page number reset - comp.text.tex

Hello, I have a 14 page document (amsart, hyperref and a bunch of other packages) in which after making a minor correction, the page numbering is reset after page 5, i.e. it starts numbering from 1 again. There is nothing in the document that should cause it to do that. I cannot reproduce it in a minimal example, but it appears to be related to the fact that there is a new \subsection at the top of page 5. 1. If I insert some vertical space prior to \subsection the page number does not reset 2. If I change the textheight or textwidth the page number does not reset 3. If I insert a \clearpage just prior to \subsection the page number does reset 4. If I create a new document (with different contents) with the same \subsection at the top of a page, the page number does not reset Does anyone have any idea/experience with this phenomenon? Other than adjusting the text height/width or inserting space, is there anything I can do? Thanks! J.

Hello, Further on this: I use pdflatex. When I disable microtype, the page numbers are correct again (it doesn't change the position of the header). More specifically, if I do this: \usepackage [expansion,protrusion=3Dfalse,tracking=3Dfalse,kerning=3Dtrue,spacing=3Dtru= e] {microtype} the page counter resets, but with \usepackage [expansion=3Dfalse,protrusion=3Dfalse,tracking=3Dfalse,kerning=3Dtrue,spaci= ng=3Dtrue] {microtype} it doesn't, even though in both cases the new subsection appears in exactly the same spot. Hmmmmm. Best, J. On Nov 25, 12:58=A0pm, Joris <pin...@gmail.com> wrote: > Hello, > > I have a 14 page document (amsart, hyperref and a bunch of other > packages) in which after making a minor correction, the page numbering > is reset after page 5, i.e. it starts numbering from 1 again. =A0There > is nothing in the document that should cause it to do that. =A0I cannot > reproduce it in a minimal example, but it appears to be related to the > fact that there is a new \subsection at the top of page 5. > > 1. If I insert some vertical space prior to \subsection the page > number does not reset > 2. If I change the textheight or textwidth the page number does not > reset > 3. If I insert a \clearpage just prior to \subsection the page number > does reset > 4. If I create a new document (with different contents) with the same > \subsection at the top of a page, the page number does not reset > > Does anyone have any idea/experience with this phenomenon? =A0Other than > adjusting the text height/width or inserting space, is there anything > I can do? > > Thanks! > > J.

Am Wed, 25 Nov 2009 09:58:28 -0800 (PST) schrieb Joris: > Hello, > > I have a 14 page document (amsart, hyperref and a bunch of other > packages) in which after making a minor correction, the page numbering > is reset after page 5, i.e. it starts numbering from 1 again. There > is nothing in the document that should cause it to do that. Well as it happens there is something that cause it. > I cannot reproduce it in a minimal example, A minimal example is simply the smallest possible document you can create to reproduce the problem. So simply shorten your document as much as possible. > but it appears to be related to the > fact that there is a new \subsection at the top of page 5. > > 1. If I insert some vertical space prior to \subsection the page > number does not reset > 2. If I change the textheight or textwidth the page number does not > reset > 3. If I insert a \clearpage just prior to \subsection the page number > does reset > 4. If I create a new document (with different contents) with the same > \subsection at the top of a page, the page number does not reset > > Does anyone have any idea/experience with this phenomenon? Well if the page 6 is the one that gets the new number 1 you should look at the bottom of page 5 not at the top. Is there something that can break if it falls directly on the page break? -- Ulrike Fischer

Ulrike, Thank you for this. I've followed up your advice and I think the culprit is \Cref from the cleveref package. I managed to produce a pretty small example which resets the page number; see below. The funny thing is that the output of \Cref only appears on page 7, not on page 5 or 6. Any way of fixing this globally, i.e. not just when the problem happens to occur? Thanks! J. \documentclass{amsart} \textwidth 6in \usepackage{mathpazo} \usepackage{setspace} \usepackage{lipsum} \usepackage{cleveref} \usepackage [expansion,protrusion=false,tracking=false,kerning=true,spacing=true] {microtype} \setstretch{1.5} \begin{document} \section{Introduction} \setcounter{page}{4} \label{hello} \clearpage \subsection{Basic Rank Condition} \lipsum[1-9] \Cref{hello} \end{document}

Joris wrote: > Ulrike, > > Thank you for this. I've followed up your advice and I think the > culprit is \Cref from the cleveref package. I managed to produce a > pretty small example which resets the page number; see below. The > funny thing is that the output of \Cref only appears on page 7, not on > page 5 or 6. Any way of fixing this globally, i.e. not just when the > problem happens to occur? > > Thanks! > > J. > > > > \documentclass{amsart} > > \textwidth 6in > > \usepackage{mathpazo} > \usepackage{setspace} > \usepackage{lipsum} > \usepackage{cleveref} > \usepackage > [expansion,protrusion=false,tracking=false,kerning=true,spacing=true] > {microtype} > > \setstretch{1.5} > > \begin{document} > > \section{Introduction} > > \setcounter{page}{4} > \label{hello} > \clearpage > > \subsection{Basic Rank Condition} > > \lipsum[1-9] > > \Cref{hello} > > \end{document} > You see the odd effect without microtype. The most obvious thing to try is setting the page dimensions properly using the geometry package, rather than setting \textwidth directly. -- Joseph Wright

