pdflatex regressions?

View: New views
9 Messages — Rating Filter:   Alert me  

pdflatex regressions?

by Hal Ashburner-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hey people,
I had to do a sever upgrade from sarge->etch->lenny on a machine that
uses pdflatex to generate some complicated reports as part of an
application it's running.
There's a .tex template that gets filled repeatedly for different
reports and when it works you get very pretty pdf documents.
pdflatex appears to have developped some issues over 5 years with
respect to errors. Now latex errors have always been hideously unhelpful
but this is a new level of unhelpful. It's reporting errors every time
\bfseries is used. running latex on the same .tex produces no such
error. It also produces a wonderful error that reads:
Dimensions can be in units of em, ex, in pt, pc cm, mm, dd, cc, nd, nc,
bp or sp; but yours is a new one!
It does this for pt and mm for reasons I don't understand.
None of these errors matter at all except that when it gets to 100 it
executes the
if (errors > 100) { // yay for magic numbers in the middle of source
 stuff_you_I_hate_you();
}

The only solution I can see that doesn't require me to learn tex and
re-write all of latex, which would be a task worthy of Sisyphus,  is to  
get the source from debian, adjust the magic number to a bazillion and
install it locally.

Any better ideas?
Anyone seen all kinds of random breakages on upgrade using tabularx and
colortbl latex packages?

Cheers,
Hal
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Parent Message unknown Re: pdflatex regressions?

by Hal Ashburner-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

The app calls pdflatex. It does this because you can include .png
graphics by path in your .tex.
Where I said latex completes without error, I was lying and didn't know it.
It also fails on 100 errors that are the same ad pdflatex, it just
writes out almost all of the output (minus the charts and images) to the
.dvi file which other than images & truncations looks like it should.


Andrew Janke wrote:

> I am sure I am missing something but why not just the more "normal" route:
>
>    foo.tex -> foo.dvi  (latex)
>
>    dvipdf foo.dvi foo.pdf
>
> and dodge pdflatex altogether?
>
>
> a
>
> On Tue, Oct 13, 2009 at 17:46, Hal Ashburner <hal.ashburner@...> wrote:
>  
>> Hey people,
>> I had to do a sever upgrade from sarge->etch->lenny on a machine that uses
>> pdflatex to generate some complicated reports as part of an application it's
>> running.
>> There's a .tex template that gets filled repeatedly for different reports
>> and when it works you get very pretty pdf documents.
>> pdflatex appears to have developped some issues over 5 years with respect to
>> errors. Now latex errors have always been hideously unhelpful but this is a
>> new level of unhelpful. It's reporting errors every time \bfseries is used.
>> running latex on the same .tex produces no such error. It also produces a
>> wonderful error that reads:
>> Dimensions can be in units of em, ex, in pt, pc cm, mm, dd, cc, nd, nc, bp
>> or sp; but yours is a new one!
>> It does this for pt and mm for reasons I don't understand.
>> None of these errors matter at all except that when it gets to 100 it
>> executes the
>> if (errors > 100) { // yay for magic numbers in the middle of source
>> stuff_you_I_hate_you();
>> }
>>
>> The only solution I can see that doesn't require me to learn tex and
>> re-write all of latex, which would be a task worthy of Sisyphus,  is to  get
>> the source from debian, adjust the magic number to a bazillion and install
>> it locally.
>>
>> Any better ideas?
>> Anyone seen all kinds of random breakages on upgrade using tabularx and
>> colortbl latex packages?
>>
>> Cheers,
>> Hal
>> --
>> linux mailing list
>> linux@...
>> https://lists.samba.org/mailman/listinfo/linux
>>
>>    

--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions?

by Daniel Pittman :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hal Ashburner <hal.ashburner@...> writes:

> I had to do a sever upgrade from sarge->etch->lenny on a machine that uses
> pdflatex to generate some complicated reports as part of an application it's
> running.

You might find a TUG more useful for this, since this is likely to be pretty
platform independent...

[...]

