Microsoft Font Validator lives!

124»

Comments

  • Dave Crossland
    Dave Crossland Posts: 1,429
    edited January 2016
    Hin-Tak Leung said:

    Oh, there are some reports that the first time a fontconfig based application is run, it needs to build the font cache, so if you have never had any other fontconfig based applications on your system (basically any Unix-like GUI application), the first time the GUI launches may take a while. I have no first hand experience on that...
    In FontForge, I used cocoadialog to show a progress bar during that wait. 

    https://mstratman.github.io/cocoadialog/
     
    https://github.com/fontforge/fontforge/blob/master/osx/FontForge.app/Contents/MacOS/fcprogress.pl

    https://github.com/fontforge/fontforge/blob/master/osx/FontForge.app/Contents/MacOS/FontForge#L116-L117

    When I ran it, it just opened the script in text edit:



    But I could solve this with right click, open with, Terminal. 




    It took about a minute to actually open, in which time I just saw the Terminal saying nothing. 

    Rather than CocoaDialog, you could just echo something to the terminal to say "FontConfig Progressing, please wait" :) 

    I suggest the default report output location be ~/Desktop, if that's possible. 
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    Dave Crossland said:
    Hin-Tak Leung said:

    Oh, there are some reports that the first time a fontconfig based application is run, it needs to build the font cache, so if you have never had any other fontconfig based applications on your system (basically any Unix-like GUI application), the first time the GUI launches may take a while. I have no first hand experience on that...
    In FontForge, I used cocoadialog to show a progress bar during that wait. 
    ...

    It took about a minute to actually open, in which time I just saw the Terminal saying nothing. 

    Rather than CocoaDialog, you could just echo something to the terminal to say "FontConfig Progressing, please wait" :) 

    I suggest the default report output location be ~/Desktop, if that's possible. 
    Thanks for the suggestion. The switch to ~/desktop as default is sitting in my repo now.

    Not very sure about the wait - I suppose I could echo a longer message than letting the ".../FontVal ; exit" part confuse people, but what ever I echo it needs to happen after that "exit" or there is no point...

    The "open with" thing is because you have modified your default file type binding. I don't think I can do anything about that. The script is meant to be a quick and dirty thing to set up to environment to launch the real thing - I might write a proper launcher later...
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    Paul van der Laan said:
    Thank you Thomas, for weighing in (no pun intented).

    BTW, something else that definitely needs to be updated: FVAL currently red flags newer versions of the OS/2 (v.4) and gasp (v.1) tables.
    What was it about gasp v1 table? There is a v0 check which limits rangebehavior compared to v1. 
  • Paul van der Laan
    Paul van der Laan Posts: 242
    Hin-Tak Leung said:
    What was it about gasp v1 table? There is a v0 check which limits rangebehavior compared to v1. 
    The additional ClearType rendering bits that are part of v.1 are marked as errors.
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited January 2016
    Paul van der Laan said:
    Hin-Tak Leung said:
    What was it about gasp v1 table? There is a v0 check which limits rangebehavior compared to v1. 
    The additional ClearType rendering bits that are part of v.1 are marked as errors.
    I don't think that's the case - yes, the current error message is not helpful, it just summarily say something is wrong, without saying *which* and *how*. Looking at the code, I suspect you have rubbish in the higher bits. Even in a v1 table, only 4 bits of the full 16 bits are used (in v0, only the lowest 2 bits are used). I think you may have rubbish in the higher 12 bits.

    Can you share a font sample?

    I just pushed an untested "better gasp error reporting" commit out to actually show the bits errors in the details section. It should say something like "version=1, range #?, rangeGaspBehavior=0x000?" in the detail box now, instead of empty. I suspect you have 0x???? where the first 3 are not zero. Can you give that a try? (sorry you'll need to build from source until I do another build next).
  • Paul van der Laan
    Paul van der Laan Posts: 242
    My comments were based on the original version of Font Validator, not your version.
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    Paul van der Laan said:
    My comments were based on the original version of Font Validator, not your version.
    Argh, that was built 2003, before the opentype spec went iso!
  • Paul van der Laan
    Paul van der Laan Posts: 242
    At the time when I made my original comment, the 2003 version was still the only one available. Haven’t looked into the new version TBH.
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    Paul van der Laan said:
    At the time when I made my original comment, the 2003 version was still the only one available. Haven’t looked into the new version TBH.
    Hmm, it wasn't the 2003 version that was opened, but a ~2009 version that was last updated around the 2nd iso spec (current spec is 3rd from last year). The os/2 change to v3 to v5 was in my branch but not in Microsoft's. So half of your sentence, the OS/2 part, was correct for the ~2009 version.
  • Paul van der Laan
    Paul van der Laan Posts: 242
    Let me rephrase then: “At the time when I made my original comment, the 2003 version was still the only one publicly available.”

