 16 Aug 2001, Ronny Buchmann wrote:
>
> > > does anybody know whether the rpm files for development snapshots provided
> > > by Kayvan can also be applied safely to a Suse system? AFAIK there are
> > > some differences between the various distributions and the rpms are
> > > primarily made for Redhat.
>
> > i think they work for suse, too, since its still rpm-3, isnt it?
> > but use also kayvans xforms rpm
>
> Indeed, the RPM version should be no problem. But how about, e.g., the
> right directory settings for installation? IIRC there was some discussion
> about this issue recently because Redhat has changed their directory
> structure from 6.4 to 7.1 (I hope I don't mix things here).
>
You're right Michael. Kayvan's RPM use very few RPM features. I have
made recently RPM's that use RPM macros only (instead of /usr/man
(RH6.2) or /usr/share/man (RH7.1) I use %{_mandir}, etc.). They compiled
fine on my old RH6.2 system and new RH7.1 system. They should compile
under any RPM based system and set up the correct standard paths for
that particular distro.
Check out: http://www.cs.wm.edu/~zvezdan/programs/RPMS/LyX/
If you're compiling the development version then install the source RPM,
change the version number in the SPEC file and rebuild the source and
binary.
You'll find the _correct_ xforms RPMs there too. The one on the LyX
site doesn't use RPM_BUILD_ROOT at all. It goes directly writing into
your /usr/X11R6/lib during the build. So, you have parts of it before
you have installed it. I fixed that and made 0.89 version RPM too, in
case somebody wants it.
Regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: Quote v Quotation
On Mon, Aug 06, 2001 at 04:11:26PM +0100, Angus Leeming wrote: > On Monday 06 August 2001 16:01, Zvezdan Petkovic wrote: > > On Mon, Aug 06, 2001 at 03:52:27PM +0100, Angus Leeming wrote: > > > In the drop down list of possible layouts. > > > > > > What's the difference? > > > > > > Angus > > > > One is used for a short single paragraph quotes. The other for a > > multiple paragraph quotations. > > Thanks. Which is which? > > A I used words _quote_ and _quotation_ above. :-) -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Quote v Quotation
On Mon, Aug 06, 2001 at 03:52:27PM +0100, Angus Leeming wrote: > In the drop down list of possible layouts. > > What's the difference? > > Angus One is used for a short single paragraph quotes. The other for a multiple paragraph quotations. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: TEXmacs
On Thu, Jul 12, 2001 at 10:42:17AM -0700, Mike Ressler wrote:
> As I recall, Zapf was involved in CM to some extent (at least giving
> advice)
This was Euler not CM. And he designed it. Knuth was giving advice on
use of Metafont.
Check out the book:
Hermann Zapf & his design philosophy : selected articles and
lectures on calligraphy and contemporary developments in type
design, with illustrations and bibliographical notes, and a
complete list of his typefaces
It talks about Euler _only_ and how Zapf first time designed using
program (with Knuths help) while designing Euler for AMS. There is no
mention of CM. Since Euler was designed much after CM, and Zapf himself
says that was first time he used Metafont (or any other font design
program), what's the only logical conclusion?
I just checked the Preface to "The METAFONTbook" and Knuth himself says
there:
\MF\ is a system for the design of alphabets suited to raster-based
devices that print or display text. The characters that you are
reading were all designed with \MF\!, in a completely precise way;
and they were developed rather hastily by the author of the system,
who is a rank amateur at such things. It seems clear that further
work with \MF\ has the potential of producing typefaces of real
^{beauty}. This manual has been written for people who would like to
help advance the art of mathematical type design.
Do you believe me now?
FWIW, The METAFONTbook is dedicated:
To Hermann Zapf
Whose strokes are the best
>, but the real "problem" is that CM is modeled after Monotype
> Modern #8. If you like that font, then you will like CM; if not, ...
>
I said in one of the earlier postings on this topic that I still like CM
better than Times. For some reason. But I'd rather not use any of them.
Give me Palatino, Garamond, Minion, etc. -- and, of course, Euler for
math.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: TEXmacs
On Thu, Jul 12, 2001 at 07:29:54PM +0200, Herbert Voss wrote:
>
> have a look at
>
> http://www.myfonts.com/Person295.html
>
> Herbert
>
As far as I understand English, that sentence says that Zapf and other
gave recommendations for improvement of Metafont _program_ _not_ CM.
Check out the book:
Hermann Zapf & his design philosophy : selected articles and
lectures on calligraphy and contemporary developments in type
design, with illustrations and bibliographical notes, and a
complete list of his typefaces
It talks about Euler _only_ and how Zapf first time designed using
program (with Knuths help) while designing Euler for AMS. There is no
mention of CM. Since Euler was designed much after CM, and Zapf himself
says that was first time he used Metafont (or any other font design
program), what's the only logical conclusion?
I just checked the Preface to "The METAFONTbook" and Knuth himself says
there:
\MF\ is a system for the design of alphabets suited to raster-based
devices that print or display text. The characters that you are
reading were all designed with \MF\!, in a completely precise way;
and they were developed rather hastily by the author of the system,
who is a rank amateur at such things. It seems clear that further
work with \MF\ has the potential of producing typefaces of real
^{beauty}. This manual has been written for people who would like to
help advance the art of mathematical type design.
Do you believe me now?
