Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Site Tag Cloud

2-category 2-category-theory abelian-categories adjoint algebra algebraic algebraic-geometry algebraic-topology analysis analytic-geometry arithmetic arithmetic-geometry book bundles calculus categorical categories category category-theory chern-weil-theory cohesion cohesive-homotopy-type-theory cohomology colimits combinatorics complex complex-geometry computable-mathematics computer-science constructive cosmology deformation-theory descent diagrams differential differential-cohomology differential-equations differential-geometry digraphs duality elliptic-cohomology enriched fibration foundation foundations functional-analysis functor gauge-theory gebra geometric-quantization geometry graph graphs gravity grothendieck group group-theory harmonic-analysis higher higher-algebra higher-category-theory higher-differential-geometry higher-geometry higher-lie-theory higher-topos-theory homological homological-algebra homotopy homotopy-theory homotopy-type-theory index-theory integration integration-theory k-theory lie-theory limits linear linear-algebra locale localization logic mathematics measure-theory modal modal-logic model model-category-theory monad monads monoidal monoidal-category-theory morphism motives motivic-cohomology nforum nlab noncommutative noncommutative-geometry number-theory of operads operator operator-algebra order-theory pages pasting philosophy physics pro-object probability probability-theory quantization quantum quantum-field quantum-field-theory quantum-mechanics quantum-physics quantum-theory question representation representation-theory riemannian-geometry scheme schemes set set-theory sheaf sheaves simplicial space spin-geometry stable-homotopy-theory stack string string-theory superalgebra supergeometry svg symplectic-geometry synthetic-differential-geometry terminology theory topology topos topos-theory tqft type type-theory universal variational-calculus

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to nForum
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 22nd 2016
    In Firefox 45 on my Debian system, the page https://ncatlab.org/nlab/show/lift looks like this:

    https://imgur.com/iioIWEx

    Is this a known issue?
    • CommentRowNumber2.
    • CommentAuthorspitters
    • CommentTimeMar 22nd 2016

    Same here. I’ve tried these links, but they don’t resolve the issues for me. More discussion at the n-cafe (couple of posts at the bottom of the page). Apparently, there is also an issue with chrome, but that one works for me.

    I’ve mentioned this before at the nforum, but I cannot find the thread at the moment.

    • CommentRowNumber3.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 22nd 2016

    Thanks for those links - they’ve helped me to solve the issue. :-) What I did was:

    • Download version 1.959 of Latin Modern Math.

    • Unpack it, and move the latinmodernmath-regular.otf file into my ~/.fonts directory. (Well, into the otf subdirectory I’d previously created therein.)

    • Edited the file /etc/fonts/conf.avail/65-fonts-lmodern.conf, commenting out the line <dir>/usr/share/texmf/fonts/opentype/public/lm-math</dir>.

    This makes fontconfig use the version of LatinModernMath-Regular in my ~/.fonts directory, rather than the one in the fonts-lmodern package (which is at the latest available version for Debian Jessie).

    • CommentRowNumber4.
    • CommentAuthorUrs
    • CommentTimeMar 22nd 2016

    Could one still take some action to ensure that not every user will have to go through this procedure, but that the problem be solved for good somehow? Will it just go away when the next version of Firefox comes properly bundled with the requisite fonts? Should one still file a bug report alerting developers that this is presently missing?

    • CommentRowNumber5.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 22nd 2016

    Well, it’s my understanding that Firefox, and Web browsers in general, don’t bundle fonts at all - instead they use the fonts available on the system on which they’re running. What’s happened is that the Firefox devs have changed how they process fonts - cf. this bug report - such that older versions of some maths fonts will be displayed with excessive line-height. Thus, what needs to happen is that more recent versions of fonts like Latin Modern Math be available on users’ systems.

    In terms of Debian, this issue has been addressed by updating the fonts-lmodern package in the ’Testing’ archive (cf. this), so users running ’Testing’ will have access to the newer version of Latin Modern Math. However, users running ’Stable’ (such as myself), or older releases of Debian, will not have access to this updated package unless it’s backported to those releases by a Debian volunteer. Hence the need for the workaround I described above.

