Обсуждение: 8.1 PDF Documentation.
I've built the PDF docs for the past two releases, but my recent attempts to build the 8.1 docs have failed. I previously used Jade 1.2.1, SP 1.3.4, and docbook-dsssl 1.76, but trying that now results in: ! pdfTeX error (ext4): \pdfendlink ended up in different nesting level than \pdfstartlink. when running pdfjadetex. I've not been able to find out much about this error. Newer versions of the tools either crash or take days to run. I've begun the days long processing, so it will be sometime (if at all) before I can generate pdfs for the 8.1 release. Kris Jurka
8.1 pdfs are finished. These versions actually got the correct pdf bookmarks and table of contents. I can probably rebuild the 7.4 and 8.0 versions in this format if there is sufficient interest. http://www.ejurka.com/pgsql/docs/ Also the website isn't listing the 8.0 US version here: http://www.postgresql.org/docs/manuals/ But it is in the cvs repository here: http://gborg.postgresql.org/cgi-bin/cvsweb.cgi/portal/files/documentation/pdf/8.0/?cvsroot=pgweb Kris Jurka
On Thu, Nov 10, 2005 at 06:01:42PM -0500, Kris Jurka wrote: > > 8.1 pdfs are finished. These versions actually got the correct pdf > bookmarks and table of contents. I can probably rebuild the 7.4 and > 8.0 versions in this format if there is sufficient interest. Way cool. Is producing these PDFs something that could be distributed? What tools actually work for this? Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 510 893 6100 mobile: +1 415 235 3778 Remember to vote!
On Thu, 10 Nov 2005, David Fetter wrote: > On Thu, Nov 10, 2005 at 06:01:42PM -0500, Kris Jurka wrote: >> >> 8.1 pdfs are finished. These versions actually got the correct pdf >> bookmarks and table of contents. I can probably rebuild the 7.4 and >> 8.0 versions in this format if there is sufficient interest. > > Way cool. Is producing these PDFs something that could be > distributed? What tools actually work for this? > Distributed? Not really. It can't be done in parts, and there's only two sizes (US and A4). It also really only needs to be done once per release (although I suppose it would be nice to do for point releases.) It took about two days each (run serially) on a 2xOpteron 246 using debian unstable's: openjade: 1.3.2-8 opensp: 1.5.1/0-4 docbook-dsssl: 1.79-3 Kris Jurka
Kris Jurka <books@ejurka.com> writes:
> It took about two days each (run serially) on a 2xOpteron 246 using debian
> unstable's:
> openjade: 1.3.2-8
> opensp: 1.5.1/0-4
> docbook-dsssl: 1.79-3
BTW, the -US file sends my (rather old) copy of acroread into what may
be an infinite loop ... I gave up waiting for it to redisplay after
trying to scroll down the bookmarks window. Even just displaying
successive pages is painfully slow.
I suspect there's something wrong with the toolset you're using.
regards, tom lane
Tom Lane wrote: > Kris Jurka <books@ejurka.com> writes: > > It took about two days each (run serially) on a 2xOpteron 246 using debian > > unstable's: > > openjade: 1.3.2-8 > > opensp: 1.5.1/0-4 > > docbook-dsssl: 1.79-3 > > BTW, the -US file sends my (rather old) copy of acroread into what may > be an infinite loop ... I gave up waiting for it to redisplay after > trying to scroll down the bookmarks window. Even just displaying > successive pages is painfully slow. > > I suspect there's something wrong with the toolset you're using. I am using Acrobat 5.0 on XP and BSD/OS and it worked fine for me, including bookmarks. It was very fast. I am still unclear why there is no more reliable SGML toolchain for PDF output after all these years. Seems things have improved because we now have bookmarks, but still, why is this functionality not more mainstream. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073
On Tue, Nov 15, 2005 at 12:13:01AM -0500, Bruce Momjian wrote: > Tom Lane wrote: > > Kris Jurka <books@ejurka.com> writes: > > > It took about two days each (run serially) on a 2xOpteron 246 using debian > > > unstable's: > > > openjade: 1.3.2-8 > > > opensp: 1.5.1/0-4 > > > docbook-dsssl: 1.79-3 > > > > BTW, the -US file sends my (rather old) copy of acroread into what > > may be an infinite loop ... I gave up waiting for it to redisplay > > after trying to scroll down the bookmarks window. Even just > > displaying successive pages is painfully slow. > > > > I suspect there's something wrong with the toolset you're using. > > I am using Acrobat 5.0 on XP and BSD/OS and it worked fine for me, > including bookmarks. It was very fast. Acrobat works beautifully for me, too. :) > I am still unclear why there is no more reliable SGML toolchain for > PDF output after all these years. Seems things have improved > because we now have bookmarks, but still, why is this functionality > not more mainstream. See, there are these tuits, and they're round, and they're a precious resource usually spent on writing and testing C code rather than on improving (and helping transition adoption of) the SGML toolchain. Maybe we can get some kind soul(s) to donate some tuits and/or pay somebody a wage for awhile to fix the toolchain etc. :) Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 510 893 6100 mobile: +1 415 235 3778 Remember to vote!