FWIW, The METAFONTbook is dedicated:
To Hermann Zapf
Whose strokes are the best
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: TEXmacs
On Thu, Jul 12, 2001 at 10:56:38AM +0200, Jean-Marc Lasgouttes wrote: > >>>>> "Zvezdan" == Zvezdan Petkovic <[EMAIL PROTECTED]> writes: > > Zvezdan> That's exactly what I meant. CM family is ugly. Notice what I > Zvezdan> say above about Euler. Metafont itself is not an issue. It's > Zvezdan> a design. CM was designed by one of the greatest computer > Zvezdan> scientist off all time, but only a would-be artist. Euler is > Zvezdan> designed by one of the greatest designers of all time -- > Zvezdan> Herman Zapf. That's what makes a difference. > > I thought the cm font had been designed with input from Zapf too... > > JMarc As far as my memory serves me, Knuth and Zapf worked together on Euler only for American Mathematical Society. Zapf never used a program to design (he used Linotype devices only before) and Knuth was an expert for such things. :-) One can hardly imagine two better choices for these two parts of the work. I read a book: Hermann Zapf & his design philosophy : selected articles and lectures on calligraphy and contemporary developments in type design, with illustrations and bibliographical notes, and a complete list of his typefaces and do not remember seeing CM mentioned there though. It had samples of all his designs at the back and Euler was there of course. I might check it out in the library again if you're _really_ interested. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: TEXmacs
On Thu, Jul 12, 2001 at 09:28:25AM +0100, Jules Bean wrote: > > > > I know that. When I say decent PDF I mean the PDF read from the screen. > > The printed page looks fine if you have a decent 600dpi printer. > > However, on a less good printer even the printed page won't look fine. > > The reason is that PDF format embeds them as Type3 fonts, which > > basically boils down to bitmaps. Hence, low resolution (as on screen or > > bad printer) amounts to a bad quality. > > Well, this isn't quite true. > > a) type 3 != bitmap. But the type 3 fonts generated by metafont are > bitmap, so that is correct here. Probably you already knew that. > I said it boils down to bitmap, implying "in this case". > b) bitmap @ (say) 600dpi, does *not* mean it looks ugly on screen. > For example, consider the output of dvips (at 600dpi) viewed in gv. > Looks absolutely fine. The fault lies entirely with acroread, which > is /very/ bad at displaying bitmaps (it's anti-aliasing/supersampling > algorithm is horrible for bitmaps). Try running gv on your pdfs, and > they will look much nicer. > I absolutely agree. > c) bitmap does not mean it looks bad printed, either. After all, > printers are bitmap devices. It just means you need a compatible > resolution (either the correct one, or an integer multiple of the > correct one). So you do need to know (guess) the output resolution at > the time the fonts are rasterised to bitmaps. > Yes. But my paper will be most probably viewed and printed by somebody who has a different printer. > > > > Fortunately, teTeX comes with all Metafont fonts in a Type1 version too. > > If one changes a single line in $TEXMFPATH/dvips/config/updmap from > > false to true, one can get Type1 used for CM, Euler, etc. by > > default. > > True. Although I've been told by some people that some of these type > 1s aren't really brilliant. They look OK to me. > It's true that they are not as good as Metafont counterparts. One can really see the distortion of some letters if you print the same page using Metafont and then Type1 fonts. > > > > I find it very important to have a readable PDF on the screen. I'm > > Of course. But I argue it's adobe that's the main problem here, not > the technology. > I agree. I still like better that some company imposes a good format on me (like PDF) than a bad format (like doc). :-) Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: TEXmacs
On Wed, Jul 11, 2001 at 07:57:12AM +0100, Jules Bean wrote: > > > I do not consider seriously Metafont fonts since they are ugly, > > starting with Computer Modern. Euler is a notable exception because > > it was designed by Herman Zapf, of course. But it's available as > > Type1 font in teTeX too. > > There's nothing fundamentally ugly about metafont fonts. It can be > argued that they're a more sophisticated font technology than Type 1, > despite predating it. Of course, if you think the CM family is ugly, > that's something else, and fair enough. You're not alone there > (although I don't agree with you). > That's exactly what I meant. CM family is ugly. Notice what I say above about Euler. Metafont itself is not an issue. It's a design. CM was designed by one of the greatest computer scientist off all time, but only a would-be artist. Euler is designed by one of the greatest designers of all time -- Herman Zapf. That's what makes a difference. Frankly, when I began using LaTeX long time ago, I thought that CM is a nice font. At least, it looked better to me than Times. However, the more I learnt about fonts the more I disliked it. I'm not a font designer, but I have read some books and talked to some people who know something about it. The general conclusion is that CM is not the greatest achievement in type art. Still, if I had to choose between CM and Times I'd use CM, for some reason. Since I have other choices I rather use something else depending on the type of writing. IEEE Transactions switched a few years ago to Palatino (Herman Zapf design again) from Times. I was quite happy about that. Palatino was also available in the default teTeX installation and I use it quite often. It looks great in combination with Euler for math. For less scientific kind of writing one can try Garamond, or one of the new beautiful fonts -- Minion. There were two excellent articles a few years ago in _Cahier GUTenberg_ (a journal of a French TeX user's group), by Thierry Bouche on the topic of math fonts and multi-master Minion fonts in TeX. http://www.gutenberg.eu.org/pub/GUTenberg/publications/publis.html Follow the links for Cahiers 25 and 26 and Thierry's articles. I remember seeing at least one of these translated in English and published in TUGboat too (http://www.tug.org). > > The importance of using Type1 cannot be overemphasised since it > > is the only way to produce a decent PDF. I hope you'll agree that > > the ability to produce a good looking PDF is the most important > > feature of any