On Nov 25, 2:59=A0pm, Joseph Wright <joseph.wri...@morningstar2.co.uk> wrote: > Joris wrote: > > Ulrike, > > > Thank you for this. =A0I've followed up your advice and I think the > > culprit is \Cref from the cleveref package. =A0I managed to produce a > > pretty small example which resets the page number; see below. =A0 The > > funny thing is that the output of \Cref only appears on page 7, not on > > page 5 or 6. =A0Any way of fixing this globally, i.e. not just when the > > problem happens to occur? > > > Thanks! > > > J. > > > \documentclass{amsart} > > > \textwidth 6in > > > \usepackage{mathpazo} > > \usepackage{setspace} > > \usepackage{lipsum} > > \usepackage{cleveref} > > \usepackage > > [expansion,protrusion=3Dfalse,tracking=3Dfalse,kerning=3Dtrue,spacing= =3Dtrue] > > {microtype} > > > \setstretch{1.5} > > > \begin{document} > > > \section{Introduction} > > > \setcounter{page}{4} > > \label{hello} > > \clearpage > > > \subsection{Basic Rank Condition} > > > \lipsum[1-9] > > > \Cref{hello} > > > \end{document} > > You see the odd effect without microtype. The most obvious thing to try > is setting the page dimensions properly using the geometry package, > rather than setting \textwidth directly. > -- > Joseph Wright Thanks Joseph. I just tried that, but I can still recreate the problem. I've written to the authors of microtype and cleveref and will post a response here. Best, J.

On Nov 25, 2:59=A0pm, Joseph Wright <joseph.wri...@morningstar2.co.uk> wrote: > Joris wrote: > > Ulrike, > > > Thank you for this. =A0I've followed up your advice and I think the > > culprit is \Cref from the cleveref package. =A0I managed to produce a > > pretty small example which resets the page number; see below. =A0 The > > funny thing is that the output of \Cref only appears on page 7, not on > > page 5 or 6. =A0Any way of fixing this globally, i.e. not just when the > > problem happens to occur? > > > Thanks! > > > J. > > > \documentclass{amsart} > > > \textwidth 6in > > > \usepackage{mathpazo} > > \usepackage{setspace} > > \usepackage{lipsum} > > \usepackage{cleveref} > > \usepackage > > [expansion,protrusion=3Dfalse,tracking=3Dfalse,kerning=3Dtrue,spacing= =3Dtrue] > > {microtype} > > > \setstretch{1.5} > > > \begin{document} > > > \section{Introduction} > > > \setcounter{page}{4} > > \label{hello} > > \clearpage > > > \subsection{Basic Rank Condition} > > > \lipsum[1-9] > > > \Cref{hello} > > > \end{document} > > You see the odd effect without microtype. The most obvious thing to try > is setting the page dimensions properly using the geometry package, > rather than setting \textwidth directly. > -- > Joseph Wright And you're right that the example I provided still works if microtype is disabled. In my main document changing the options to microtype changed the behavior, which made me erroneously believe that microtype was to blame. Best, J.