Bruce Momjian wrote: > I am still unclear why there is no more reliable SGML toolchain for PDF > output after all these years. Seems things have improved because we now > have bookmarks, but still, why is this functionality not more mainstream. There is no SGML toolchain because SGML has been abandoned as the DocBook markup language. The toolchain that everybody else uses is based on XML DocBook. It has been proposed that we abandon SGML and jump to XML, but the idea has been shot down. And it's possible to convert our SGML to XML automatically and process it with the newer toolchain. -- Alvaro Herrera http://www.CommandPrompt.com/ PostgreSQL Replication, Consulting, Custom Development, 24x7 support
On Nov 15, 2005, at 20:23 , Alvaro Herrera wrote: > Bruce Momjian wrote: > >> I am still unclear why there is no more reliable SGML toolchain >> for PDF >> output after all these years. Seems things have improved because >> we now >> have bookmarks, but still, why is this functionality not more >> mainstream. > > There is no SGML toolchain because SGML has been abandoned as the > DocBook markup language. The toolchain that everybody else uses is > based on XML DocBook. It has been proposed that we abandon SGML and > jump to XML, but the idea has been shot down. And it's possible to > convert our SGML to XML automatically and process it with the newer > toolchain. It's my understanding that it hasn't been shot down as much as no one has shown that it will serve our needs. IIRC, there are two issues: > Marked sections: SGML DocBook (which is currently used) supports > parameter entities to be used to include or exclude certain > sections of material depending on the desired output. Profiling in > XML DocBook may be able to accomplish the same task. > > XML DocBook toolchain: making sure the tools to produce the > documentation in its myriad forms (e.g., html, manpages, PDF) works. (Just for history repeating itself, this was brought up Aug 2004 as well :) ) I don't think anyone's standing in the way. It's just that no one's shown that the XML toolchain can do the work for us. Michael Glaesemann grzm myrealbox com
Am Dienstag, 15. November 2005 06:13 schrieb Bruce Momjian: > I am still unclear why there is no more reliable SGML toolchain for PDF > output after all these years. Seems things have improved because we now > have bookmarks, but still, why is this functionality not more mainstream. I don't understand how you get from "ancient Acrobat can't display large PDF file" to "SGML toolchain is unreliable".
Alvaro Herrera wrote: > Bruce Momjian wrote: > > > I am still unclear why there is no more reliable SGML toolchain for PDF > > output after all these years. Seems things have improved because we now > > have bookmarks, but still, why is this functionality not more mainstream. > > There is no SGML toolchain because SGML has been abandoned as the > DocBook markup language. The toolchain that everybody else uses is > based on XML DocBook. It has been proposed that we abandon SGML and > jump to XML, but the idea has been shot down. And it's possible to > convert our SGML to XML automatically and process it with the newer > toolchain. OK, that makes sense. We went from LaTeX to SGML long ago with the idea that we could easily generate any output format, but that promise has never been fully realized. Could the SGML be automatically converted to XML during the print process and then converted to PDF? -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073
Peter Eisentraut wrote: > Am Dienstag, 15. November 2005 06:13 schrieb Bruce Momjian: > > I am still unclear why there is no more reliable SGML toolchain for PDF > > output after all these years. Seems things have improved because we now > > have bookmarks, but still, why is this functionality not more mainstream. > > I don't understand how you get from "ancient Acrobat can't display large PDF > file" to "SGML toolchain is unreliable". Well, historically we have had trouble generating PDF (particuarly bookmarks), and it took _days_ to generate the output, which certainly seems strange. In fact, this is the first verion that has bookmarks, and that is 1/2 the value of a PDF file for me, so I see this release as a great leap forward for us. What I do not understand is why it has taken >5 years for this to happen. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073