typesetting program nowadays. > > Point of information: Metafont fonts produce perfectly decent PDFs > (although, admittedly, PDFS best suited to particular resolutions, > since there is no way of embedding the 'outline' version of the > metafont into the PDF, it has to be a particular resolution > version). Acrobat reader simply fails to display them in a remotely > pleasant way: this may be related to the fact that Adobe make money > selling type 1 fonts. You'll note that they print fine. > I know that. When I say decent PDF I mean the PDF read from the screen. The printed page looks fine if you have a decent 600dpi printer. However, on a less good printer even the printed page won't look fine. The reason is that PDF format embeds them as Type3 fonts, which basically boils down to bitmaps. Hence, low resolution (as on screen or bad printer) amounts to a bad quality. Fortunately, teTeX comes with all Metafont fonts in a Type1 version too. If one changes a single line in $TEXMFPATH/dvips/config/updmap from false to true, one can get Type1 used for CM, Euler, etc. by default. I find it very important to have a readable PDF on the screen. I'm subscribed to both IEEE Computer Society and ACM digital libraries. I don't like printing so many articles (saves some trees I guess :-). I suppose people would begin reading my own articles off the screen at first. They might just as well dismiss them if they are unreadable, rather than decide to print them. Telling them that they look quite well when printed is simply not good enough. We are responsible for presentation of our work in any form. If it looks bad on screen, people might get the impression that we didn't make an effort to make it look better. Hence, having a good-looking PDF on screen is an imperative. Hence Type1 fonts. I hope I made my point more clearly now. :-) Finally, I would agree with you that Metafont is a fine technology, but Type1 has become de facto standard for printing industry. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: TEXmacs
On Tue, Jul 10, 2001 at 10:13:30AM +0200, Rainer Dorsch wrote: > > Hi, > > I am just wondering, if there are any suggestions to Question 5 of > > http://www.texmacs.org/Web/FAQ.html > I've made them to Mr. van der Hoeven directly as can be seen in a copy of that email here: * >From [EMAIL PROTECTED] Tue Jul 10 14:26:29 2001 Date: Tue, 10 Jul 2001 14:26:29 -0400 From: Zvezdan Petkovic <[EMAIL PROTECTED]> To: Joris van der Hoeven <[EMAIL PROTECTED]> Subject: TeXmacs advantages over LyX list Message-ID: <[EMAIL PROTECTED]> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Status: RO Content-Length: 3243 Lines: 65 Dear Mr. van der Hoeven, I have come across a Web page for TeXmacs and in the FAQ section you state the advantages of TeXmacs over LyX as follows below. I have put my comments in between. 1. TeXmacs is fully WYSIWYG. I hope you will accept that this is a matter of taste. What's an advantage for one person is a disadvantage for another. In my opinion this is merely a feature that some people might prefer. Still, for such people, this is an acceptable claim. 2. TeXmacs has a professional typesetting quality and nicer fonts. The meaning of this is not quite clear to me. LyX uses TeX for typesetting. Are you claiming that you are producing better and more professional printed output than TeX? If you are referring to rendering on the screen than I wish you have used that word instead. The word "typesetting" refers to a printed media (still). But even that claim is not so strong to me. Who has nicer fonts obviously depends on which fonts are installed on a system. XFree86 4.x has brought us an anti-aliased rendering engine of a very good quality. Is yours better? Can you show True Type fonts too? teTeX provides all nice fonts in a Type1 format, and hence available for rendering under X too. Thus, any Type1 and TrueType font (including teTeX ones) can be rendered in LyX (or any other X application) through the regular X mechanism. Qt supports anti-aliasing through X in an excellent way and it looks really great in native KDE applications. I hope GTK will do the same in the near future. LyX will soon support both of these toolkits (although there is a KLyX already). I do not consider seriously Metafont fonts since they are ugly, starting with Computer Modern. Euler is a notable exception because it was designed by Herman Zapf, of course. But it's available as Type1 font in teTeX too. The importance of using Type1 cannot be overemphasised since it is the only way to produce a decent PDF. I hope you'll agree that the ability to produce a good looking PDF is the most important feature of any typesetting program nowadays. To conclude: Can TeXmacs do typesetting better than TeX and rendering of Type1 and TrueType fonts better than Freetype X module? If not, this claim is invalid. 3. TeXmacs comes with the Guile/Scheme extension language. 4. You can use TeXmacs as an interface to computer algebra systems. These are real advantages for people who need them. Finally, what are the disadvantages of your program. Well, take a look at your own list of suggestions -- they'll be quite obvious (no floats, no two columns, ...). Besides, you are using a non-standard format with still imperfect conversion _to_ LaTeX. That is quite unacceptable to any scientist. Non-scientists will use something Word-like anyway. You do not support Docbook SGML DTD, also. Future is in XML, MathML, etc. I do respect your work and I believe that putting a sincere list of advantages and disadvantages would be really beneficial for it. Yours sincerely, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/ *** -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www.devel.lyx.org
On Sat, Jul 07, 2001 at 08:55:43AM +0200, Jürgen Spitzmüller wrote: > Have you noticed that the devel page is somehow broken since a few > days? (Menue not as float but somewhere in the mid of the side and very > messy if you switch to top navbar). > Tested with konq 2.1.2, Opera 5, Netscape 4.7 and Mozilla 0.91 > Yes, the developers pages are in transition. They still use tables instead of pinmenu.css. They will be done soon as far as I understood Allan when I asked the same question. :-) Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Red hat 7.0 compiler