    I can’t speak to what would need to be done in the context of other operating systems (e.g. Ubuntu, OS X).

    • CommentRowNumber6.
    • CommentAuthorspitters
    • CommentTimeMar 22nd 2016

    Ubuntu 15.10 has fonts-lmodern 2.004, recent enough, but this does not solve the issue for me. I haven’t had the time to properly find out which fonts are actually used, but the fonts are available on my system.

    • CommentRowNumber7.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 22nd 2016

    Well, the version of the package is not the same as the version of the fonts within the package. My system has version 2.004.4-5 of the fonts-lmodern package as well, but the version of the Latin Modern Math font within that packge isn’t recent enough. The closeness of the package version number to the font version number initially misled me also; but no, the latest version of Latin Modern Math is indeed 1.959, whereas according to the Debian bug report, the version in fonts-lmodern 2.004.4-5 is 1.950.

    So, given this, I’d like to suggest trying the workaround I described above (modulo any differences between Debian packaging/config and Ubuntu packaging/config).

    • CommentRowNumber8.
    • CommentAuthorUrs
    • CommentTimeMar 23rd 2016

    but the version of the Latin Modern Math font within that packge isn’t recent enough

    Is that an information that could still be filed somewhere such that somebody would fix it globally?

    • CommentRowNumber9.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 23rd 2016

    I might not be understanding what you mean by “fix it globally”, but insofar as I read it as “fix it in one place so that all site users won’t face this issue”, the answer is basically: “No”.

    As I’ve noted, this is an operating-system level issue, such that how the issue can be addressed differs from OS to OS. The fonts-lmodern package to which I’ve referred is specific to the Debian OS (and apparently to at least some Debian derivatives, such as Ubuntu); it’s not something one installs on (for exampel) OS X or Windows.

    In the case of Ubuntu, there’s a reasonable chance that the workaround I described for my Debian system will work without changes for an Ubuntu system. (Not necessarily, though; in a number of significant areas, Ubuntu does things differently from Debian.) However, I don’t know whether the fix to the fonts-lmodern package in Debian will automatically flow through (eventually) to Ubuntu; that would be something to ask on Ubuntu user fora.

    Further, fixing this issue on other OSes such as Windows, Android, iOS or OS X, will require still other processes, since they manage software (and fonts in particular) in different ways from Debian, Ubuntu, and each other.

    Off the top of my head, the closest one might get to be able to addressing the issue in a more general way is to set up the nLab wiki so that it uses Latin Modern Math as a “Web font”. Browsers supporting Web fonts can download a font specified by a Web site “on the fly”, so that it can be used without site visitors having to install it themselves. However, this requires that:

    • the font in question have a license which allows it to be used as a Web font; and

    • site visitors have a sufficiently recent browser, one which supports “Web font” functionality.

    So even taking this route, it won’t necessarily result in the issue being addressed for all site users.

    • CommentRowNumber10.
    • CommentAuthorUrs
    • CommentTimeMar 23rd 2016
    • (edited Mar 23rd 2016)

    I may be naive, but my understanding is that we are talking about a bug, and that when faced with a bug affecting a public site we are not content with some select users (such as you and Bas) knowing how to fiddle a workaround, but we want the bug to be fixed such that the generic user surfing to our site doesn’t get the first impression that it’s basically broken.

    All I am trying to suggest is that besides you and Bas sorting out how to fix the issue on your local system, thereby reducing the number of bug-free experiences by 2, maybe you might thankfully have a further second to spare such as to trigger the issue to be fixed for all present and future users.

    I know that you may have better things to do, which is fair enough, but since you are looking into it already, you might just know.

    What you reported in #7 sounded like a bug in package bundeling. If that is so, I am wondering if you could “file a bug report” on that problem somewhere, so that it gets fixed.

    • CommentRowNumber11.
    • CommentAuthorZhen Lin
    • CommentTimeMar 23rd 2016
    • (edited Mar 23rd 2016)

    Reporting it as a bug to one Linux distribution will only fix it for the users of that distribution. There really is no global solution except for forcing the use of web fonts, as suggested in #9.