> It's reporting errors every time \bfseries is used. running latex on the
> same .tex produces no such error. It also produces a wonderful error that
> reads: Dimensions can be in units of em, ex, in pt, pc cm, mm, dd, cc, nd,
> nc, bp or sp; but yours is a new one!  It does this for pt and mm for
> reasons I don't understand.

Well, presumably it is seeing the measure incorrectly for some reason.  Sadly,
I can't speculate as to why.

> None of these errors matter at all except that when it gets to 100 it
> executes the if (errors > 100) { // yay for magic numbers in the middle of
> source stuff_you_I_hate_you();
> }

IIRC, non-stop mode avoids the abort and just runs through?  Certainly, one of
the run-modes of TeX does that...

[...]

> Any better ideas?  Anyone seen all kinds of random breakages on upgrade
> using tabularx and colortbl latex packages?

Bite the bullet and port it to some typesetting or template system you do
understand?  I confess that, unhelpful as it is to say, I used ConTeXt because
of this sort of problem with LaTeX.

        Daniel
--
✣ Daniel Pittman            ✉ daniel@...            ☎ +61 401 155 707
               ♽ made with 100 percent post-consumer electrons
   Looking for work?  Love Perl?  In Melbourne, Australia?  We are hiring.
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions?

by Andrew Janke :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

> Bite the bullet and port it to some typesetting or template system you do
> understand?  I confess that, unhelpful as it is to say, I used ConTeXt because
> of this sort of problem with LaTeX.

And on this note, I'd be interested to know what others use (beyond
ConTeXt!) for this.  Myself I tend to use inkscape and sed for complex
graphics heavy stuff.

Just make your template doc, sed the text into placeholders then just go:

   $ inkscape /tmp/fred.svg --output-pdf=out.pdf

I use this all the time in web form and other automation things.  Like
printing 3000 individually numbered raffle tickets with stubs 4-up on
A4 for the kids school last night.. :)


a
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions?

by Hal Ashburner-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

On 13/10/09 18:27, Andrew Janke wrote:

>> Bite the bullet and port it to some typesetting or template system you do
>> understand?  I confess that, unhelpful as it is to say, I used ConTeXt because
>> of this sort of problem with LaTeX.
>>      
> And on this note, I'd be interested to know what others use (beyond
> ConTeXt!) for this.  Myself I tend to use inkscape and sed for complex
> graphics heavy stuff.
>
> Just make your template doc, sed the text into placeholders then just go:
>
>     $ inkscape /tmp/fred.svg --output-pdf=out.pdf
>
> I use this all the time in web form and other automation things.  Like
> printing 3000 individually numbered raffle tickets with stubs 4-up on
> A4 for the kids school last night.. :)
>    
As far as I know this is a really limited solution. It's fine but for
the simple cases where this will work you could happily roll your own
using just about anything.

<rant>
Where it falls down is where you wish to produce several thousand
documents with a mix of both constant and document specific text of
arbitrary and varying lengths, both constant and document specific
images and charts, tables that will all be arbitrary and individual
lengths where each table will spill over more than one landscape A4
page. Now all of this has to look exactly like the graphic designer drew
with their software which does not produce anything useful. And the app
has to run on linux.

Context might be better than latex, could hardly be worse, I guess.
As for re-writing the template in $something_else, this is simply not
practical. It's not a labour of love and the client sure won't pay for
it. If we're on impractical solutions I suggest expunging latex in its
entirety. Ah I can dream, can't I...

Back in reality I'll fight my way through the idiocy of latex spending
too long on it and learning a bunch of stuff I really don't want to
know. A server upgrade probably shouldn't break what latex accepts and
what it doesn't. But while I'm here I will mention to anyone who can be
bothered listening to rants that the use of latex for anything other
than the generation of a single A4 document in portrait mode with two
columns and a template provided by the journal in which you are
publishing given you have a bunch of latex experts around who can fix
the inevitable problems for you is considered harmful.
</rant>

tl;dr latex, just don't. Automated document & report generation that
looks pretty is an unsolved problem on unix.
And if I'm wrong I'd love to know.