On Mon, Jul 02, 2001 at 07:52:53AM +0100, Yann MORERE wrote: > > i tried to update, without success... > > sorry but no 7.1 release for my alpha :-(. > kgcc is your friend. Get it and I believe you'll be able to compile. It's a well known issue and LyX is not the only application you won't be able to compile with 7.0 gcc. It's not a real gcc, it's Red Hat `creation'. As far as I know, such applications compile with kgcc (and work!) Good luck. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Red hat 7.0 compiler
On Mon, Jul 02, 2001 at 12:08:25AM +0300, Gady Kozma wrote: > > Trying to compile lyx 1.1.6fix2 on Red Hat 7.0 with the supplied compiler > (it calls itself gcc 2.96 2731, but Red Hat probably patched the hell > out of it) creates an application the crashes with almost every save. It > crashes in operator<< called from somewhere in the mathed. > > Removing the -O2 from the mathed+main Makefiles seems to solve the problem. > > So, what's the recommended compiler for compiling lyx? Is this a known > problem? Is this a lyx or a gcc problem? > > Gady Red Hat 7.0 has a well known problem with a compiler. Did you get the update from the Red Hat update pages/site? Also, there is a _real_ gcc compiler for Red Hat 7.0 that is called kgcc. Try that one. It might be worth upgrading to 7.1 also, but I believe Red Hat hacked gcc in it too (there came an update recently). -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: Long list of bugs
On Thu, Jun 28, 2001 at 04:37:58PM +0200, Lars Gullik Bjønnes wrote: > > | Or maybe provide something else that can reasonably substitute the > | feature. > > KDE,GNOME,CDE,Windown etc... > What if I want to use my good ol' 90K small wmx? -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: http://www.devel.lyx.org/~rae/www-user/
On Thu, Jun 28, 2001 at 09:04:46AM +0200, A.Gulino wrote: > > > Finally, what's the look of the pages on www.w3c.org/Style/CSS in your > > browser? > > good, without problems Allan's previous message explains everything. Thanks! -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: http://www.devel.lyx.org/~rae/www-user/
On Wed, Jun 27, 2001 at 11:03:29AM +0200, A.Gulino wrote: > > after using Disable CSS I can read it. > But, how many peoples know, that "disable CSS" resolve the problem? > I forgot to add the most important thing in the previous e-mail. If your browser still renders the pages badly after you apply the default options to it we shall put Netscape 4.51 to the list of nocss browsers. That way people won't need to know about "disable CSS". For example, if you try lynx http://www.devel.lyx.org/~rae/www-user/ you'll be automatically offered two cookies: csidebar=0 and clong=0. That way you get the presentation that is the best. The purpose of testing is to show which browsers should go to the nocss list. I hope this clears out the things. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: http://www.devel.lyx.org/~rae/www-user/
On Wed, Jun 27, 2001 at 11:03:29AM +0200, A.Gulino wrote: > this page is terrible > http://www.devel.lyx.org/~rae/www-user/ > words = very_lightgray > backgroudn = white > result=not readable > I would say: This page looks terrible in my browser. Very important difference! > I'm using Netscape comunicator 4.51 > using: Preference->Fonts: Use my default fonts, overriding document-specified > fonts > Most of the people leave Use document fonts option on, because than they see the page as author intended. Besides most people update their browsers (or their sysadmins do that) whenever there is a security advisory. > after using Disable CSS I can read it. > But, how many peoples know, that "disable CSS" resolve the problem? > > ciao, antonio And I ask you: How many people uses Netscape 4.51? + How many of them has Use my default fonts option on? Switch off that option. Switch on CSS. Try it again. If it doesn't work update your Netscape to 4.77. It's free. It's recommended because all updates were _security_ advisories. Finally, what's the look of the pages on www.w3c.org/Style/CSS in your browser? -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
On Wed, Jun 27, 2001 at 01:58:23PM +0200, Jean-Marc Lasgouttes wrote: > >>>>> "Juergen" == Juergen Vigna <[EMAIL PROTECTED]> writes: > > Juergen> On 27-Jun-2001 Lars Gullik Bjønnes wrote: > >> | It made them difficult to read. > >> > >> Hmm, I never thought my eyes particulary superior... > >> > > Juergen> Me too, but maybe our glasses make us to be superior... > > I do not wear glasses, I liked the background too. Maybe developpers > meetings make us superior :) > > JMarc If most of you like it, it shouldn't be difficult to fix lyx.css. Very small change. And now, when we do not have tables any more, it shouldn't create the "black parts problem in Netscape" as it did before. That was the benefit of switching to s. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
On Tue, Jun 26, 2001 at 02:34:44PM -0400, Paul D. Smith wrote: > %% Martin Vermeer <[EMAIL PROTECTED]> writes: > > mv> You can directly access them in your browser... > > mv> http://www.lyx.org/pinmenu.css > mv> http://www.lyx.org/lyx.css > > That was the first thing I tried, and it doesn't work for me. > You must be using Netscape 4.x. > Netscape 4.77 on a Debian GNU/Linux box. > I knew it. :-) Change the browser. Netscape is lame in any respect. Especially CSS. Seriously, try Konqueror, Mozilla, whatever... because then you'll see the real effect of pinned menu. It stays in the top left corner while the text scrolls by. And you'll be able to see the source. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
On Tue, Jun 26, 2001 at 08:54:52PM +0300, Martin Vermeer wrote: > > > > I just finished setting up a framework for a local site, and while I > > used CSS for style-type stuff (fonts/sizes/etc.) I couldn't figure out > > any way to use them to create the navbar, etc. Instead, I'm using SSI > > for that. I'm no guru, for sure, I just know enough to be dangerous :). > > I'm just curious if that _local_ site is a personal/non-profit org or a company site. > > I'd be really interested in seeing the style sheets you are using to > > create the navbar stuff. That's easy. > > > > Thx... You're welcome. > > You can directly access them in your browser... > > http://www.lyx.org/pinmenu.css > http://www.lyx.org/lyx.css > > ...which you can see are used by viewing the source. > Martin, thanks for pointing this out to him. It's much easier than CVS in my opinion. FWIW, I wrote pinmenu.css. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