    • CommentRowNumber12.
    • CommentAuthorUrs
    • CommentTimeMar 23rd 2016

    Okay, thanks, now I see. I’ll forward that information.

    • CommentRowNumber13.
    • CommentAuthorRodMcGuire
    • CommentTimeMar 23rd 2016

    It might be worthwhile to report the problem and fix to https://groups.google.com/forum/#!forum/mathjax-users. You have to join that group to post.

    (that group has discussed the “small vertical line” problem)

    • CommentRowNumber14.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 23rd 2016

    So, to summarise:

    (i) There was a bug in Firefox around how Firefox uses fonts, which has been fixed.

    (ii) As a result of this fix, certain fonts, such as older versions of Latin Modern Math, render with excessive line height.

    (iii) Newer versions of Latin Modern Math have been fixed so that this excessive line height issue does not occur.

    (iv) The problem is, people have older versions of Latin Modern Math installed on their systems. So users need the newer versions installed on their systems, and to make sure the newer font version is used by the fixed versions of Firefox.

    (v) There is no cross-OS way of doing (iv).

    (vi) On some versions of Debian, the fonts-lmodern package has been updated with the fixed version of Latin Modern Math. Since this update is not yet available for other versions of Debian (such as Debian Stable / ’Jessie’), pending a Debian volunteer doing the necessary work, the workaround described in #3 is necessary for at least some of those versions.

    (vii) For Ubuntu, there does not appear to be an updated version of fonts-lmodern package containing the fixed version of Latin Modern Math (or at least, not for all versions of Ubuntu). Ubuntu users could lodge a bug report about this (if it’s not already been done). In the meantime, the workaround described in #3 might work.

    (viii) For other operating systems, users will need to either:

    (a) (in the case of OSes like Linux or *BSDs) file a bug with the operating system maintainers; or

    (b) manually download, install, and possibly configure, the fixed version of Latin Modern Math according to the procedures appropriate for their operating system.

    (ix) Points (iv) to (viii) might be able to be avoided to a certain extent if the nLab site makes use of Latin Modern Math as a Web font, so that users don’t need to have the latest version of that font installed on their system.

    • CommentRowNumber15.
    • CommentAuthorspitters
    • CommentTimeMar 29th 2016
    • (edited Mar 29th 2016)

    Re #3, Thanks. That worked for me.

    • CommentRowNumber16.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 30th 2016

    You’re welcome! Glad it sorted things out for you. :-)

    • CommentRowNumber17.
    • CommentAuthorUrs
    • CommentTimeMar 30th 2016

    Over via email, our admin, Adeel Khan, ends up agreeing with #8, #10 in thinking that it’s a bug which needs to be fixed by the operating system maintainers. On the risk of getting on everyone’s nerves I’ll ask again: might anyone care about making sure that a corresponding bug report is being issued for them to notice?

    • CommentRowNumber18.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 30th 2016

    OS maintainers will only consider it a bug if their OS ships the Latin Modern Math font as standard. Is that the case for the various versions of Windows and OS X?

    • CommentRowNumber19.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 30th 2016

    (Keeping in mind that Microsoft and Apple only support more recent versions of their OSes - for example, support for Win XP ended in April 2014, and this comment suggests that only the current and previous releases of OS X are actively supported by Apple.)

    • CommentRowNumber20.
    • CommentAuthorUrs
    • CommentTimeMar 30th 2016

    Maybe it’s the Firefox developers who should consider it as a bug of their software package?

    Or somebody. Why should there be any subtlety. Firefox is supposed to properly display pages containing MathML, and on some systems it doesn’t. The nLab could try to work around it by using Web fonts, but whether we work around it or not, there is something that isn’t working which ought to be working – on our site and potentially on any other site using MathML . Why should that not be a bug that may be reported and fixed?

    Let me maybe explain why I care about this: I have various lengthy nLab pages which presently only display on Firefox, because on all other browsers, that fall back to MathJax, these pages take minutes to render, much longer than users are prepared to wait until deciding that the page must be broken. So if even the last bastion, Firefox, crumbles, then nothing is left.