Cheers,
Hal
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions?

by Daniel Pittman :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hal Ashburner <hal.ashburner@...> writes:
> On 13/10/09 18:27, Andrew Janke wrote:
>
>>> Bite the bullet and port it to some typesetting or template system you do
>>> understand?  I confess that, unhelpful as it is to say, I used ConTeXt because
>>> of this sort of problem with LaTeX.
>>>
>> And on this note, I'd be interested to know what others use (beyond
>> ConTeXt!) for this.  Myself I tend to use inkscape and sed for complex
>> graphics heavy stuff.

The idea had never occurred to me to do that; I *must* keep it in mind for the
future, if I have graphics-focused rather than text-focused layout to do...

[...]

> As far as I know this is a really limited solution. It's fine but for the
> simple cases where this will work you could happily roll your own using just
> about anything.

*nod*

> <rant>

[...]

> Context might be better than latex, could hardly be worse, I guess.  As for
> re-writing the template in $something_else, this is simply not practical.

Yah.  It sucks, really, and I expected that when I made the comment.  On a
couple of occasions I actually...

> It's not a labour of love and the client sure won't pay for it.

...ended up selling the client on the fact that it was /so/ much easier for me
to deal with the ConTeXt side of things[1], and because it could deliver so
much more[2] for their needs.

[...]

> Back in reality I'll fight my way through the idiocy of latex spending too
> long on it and learning a bunch of stuff I really don't want to know. A
> server upgrade probably shouldn't break what latex accepts and what it
> doesn't.

Actually, it implies that you now have a newer version of LaTeX, and the
underlying TeX engine, and all the add-on packages that extend LaTeX to do
things that are not just your A4 two-column science paper.

Like any upgrade that is going to end up with incompatible changes *somewhere*
in the life of a multi-decade old product.  You may find the LaTeX change
documentation[3] gives some guidance to fixing the outdated code...

[...]

> tl;dr latex, just don't. Automated document & report generation that looks
> pretty is an unsolved problem on unix.  And if I'm wrong I'd love to know.

You are wrong. ;)

More seriously, ConTeXt is non-trivial but very effective.  It can do vastly
more than LaTeX does, and usually has a reasonable low cost.  Not so much fun
to typeset for a science journal, but otherwise...

Also, while non-free, PrinceXML is an *excellent* tool and I can heartily
recommend it.[4]  It uses HTML (or XML) and CSS to do layout.

Finally, wkhtmltopdf shows great promise for doing the same job, using the
WebKit engine to drive the HTML and CSS side of things.  It isn't quite there
yet, though.

        Daniel

Footnotes:
[1]  I strongly prefer it to LaTeX in, er, context because it comes from a
     layout perspective rather than an article perspective; while LaTeX has
     grown the capability to do cool things, ConTeXt started there — and is
     now driving some of the more advanced TeX capabilities like LuaTeX.

[2]  The good integration of JavaScript and PDF fields, at the time, along
     with the improved PDF output stuff, was great.  LaTeX has mostly caught
     up now, I think.

[3]  ...or the documentation for whatever add-on provides the specific command
     you are having trouble with.

[4]  They have a free-for-private-use version that sticks a logo on the first
     page, which is not too obnoxious, if you want to trial it first.  License
     fees are not too unreasonable...

--
✣ Daniel Pittman            ✉ daniel@...            ☎ +61 401 155 707
               ♽ made with 100 percent post-consumer electrons
   Looking for work?  Love Perl?  In Melbourne, Australia?  We are hiring.
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions?

by Andrew Janke :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

On Tue, Oct 13, 2009 at 21:47, Daniel Pittman <daniel@...> wrote:
>>> And on this note, I'd be interested to know what others use (beyond
>>> ConTeXt!) for this.  Myself I tend to use inkscape and sed for complex
>>> graphics heavy stuff.
>
> The idea had never occurred to me to do that; I *must* keep it in mind for the
> future, if I have graphics-focused rather than text-focused layout to do...

For longer (text) documents I tend to play the same games with
scribus. It exports SVG also and page layout of the template is a bit
easier.