On Nov 25, 2:22=A0pm, Joris <pin...@gmail.com> wrote: > On Nov 25, 2:59=A0pm, Joseph Wright <joseph.wri...@morningstar2.co.uk> > wrote: > > > > > Joris wrote: > > > Ulrike, > > > > Thank you for this. =A0I've followed up your advice and I think the > > > culprit is \Cref from the cleveref package. =A0I managed to produce a > > > pretty small example which resets the page number; see below. =A0 The > > > funny thing is that the output of \Cref only appears on page 7, not o= n > > > page 5 or 6. =A0Any way of fixing this globally, i.e. not just when t= he > > > problem happens to occur? > > > > Thanks! > > > > J. > > > > \documentclass{amsart} > > > > \textwidth 6in > > > > \usepackage{mathpazo} > > > \usepackage{setspace} > > > \usepackage{lipsum} > > > \usepackage{cleveref} > > > \usepackage > > > [expansion,protrusion=3Dfalse,tracking=3Dfalse,kerning=3Dtrue,spacing= =3Dtrue] > > > {microtype} > > > > \setstretch{1.5} > > > > \begin{document} > > > > \section{Introduction} > > > > \setcounter{page}{4} > > > \label{hello} > > > \clearpage > > > > \subsection{Basic Rank Condition} > > > > \lipsum[1-9] > > > > \Cref{hello} > > > > \end{document} > > > You see the odd effect without microtype. The most obvious thing to try > > is setting the page dimensions properly using the geometry package, > > rather than setting \textwidth directly. > > -- > > Joseph Wright > > Thanks Joseph. =A0I just tried that, but I can still recreate the > problem. =A0I've written to the authors of microtype and cleveref and > will post a response here. =A0Best, J. Cleveref is doing calculations with the registers \count0 and count1. Somehow, the output routine is triggered before the end of the group that is supposed to restore those counters. Since the page counter is normally \count0, and since the output routine globally steps the page counter, \count0 becomes globally reset. If the commands \countdef\count@consecutive=3D0% \countdef\count@group=3D1% are omitted from the definition of \@cref in cleveref.sty, and if those registers are allocated elsewhere with \newcount\count@consecutive \newcount\count@group Then the page reset doesn't occur. Count register 0 actually doesn't seem to be needed. It seems to be used only to pass along the result of some \if-test. This could be done any number of other ways. Still, it would be interesting to know why the output routine is triggered in the middle of this command. Tiny changes in the length of the text of the example (by changing fonts or removing microtype) seem to change the timing of the output routine and therefore whether this page reset happens. There is another command in cleverref that does arithmetic with these count registers (\cref@isrefconsecutive), but it seems to be unrelated to this problem. I hope this helps. Dan

On Nov 25, 12:12=A0pm, Ulrike Fischer <ne...@nililand.de> wrote: > Am Wed, 25 Nov 2009 09:58:28 -0800 (PST) schrieb Joris: > > but it appears to be related to the > > fact that there is a new \subsection at the top of page 5. > > > 1. If I insert some vertical space prior to \subsection the page > > number does not reset > > 2. If I change the textheight or textwidth the page number does not > > reset > > 3. If I insert a \clearpage just prior to \subsection the page number > > does reset > > 4. If I create a new document (with different contents) with the same > > \subsection at the top of a page, the page number does not reset > > > Does anyone have any idea/experience with this phenomenon? > > Well if the page 6 is the one that gets the new number 1 you should > look at the bottom of page 5 not at the top. Is there something that > can break if it falls directly on the page break? The page number is printed _after_ the page (and more) is accumulated: during the output routine. Therefore, anything changing the page number from 6 to 1 would likely be on page 6, or the beginning of page 7 (or in the output routine). Probably anything on page 5 would show up as a change in the number of page 5. Dan

On Nov 25, 5:50=A0pm, Dan <lueck...@uark.edu> wrote: > On Nov 25, 2:22=A0pm, Joris <pin...@gmail.com> wrote: > > > > > On Nov 25, 2:59=A0pm, Joseph Wright <joseph.wri...@morningstar2.co.uk> > > wrote: > > > > Joris wrote: > > > > Ulrike, > > > > > Thank you for this. =A0I've followed up your advice and I think the > > > > culprit is \Cref from the cleveref package. =A0I managed to produce= a > > > > pretty small example which resets the page number; see below. =A0 T= he > > > > funny thing is that the output of \Cref only appears on page 7, not= on > > > > page 5 or 6. =A0Any way of fixing this globally, i.e. not just when= the > > > > problem happens to occur? > > > > > Thanks! > > > > > J. > > > > > \documentclass{amsart} > > > > > \textwidth 6in > > > > > \usepackage{mathpazo} > > > > \usepackage{setspace} > > > > \usepackage{lipsum} > > > > \usepackage{cleveref} > > > > \usepackage > > > > [expansion,protrusion=3Dfalse,tracking=3Dfalse,kerning=3Dtrue,spaci= ng=3Dtrue] > > > > {microtype} > > > > > \setstretch{1.5} > > > > > \begin{document} > > > > > \section{Introduction} > > > > > \setcounter{page}{4} > > > > \label{hello} > > > > \clearpage > > > > > \subsection{Basic Rank Condition} > > > > > \lipsum[1-9] > > > > > \Cref{hello} > > > > > \end{document} > > > > You see the odd effect without microtype. The most obvious thing to t= ry > > > is setting the page dimensions properly using the geometry package, > > > rather than setting \textwidth directly. > > > -- > > > Joseph Wright > > > Thanks Joseph. =A0I just tried that, but I can still recreate the > > problem. =A0I've written to the authors of microtype and cleveref and > > will post a response here. =A0Best, J. > > Cleveref is doing calculations with the registers \count0 and count1. > Somehow, the output routine is triggered before the end of the group > that is supposed to restore those counters. Since the page counter is > normally \count0, and since the output routine globally steps the > page > counter, \count0 becomes globally reset. > > If the commands > =A0 \countdef\count@consecutive=3D0% > =A0 \countdef\count@group=3D1% > are omitted from the definition of \@cref in cleveref.sty, and if > those > registers are allocated elsewhere with > =A0 \newcount\count@consecutive > =A0 \newcount\count@group > > Then the page reset doesn't occur. Count register 0 actually > doesn't seem to be needed. It seems to be used only to pass > along the result of some \if-test. This could be done any number > of other ways. > > Still, it would be interesting to know why the output routine is > triggered in the middle of this command. > > Tiny changes in the length of the text of the example (by changing > fonts or removing microtype) seem to change the timing of the > output routine and therefore whether this page reset =A0happens. > > There is another command in cleverref that does arithmetic with > these count registers (\cref@isrefconsecutive), but it seems to > be unrelated to this problem. > > I hope this helps. > > Dan Dan, Thank you very much for both of your messages. This helps a lot. I will try the changes that you are suggesting. Thanks again. Best, J.