There's absolutely _NO_ JavaScript there. Just Netscape is so stupid that it can't show CSS unless JavaScript is on. It looks the same as if CSS is switched off. That's Netscape bug. I just checked in Konqueror and Mozilla with JavaScript off. Works quite fine. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
On Tue, Jun 26, 2001 at 02:34:34AM +0200, Paul Seelig wrote: > Just a minor nitpick: Too bad that for rendering it using stylesheets > within Netscape one has to enable JavaScript to make it work. :-( > > Cheers, P. *8^) I'm flabbergasted. You're quite right. What JavaScript has to do with CSS I have no idea. And I do not believe that people at Netscape will lose any sleep over that _BUG_! Despite all the effort to make it look right in Netscape, one switch can destroy it. Not to mention that despite all the effort it still looks inferior compared to more modern browsers (because the menu scrolls with the page, while in other browser it stays pinned down at the top left corner). Ah, well... -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
On Mon, Jun 25, 2001 at 07:25:33PM +0200, Sasa Janiska wrote: > > Here I see just a big stripe over main menu (running SuSE 7.1 & > Konqueror 1.9.8). > > But I appreciate your effort to follow the W3C's standards. > Thanks Sasa. As I explained in the previous message, I hope you understand that the best solution is an upgrade of KDE. The latest Konqueror (and all other browsers from the list) doesn't have any problems with LyX pages. I believe this is not the only page you have trouble seeing in that version of Konqueror (as far as I remember it). Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
On Mon, Jun 25, 2001 at 09:56:45AM -0700, [EMAIL PROTECTED] wrote:
> Look pretty good, but why is everything double spaced?
>
Because currently paragraph spacing is defined as:
p {
line-height: 1.4em;
}
Changing that to smaller value or just commenting out that rule in a CSS
file would return spacing to default. Now you see why CSS is so Good
Thing (TM). If most of the people wants the regular spacing it's an easy
fix.
No, I didn't write that rule. :-) I wouldn't mind the regular spacing.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: New LyX website released
We have made all possible effort to make this thing work on all important browsers. I personally test it first in Konqueror, but, yes, I do have Konqueror 2.1.1 on KDE 2.1.2. I believe you are aware that Konqueror below 2.x version was a rather buggy browser, especially in regard to CSS. Since you are using a free software I also believe it won't be a trouble for you to update to a new KDE/Konqueror. Believe me you'll get many benefits: anti-aliased fonts (if you also have XFree86 4.x), excellent support for printing non-latin1 encodings, very good standard support in Konqueror 2.1.1. It would be really difficult (if possible at all) to make this work in a discontinued version of Konqueror, and I hope you'll agree that it's not that important considering the number of users of that version. I also hope you won't take this as a lame excuse. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: LDN validations
On Mon, Jun 25, 2001 at 10:26:43PM +1000, Allan Rae wrote: > > Netscape4 and IE5.5 centre the table. > > Galeon0.10.2 (mozilla-0.8) and Opera5 both left align the table. > I am looking in the same issue for other reasons. Anything that's not a simple paragraph (might contain images) destroys the "centredness" of . There are some indications in CSS specification that table body element's presentation should be defined in CSS. I'm not sure whether Netscape will like it. I'll play with this these days. Definitely not a pressing issue. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
On Mon, Jun 25, 2001 at 07:17:20PM +1000, Allan Rae wrote:
>
> I think I've fixed all the points you made. I'm not sure I like the look
> of quotes with authors name on the right. It looks lost or at least is
> easily missed. I'll put up a page for comparison but for the moment I
> have the authors name inside the blockquote so that it is left aligned
> with the authors text. (this is the typographical standard for letters in
> Australia BTW -- so maybe that is why I prefer the left to the right)
>
I'm not picky about this. The only important thing is that author's name
is enclosed in so that font looks right in Netscape 4.x.
And I agree that these look really lost on the right side. What about:
p.quote-author {
text-align: right;
margin-right: 30%; /* or whatever looks OK */
}
and then &8212; Anonymous (&8212; is an em
dash in case you want it there). Just an idea to ponder. As I said it's
not important.
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: www page background
One thing we have to be particularly careful about is Internet Explorer's parsing of XML statements at the top. If I had a say, I would put the guy who wrote IE parser to the pillory so everybody can throw rotten eggs at him. :-) This prologue works in IE: http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";> http://www.w3.org/1999/xhtml"; lang="en" xml:lang="en"> as it should one would say. Not so fast Johnny. If you make _one more space_ between "DOCTYPE html" and "PUBLIC" the page looks like a huge single paragraph, underlined, unpredictable colour. If you add one more stylesheet (as we needed to do) below the first one. the same ugly result. Then I tried moving both stylesheet lines below DOCTYPE instruction. No joy. I tried everything on the same line, to no avail. Finally, I moved them below tag. http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";> http://www.w3.org/1999/xhtml"; lang="en" xml:lang="en"> It works! It even validates. But as far as I know should come _before_ the first tag. The other solution (applied in the first example) is cat pinmenu.css >>lyx.css and use of a single stylesheet. It at least respects the rules. Allan will probably decide which solution we choose. In my opinion, both are just workarounds. The first one respects the before-first-tag rule, but is inflexible since we always have to keep a single stylesheet. The second is a real workaround, and it's not clear to me how this can validate (and it does). And yes, _NO OTHER BROWSER_, including the lame Netscape 4.x, had any problems with parsing: http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";> http://www.w3.org/1999/xhtml"; lang="en" xml:lang="en"> Microsoft, should I say more... -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
There was too much redundant code in the css I sent in the previous
message. And topmenu is a much better name for our new div than
pinmenutop.