Still I have been surprised at what I can achieve with inkscape for
business flyers with custom images and the likes from a web front end
with such a small amount of fuss, sponge[1] comes in most handy here.


a

[1] apt-get install moreutils
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions?

by David Tulloh-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Andrew Janke wrote:
>> Bite the bullet and port it to some typesetting or template system you do
>> understand?  I confess that, unhelpful as it is to say, I used ConTeXt because
>> of this sort of problem with LaTeX.
>>    
>
> And on this note, I'd be interested to know what others use (beyond
> ConTeXt!) for this.  Myself I tend to use inkscape and sed for complex
> graphics heavy stuff.
>  
I've generated Open Office, often to convert into MS Word.  I have used
straight template substitution driven by Make and full generation using
the Perl CPAN libraries.  This is very useful for test reports from
automated testing (with a small amount of hand editing at the end) and
large documents (hundreds of pages) generated from test databases with
multiple graphs etc.

For precise layout work I have also generated pure postscript in the
past, I've written bash scripts to generate CD covers and CD wallet
faces.  For simple drawing of lines and text postscript is really very
simple, you can even push some of your calculations into the postscript
engine.  SVG may be more suitable these days but I haven't investigated
it myself.


David
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux

Re: pdflatex regressions? [SEC=PERSONAL]

by Roppola, Antti - BRS-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Unfortunately my personal experience is that anything that is simpler
than .tex is typesetting incomplete. OTOH, if you don't need a full
typesetting environment, near enough is likely to be good enough.

Word processing documents can be vague on structure. SVG will expect you
to handle a bunch of stuff yourself.

I have had some good results with manually coding Postscript and using
external apps to generate nested content and then importing them into
pdflatex. FYI, we ended up using Perl to generate and fill the .tex
templates.

Manually coding up Poscript to do things like sensibly break items
across columns (i.e. so the sub-heading wasn't the last element in a
column with its child entry alone at the top of the next) gave me a new
appreciation of why one should use a library to do this sort of thing.

We tinkered with Cocoon once, but I've not looked at it for years.

Cheers,

Antti

-----Original Message-----
From: linux-bounces@...
[mailto:linux-bounces@...] On Behalf Of David Tulloh
Sent: Tuesday, 13 October 2009 11:49 PM
To: Andrew Janke
Cc: linux@...
Subject: Re: [clug] pdflatex regressions?

Andrew Janke wrote:
>> Bite the bullet and port it to some typesetting or template system
>> you do understand?  I confess that, unhelpful as it is to say, I used

>> ConTeXt because of this sort of problem with LaTeX.
>>    
>
> And on this note, I'd be interested to know what others use (beyond
> ConTeXt!) for this.  Myself I tend to use inkscape and sed for complex

> graphics heavy stuff.
>  
I've generated Open Office, often to convert into MS Word.  I have used
straight template substitution driven by Make and full generation using
the Perl CPAN libraries.  This is very useful for test reports from
automated testing (with a small amount of hand editing at the end) and
large documents (hundreds of pages) generated from test databases with
multiple graphs etc.

For precise layout work I have also generated pure postscript in the
past, I've written bash scripts to generate CD covers and CD wallet
faces.  For simple drawing of lines and text postscript is really very
simple, you can even push some of your calculations into the postscript
engine.  SVG may be more suitable these days but I haven't investigated
it myself.


David
--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux


------
IMPORTANT - This message has been issued by The Department of Agriculture, Fisheries and Forestry (DAFF). The information transmitted is for the use of the intended recipient only and may contain sensitive and/or legally privileged material. It is your responsibility to check any attachments for viruses and defects before opening or sending them on.

Any reproduction, publication, communication, re-transmission, disclosure, dissemination or other use of the information contained in this e-mail by persons or entities other than the intended recipient is prohibited. The taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you have received this e-mail in error please notify the sender and delete all copies of this transmission together with any attachments. If you have received this e-mail as part of a valid mailing list and no longer want to receive a message such as this one advise the sender by return e-mail accordingly. Only e-mail correspondence which includes this footer, has been authorised by DAFF

------

--
linux mailing list
linux@...
https://lists.samba.org/mailman/listinfo/linux