Dan <luecking@uark.edu> writes: >On Nov 25, 2:22=A0pm, Joris <pin...@gmail.com> wrote: >> On Nov 25, 2:59=A0pm, Joseph Wright <joseph.wri...@morningstar2.co.uk> >> wrote: >Cleveref is doing calculations with the registers \count0 and count1. >Somehow, the output routine is triggered before the end of the group >that is supposed to restore those counters. Since the page counter is >normally \count0, and since the output routine globally steps the >page >counter, \count0 becomes globally reset. > >If the commands > \countdef\count@consecutive=0% > \countdef\count@group=1% >are omitted from the definition of \@cref in cleveref.sty, and if >those >registers are allocated elsewhere with > \newcount\count@consecutive > \newcount\count@group i had spotted those, and was then stuck on the issue of how any assignment could possibly escape from within the group the definition happens in. >Then the page reset doesn't occur. Count register 0 actually >doesn't seem to be needed. It seems to be used only to pass >along the result of some \if-test. This could be done any number >of other ways. > >Still, it would be interesting to know why the output routine is >triggered in the middle of this command. that hadn't occurred to me. i don't see how it could happen, though. (in my research, long ago, one used to worry all the time about interruption, but tex isn't really like the kernel of an operation system.) >Tiny changes in the length of the text of the example (by changing >fonts or removing microtype) seem to change the timing of the >output routine and therefore whether this page reset happens. > >There is another command in cleverref that does arithmetic with >these count registers (\cref@isrefconsecutive), but it seems to >be unrelated to this problem. > >I hope this helps. i think tony cubitt ought to be alerted. his concern for register usage is less significant now than it once was. -- Robin Fairbairns, Cambridge

On Nov 25, 6:07=A0pm, r...@cl.cam.ac.uk (Robin Fairbairns) wrote: > =A0Dan <lueck...@uark.edu> writes: > >On Nov 25, 2:22=3DA0pm, Joris <pin...@gmail.com> wrote: > >> On Nov 25, 2:59=3DA0pm, Joseph Wright <joseph.wri...@morningstar2.co.u= k> > >> wrote: > >Cleveref is doing calculations with the registers \count0 and count1. > >Somehow, the output routine is triggered before the end of the group > >that is supposed to restore those counters. Since the page counter is > >normally \count0, and since the output routine globally steps the > >page > >counter, \count0 becomes globally reset. > > >If the commands > > =A0\countdef\count@consecutive=3D0% > > =A0\countdef\count@group=3D1% > >are omitted from the definition of \@cref in cleveref.sty, and if > >those > >registers are allocated elsewhere with > > =A0\newcount\count@consecutive > > =A0\newcount\count@group > > i had spotted those, and was then stuck on the issue of how any > assignment could possibly escape from within the group the definition > happens in. > > >Then the page reset doesn't occur. Count register 0 actually > >doesn't seem to be needed. It seems to be used only to pass > >along the result of some \if-test. This could be done any number > >of other ways. > > >Still, it would be interesting to know why the output routine is > >triggered in the middle of this command. > > that hadn't occurred to me. =A0i don't see how it could happen, though. > (in my research, long ago, one used to worry all the time about > interruption, but tex isn't really like the kernel of an operation > system.) > > >Tiny changes in the length of the text of the example (by changing > >fonts or removing microtype) seem to change the timing of the > >output routine and therefore whether this page reset =A0happens. > > >There is another command in cleverref that does arithmetic with > >these count registers (\cref@isrefconsecutive), but it seems to > >be unrelated to this problem. > > >I hope this helps. > > i think tony cubitt ought to be alerted. =A0his concern for register > usage is less significant now than it once was. > -- > Robin Fairbairns, Cambridge Thanks Robin. I've done that with proper attribution. Best, J.