The new css is attached. For markup check:
http://www.cs.wm.edu/~zvezdan/LyX/www-user/
http://www.cs.wm.edu/~zvezdan/LyX/www-user/indexs.html
http://www.cs.wm.edu/~zvezdan/LyX/www-user/indext.html
I also tested in Konqueror. No joy. So
is the only solution.
I think we are ready for a nice web site. Implement the changes 1-7 I
indicated in the previous email. Change the markup for the topbar
version. Test in Opera (I already did Konqueror, Mozilla, Netscape, Lynx
and IE). Commit the changes.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
/*
* Pinned menu CSS
*
* Borrowed from the idea of Bert Bos <[EMAIL PROTECTED]> published at:
* http://www.w3.org/Style/Examples/007/menus.html
* and adjusted to work with Netscape 4.x, and better browsers as well.
*
* Usage:
*
* Menu option
*
*
*
* Page text follows
*
*
* (C) 2001 Zvezdan Petkovic <[EMAIL PROTECTED]>
*/
div.pinmenuwide {
/*
font-size: 80%;
font-weight: bold;
*/
text-align: left;
position: fixed;
top: 1em;
width: 12em;
margin-left: 1em;
margin-right: auto;
/* everything below is needed because of Netscape 4.x only */
float: left;
color: rgb(238,238,238);
background: rgb(53,99,144);
padding: 0.2em 0.2em 1.2em 0.2em;
}
div.pinmenushort {
/*
font-size: 80%;
font-weight: bold;
*/
text-align: left;
position: fixed;
top: 1em;
width: 7em;
margin-left: 1em;
margin-right: auto;
/* everything below is needed because of Netscape 4.x only */
float: left;
color: rgb(238,238,238);
background: rgb(53,99,144);
padding: 0.2em 0.2em 1.2em 0.2em;
}
div.topmenu {
text-align: center;
width: 99%;
border: none;
/* everything below is needed because of Netscape 4.x only */
color: rgb(238,238,238);
background: rgb(53,99,144);
padding: 0.2em 0.2em 1.2em 0.2em;
}
div.pinmenuwide p, div.pinmenushort p, div.topmenu p {
margin: 0;
color: rgb(238,238,238);
background: rgb(53,99,144);
font-family: Verdana, Arial, Helvetica, sans-serif;
}
div.pinmenuwide a:hover, div.pinmenushort a:hover, div.topmenu a:hover {
color: rgb(235,188,94);
background: rgb(36,62,108);
}
div.page {
margin-left: 15em; /* change according to menu */
}
div.page-wide {
margin-left: 9em; /* change according to menu */
}
a:hover {
color: rgb(235,188,94);
background: rgb(36,62,108);
}
.right {
text-align: right;
}
.rfloat {
float: right;
}
Re: www page background
I forgot to attach pinmenu.css to the previous message but that's fine
because I have a solution for the top navbar now. Looks good in
Konqueror this time (and Mozilla, Netscape, IE, and Lynx).
http://www.cs.wm.edu/~zvezdan/LyX/www-user/indext.html
Just don't ask me why "width: 99%" and not 100% -- those browser
differences :-(
You might want to make this one "position: fixed" too, but be warned
that you might need a after because Mozilla hides the
title below the menu.
Cheers,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
/*
* Pinned menu CSS
*
* Borrowed from the idea of Bert Bos <[EMAIL PROTECTED]> published at:
* http://www.w3.org/Style/Examples/007/menus.html
* and adjusted to work with Netscape 4.x, and better browsers as well.