    • CommentRowNumber21.
    • CommentAuthorAlexisHazell
    • CommentTimeMar 30th 2016

    Well, as per #14(i), this issue has come about because Firefox have fixed a bug, #1170782. Fixing this bug caused the excessive line height problem due to some fonts (such as Latin Modern Math) not being implemented adequately; this resulted in bug #1235757 being reported. From the comments on that bug report, it seems to me that Mozilla developers are taking the position that Firefox is now doing the right thing, such that font implementations need to deal with that. Perhaps you could try explaining to the Mozilla developers, via that bug report, the difficult situation you’re facing? You might be able to convince them that they should try to find a workaround (other than reversing the changes applied in #1170782, which they might not be willing to do).

    (As a software developer myself, I have sympathy for both sides here. It’s frustrating, as a developer, to have to be able to handle other software or data that doesn’t work as it should, and to thus have to make one’s own code more complicated, more difficult to reason about, and thus more difficult to maintain. On the other hand, as a user, I want things to “just work”, and to not have to deal with a responsibility tennis match between developers.)

  1. Hi, why isn’t it possible to fix the bug without changes in the maintainer’s system by providing the font file by the server and include them? Normally, this is done by the css command like @font-face { font-family: ’correct-font’; src: url(’path/to/latinmodernMath.otf’) format(’opentype’); }

    • CommentRowNumber23.
    • CommentAuthorAlexisHazell
    • CommentTimeJan 15th 2017

    This was addressed in #9. The problem is that it requires browsers supporting Web font functionality, and not all browsers, particularly older browsers, do so[1]. Further, not all users are able to use a Web-font-enabled browser (e.g. because they work for an organisation mandating use of a particular browser). Having said that, Adeel could possibly:

    • examine site data to determine which browser versions are used by site visitors overall,

    • poll regular site users to determine which browser(s) they’re using,

    and make a decision to use or not use Web fonts accordingly.

    [1] Cf. this, which suggests that ~6% of browsers in use don’t support Web fonts.

    • CommentRowNumber24.
    • CommentAuthorMike Shulman
    • CommentTimeJan 15th 2017

    FWIW, a friend recently upgraded a very old computer from Ubuntu 12.04 to 14.04 and then to 16.04. The upgrade to 14.04 created this problem, but then the second upgrade to 16.04 fixed it without any further intervention. So I don’t know what the status of the bugs related to this are, but it seems to have been fixed somehow in Xenial without the user needing to install fonts manually.

    • CommentRowNumber25.
    • CommentAuthorTobyBartels
    • CommentTimeJan 16th 2017

    I found that 14.04 introduced so many problems that I went back to 12.04, and I still haven't tried 16.04. If it's fixing problems introduced in 14.04, then maybe I should!

    • CommentRowNumber26.
    • CommentAuthorMike Shulman
    • CommentTimeJan 16th 2017

    My own experience with 16.04 has been pretty good.

  2. Re #23: As I understand, the @font-face command is simply ignored if not supported by the browser, so the solution would work for 94% (according to your numbers) and don’t cause troubles for the remaining 6%.

    Anyway for my personal usage: how would the exact @font-face command look like and where do I have to put it in? My office browser (firefox) supports web fonts but technical support has not so far managed to install the up-to-date font. So I wanted to fix the problem myself by overwriting some css file but all reasonable attempts didn’t work out.

    • CommentRowNumber28.
    • CommentAuthorAlexisHazell
    • CommentTimeJan 18th 2017

    True, using Web fonts properly shouldn’t cause additional issues for the 6%; it just won’t solve the original issue for them, which is a particular problem if regular site contributors are amongst that 6%.

    To use Web fonts, there needs to be a publicly-accessible location online from which the fonts are served, in the appropriate format. (For example, OTF or TTF; I’m not sure fonts contained in an archive, e.g. a zip archive, would work.) Unless the relevant fonts are already available on the nLab server (check with Adeel), you’ll need the URL of such a location. Then, in the appropriate CSS file, you’d use something like:

    @font-face {
        font-family: "[font name]";
        src: url("[url]") format("[font format, e.g. truetype");
    }
    
  3. So what is the appropriate CSS file? (I tried all I could find) Or perhaps I got the font name wrong?