*
* Usage:
*
* Menu option
*
*
*
* Page text follows
*
*
* (C) 2001 Zvezdan Petkovic <[EMAIL PROTECTED]>
*/
div.pinmenuwide {
/*
font-size: 80%;
font-weight: bold;
*/
text-align: left;
position: fixed;
top: 1em;
width: 12em;
margin-left: 1em;
margin-right: auto;
/* everything below is needed because of Netscape 4.x only */
float: left;
color: rgb(238,238,238);
background: rgb(53,99,144);
padding: 0.2em 0.2em 1.2em 0.2em;
}
div.pinmenushort {
/*
font-size: 80%;
font-weight: bold;
*/
text-align: left;
position: fixed;
top: 1em;
width: 7em;
margin-left: 1em;
margin-right: auto;
/* everything below is needed because of Netscape 4.x only */
float: left;
color: rgb(238,238,238);
background: rgb(53,99,144);
padding: 0.2em 0.2em 1.2em 0.2em;
}
div.pinmenutop {
text-align: center;
width: 99%;
border: none;
/* everything below is needed because of Netscape 4.x only */
color: rgb(238,238,238);
background: rgb(53,99,144);
padding: 0.2em 0.2em 1.2em 0.2em;
}
div.pinmenuwide p {
margin: 0;
color: rgb(238,238,238);
background: rgb(53,99,144);
font-family: Verdana, Arial, Helvetica, sans-serif;
}
div.pinmenushort p {
margin: 0;
color: rgb(238,238,238);
background: rgb(53,99,144);
font-family: Verdana, Arial, Helvetica, sans-serif;
}
div.pinmenutop p {
margin: 0;
color: rgb(238,238,238);
background: rgb(53,99,144);
font-family: Verdana, Arial, Helvetica, sans-serif;
}
div.page {
margin-left: 15em; /* change according to menu */
}
div.page-wide {
margin-left: 9em; /* change according to menu */
}
div.pinmenuwide a:hover {
color: rgb(235,188,94);
background: rgb(36,62,108);
}
div.pinmenushort a:hover {
color: rgb(235,188,94);
background: rgb(36,62,108);
}
a:hover {
color: rgb(235,188,94);
background: rgb(36,62,108);
}
.right {
text-align: right;
}
.rfloat {
float: right;
}
Re: www page background
On Sun, Jun 24, 2001 at 02:54:21PM -0400, Zvezdan Petkovic wrote: > On Sun, Jun 24, 2001 at 08:26:34PM +0200, Herbert Voss wrote: > > > > in Netscape 4.77 without css the menu is above the text. > > > > And it should be. In Lynx it looks exactly like that. In such cases top > navigation should be switched on. > I should have said: Allan's php will switch top navigation bar on. I'm just giving an example page so Allan can pick up CSS. Regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
On Sun, Jun 24, 2001 at 08:26:34PM +0200, Herbert Voss wrote: > > in Netscape 4.77 without css the menu is above the text. > And it should be. In Lynx it looks exactly like that. In such cases top navigation should be switched on. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
Allan,
here's the list of issues I've resolved in order of importance:
1. pinmenu.css fixed to work in Netscape, Mozilla, Konqueror, IE and
Lynx (wide and short versions). It's attached. See:
http://www.cs.wm.edu/~zvezdan/LyX/www-user/
http://www.cs.wm.edu/~zvezdan/LyX/www-user/indexs.html
2. Whatever used class="right" should use class="rfloat" now. Reason: I
needed name right for slightly different purpose.
3. Get rid of img.right in lyx.css, because it uses "align: right"
instead of "text-align". Anyway, it should float to be correctly
shown in Konqueror. Use or directly
. I tested the first option only.
4. Now why I needed .right. In news/20010523.php3 your quotes have the
author listed to the left. I really do not remember if this was a
decision or an accident. If you need more customary right alignment
use Jurgen (where .right { text-align: right }).
Even if you want them to the left use Jurgen because the font
gets screwed in Netscape otherwise. Take a look at:
http:/www.cs.wm.edu/~zvezdan/LyX/news/20010523.html
Only one quote is formatted with . See the diff
compared to others, and mark-up in the source.
5. Use local valid-xhtml1.gif because it is transparent as opposed to
w3c version which has a white background. You can pick GIMP fixed
version from my site.
6. Get rid of spaces around images. See the diff between index.html and
indexs.html when you hover over images and see the diff in the
source.
7. Add for pinmenu.css. In fact reformat a little
your meta lines so they are more readable:
http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";>
http://www.w3.org/1999/xhtml"; lang="en" xml:lang="en">
What remains to be done:
Did you ever see top navbar in Konqueror. If not you are spared of big
surprise. It takes a whole page in height and two pages in width. I do
not know why, but I'm working now on fixing it in a manner.
I'll be back as soon as I do that.
Best regards,
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: www page background
On Sun, Jun 24, 2001 at 06:46:43PM +1000, Allan Rae wrote: > I fixed the menu problem. But somehow Netscape-4.77 has decided that it > won't draw the body correctly anymore it now leaves a huge gap between the > sidebar and the body. *sigh* > There's a problem with Netscape 4.x that under some conditions in measures left margin starting from the right edge of the pinmenu . Thus it created 15em gap. I struggled a lot to fix it and as far as I can remember both divs have to have margin-* setup correctly (no use of left or right only). Now, when you were changing div.* p to .pin, div.* that causeed the problems. In there, the first line is margin: 0 which screws up the margins already set up for divs. I think you can take div.* safely out of that line. The reason I set up div.* p is that the colour of centered things remains the same. When one says it recognises that it's inside div and takes the colours and top/bottom margins of that p. Now, with the change, when you say it takes the colours of and top/bottom margins of a normal + centering. > > I had to make the definition in pinmenu into a to get > the menubar rendered approximately right. It doesn't make sense because > everything was otherwise identical to Zvezdan's test page! > Did you have to do this because of Opera? If not, then I think it should go back to div.* p (just make two copies with wide and narrow div.) I'm looking into this, right now. > Netscape 4.77 > = > > sidebar has bottom line in wrong background colour > body of text has it's left margin about twice the width of the navbar. > problem explained above > Before the change the body text was where it should be! > Now the sidebar renders okay but separation from main body is stuffed. > margin: 0 above > ### > > I can't run Konquerer at the moment 'cos I have to rebuild the KDE libs > with the new (gcc 2.95.3) libstdc++ so can someone else look at that? > Michael? > Short menu is too narrow in Konqueror. "Top navbar" goes out the right edge. I suppose we should go with 7em instead of 6. You still didn't apply the fix to . In Konqueror there is a blue line over LyX logo created by an h3 Recent LyX news. Try . It fixes the problem. Makes sense since is an inline element and h3 doesn't clear around. is a block element and h3 has to clear around it. > ### > > So if I revert the change then the menubar is stretched > out but everything otherwise looks right. Any ideas? > I'm looking into this. > ### > > I'm pretty sure I have nearly all LDNs rendering okay and completely valid > XHTML now as well. If it weren't for that stupid Netscape rendering bug > or at least the weirdness with the sidebar I'd commit these changes to cvs > so we could enjoy some really sexy web pages. > > Allan. (ARRae) > -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
On Sun, Jun 24, 2001 at 04:31:23PM +1000, Allan Rae wrote:
>
> Aren't you supposed to be able to do for
> multiple inheritance?
>
Something similar I suppose, but it never
worked in Netscape 4.x, so I never use it. One usually can achieve the
same with:
--
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/
Re: www page background
On Sun, Jun 24, 2001 at 03:11:04PM +1200, Michael Koziarski wrote: > In opera it looks ... wrong :) There's a screenshot up at > www.koziarski.org/LyX. Looks good in mozilla. Try it again please. I do not have Opera. I fixed width of pinmenu to 12em. I had the same problem with Mozilla and right aligned menu. Fixing width fixed it. I tested again in Mozilla, Netscape, and Konqueror. Let's wait Allan's reply to decide about moving the menu to the left corner. I'm for it. Cheers, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
Allan, I think I know what's the cause of most of the problems we had with Netscape's understanding of CSS. is just the nastiest manifestation of it. Namely, we use for the layout purposes. That is something we had to do before, but with CSS there are other ways. W3C guys openly encourage change in that direction. I have made a page that uses instead of for layout. Take a look at: http://www.cs.wm.edu/~zvezdan/LyX/www-user/ In Netscape it'll look the same as yours (except the navbar background), but in more modern browsers you will notice that the menu is pinned up. The text scrolls, but the menu stays in place "sticked" to the window, and easily accessible. I like the feature, and it it took me some time to make it to at least appear in Netscape. The css is included as a second one in XHTML header (cascaded) and it's called pinmenu.css. Take a look at XHTML. It's much more readable than the version. I put on the page a test of hfill in a title using class="right", and a test of . Doesn't crash Netscape now! One fix also. In Konqueror LyX mascot was not right aligned. I put the statement in a . The only change I would do is "top: 1em" in the div.pinmenu class. It's currently 10em to be in approximately same place as before. 1em would pin it in the upper left corner. It can also be changed to go to the right corner if you want. Tell me what you think. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: www page background
On Sat, Jun 23, 2001 at 12:45:11PM +1000, Allan Rae wrote: > What would like to see? > > I'm happy to leave the background but set a solid fill colour for the > tables -- something I did for my own LyX page but then I'd like to pinch > some of those other ideas from windowmaker and bbkeys websites I pointed > out before. > > I'd prefer a solid colour under the text so we can actually read it -- and > people with poor eyesight don't have to struggle so much. > As I pointed out in another e-mail this would prevent some of the troubles we currently have with Netscape 4 understanding of CSS. I struggled for a while with this, probably a year ago, and decided to go with a solid colour. Some of my pages have light background (Navajo White [rather sandy colour]) and some have dark background (Indigo blue) and white text. Much less CSS troubles with Netscape, believe me. -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: LDN validations
On Sat, Jun 23, 2001 at 12:12:50PM +1000, Allan Rae wrote: > > Did you try this yourself? > I didn't. I'll try to find the solution tomorrow morning. I had a tough day at work and its time to get some sleep. > This was the first thing I tried last night and guess what happens when > you do this? > > Netscape renders the background of the main table (the main body of text) > black! But at least the list stats table comes out with white text. > My experience is that this is caused by a background image. Because of it we have to put "background: transparent" whenever we change colour of something. Well, Netscape sometimes renders that transparent as black. The same thing happens with that rfloat I sent you in the other email. I had to put explicitly color/background-color pair to avoid black background in the right floated part. I assume something similar should work for the tables, but than -- bye, bye background image. You get plain cyan/green (whatever) colour. > Opera does something even stranger. Instead of centering the table it > suddenly decides it should left align it! > Konqueror does the same. CSS1 specification gives an example with "text-align: centre" used on a element and says that everything inside that block should be centred. As long as you have only paragraphs and images inside that it works. The table however screws up Konqueror. In my opinion it's their bug. Even Netscape 4.7 is able to render that correctly (and Netscape 6/Mozilla, IE 5.x of course). > Maybe I'll take a closer look at the strict validation that you suggest. > Worth doing for the future. The mark-up becomes very clean and maintainable. > > This is a nicety not essential although I'm sure Matej Cepl would > appreciate having his name spelled correctly rather than with no accents > like here. He appears 8 times in four different LDNs. > > Actually, Zvezdan what is the right XHTML compliant string to use for your > name? I went to your website and your accented name only appears in > images as far as I could work out.. > Unfortunately, I do not know if there is ć in XHTML standard. I almost got used to having my name "misspelled" here in USA. The only trouble is that Americans usually mispronounce it with "k" sound at the end instead of "ch" sound. Well, ... I'll post tomorrow if I find the solution for table background problem. Best regards, -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
Re: LDN validations
On Fri, Jun 22, 2001 at 10:54:11PM +1000, Allan Rae wrote: > Why does the mail-list stats table in the tips section of > news/20010117.php3 render with black text in Netscape? > > How do I fix it so the text is white? > Try to validate the things as Strict instead of transitional XHTML, you'll see. :-) Here's a tip. After Sponsorship for 5th Developers Meeting you have: (Good news -- Jun 8, 2001) Transitional XHTML lets you by with this error. It implies after . Strict doesn't. Put (Good news -- Jun 8, 2001) and you'll get the colour. Simple rule: Don't leave any text outside of some block element (usually ) -- Zvezdan Petkovic <[EMAIL PROTECTED]> http://www.cs.wm.edu/~zvezdan/