    Do we really need to discuss semantics?
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited January 2016
    Paul van der Laan said:
    Let me rephrase then: “At the time when I made my original comment, the 2003 version was still the only one publicly available.”

    Do we really need to discuss semantics?
    Hmm, that was what confused me: your comment was made as a response to this thread, which was an announcement about newer versions than the 2003 one. So your comment was made *after* the newer ones were publicly available.

    I think we disagreed on what "current" means. "Current" for you was the 2003 version, whereas it means to me something a bit closer to the code in the public repository and the posted snapshot binaries at the time. "2003" wasn't current, quite far from it.
  • Paul van der Laan
    Paul van der Laan Posts: 242
    I say potato, you say potato. C’est la vie.
  • LeMo aka PatternMan aka Frank E Blokland
    LeMo aka PatternMan aka Frank E Blokland Posts: 723
    And some say the same but write it differently…
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    Remember, it is a project being progressed by limited resources. Time used in chasing issues which aren't issues, is time lost, and time that could be put to other use.
  • Gustavo Martinez
    Gustavo Martinez Posts: 2
    First time poster. Long time reader. I want to install Microsoft Validator on Windows but don't know how. Can someone provide step by step instructions for the average Windows user who's not savvy with Github? (I saw Thomas Phinney suggesting something similar for the average Mac user for the Microsoft Validator for Mac). Thank you.
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited January 2016
    @Gustavo Martinez just download, unzip, and run the "FontVal.exe" within the unzipped folder. quite simple.

    http://sourceforge.net/projects/hp-pxl-jetready/files/Microsoft Font Validator/FontVal-bin-2016_01_22.zip/download

    A bit of technical detail: 

    There is no need for an installer any more - it was needed in the 2003 version as far as I see because the digital signature checking has to register itself for the windows crypto stuff (called "windows trust API) to actually does the checking. Since I rewrote that part in November in a platform neutral manner (Microsoft did not open that part of the FontVal source code anyway), the executable no longer depends on interacting with windows' trust API. In any case, the old windows installer packaging axil files weren't open either, and not needed and nobody is reimplementing it (yet).

    Please feel free to click the donate button if you find it useful.
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    Gustavo Martinez said:
    I saw Thomas Phinney suggesting something similar for the average Mac user for the Microsoft Validator for Mac
    The first mac version was a bit troublesome with the command line interface and the need to load extra libraries. The 2nd mac version (just out) with the GUI works mostly like the windows one, just unpack the download, find the red icon inside and click it.

    The main difference between windows and Mac experience on "installation" is now that for one, "unpack" means unzip, for the other "unpack" means double-click to mount disk image on Mac desktop. After unpacking,  it is just find that red icon and double-click to launch the application itself.
  • Gustavo Martinez
    Gustavo Martinez Posts: 2
    Thank you! The link you provided goes directly to download the app. and as you said is quite simple from there. The problem I was facing is I was trying to use the link on the first post on this thread that takes you to Github and I just couldn't make it work from there.
  • John Hudson
    John Hudson Posts: 3,186
    edited February 2016
    just download, unzip, and run the "FontVal.exe" within the unzipped folder. quite simple.
    http://sourceforge.net/projects/hp-pxl-jetready/files/Microsoft Font Validator/FontVal-bin-2016_01_22.zip/download

    I can run this file, but if I associate .otf and .ttf files with it, such that I can right click on these and select 'Open with...', the program doesn't run.

    (Win7)
  • John Hudson
    John Hudson Posts: 3,186
    It would be nice if the 'About Font Validator' pop-up in the Help menu presented the current version number instead of 1.0.00
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited February 2016
    John Hudson said:
    just download, unzip, and run the "FontVal.exe" within the unzipped folder. quite simple.
    http://sourceforge.net/projects/hp-pxl-jetready/files/Microsoft Font Validator/FontVal-bin-2016_01_22.zip/download

    I can run this file, but if I associate .otf and .ttf files with it, such that I can right click on these and select 'Open with...', the program doesn't run.

    (Win7)
    I'll think about how one might accommodate that usage.

    Logically, the main thing it should open (although it does not, and never did) by file association is a project file for multi-font testing though. And the other thing it should open by file association, and again does not and never did, is xml reports.

    Filed upstream as:
    https://github.com/Microsoft/Font-Validator/issues/12

    (in general I prefer issues to be filed upstream, unless it is specific to changes I made - just so that Microsoft folks can get to them)
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited February 2016

    John Hudson said:
    It would be nice if the 'About Font Validator' pop-up in the Help menu presented the current version number instead of 1.0.00
    That's mainly because nobody bothered to change it. Internally it identify itself as 1.1.<build#>, I think, and it is relatively easy to change.

    I have thought about bumping the version up to 2.0 when freetype based rasterisation - the last piece Microsoft did not open - happens though.

    Here to another issue.

    https://github.com/Microsoft/Font-Validator/issues/13 
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited February 2016
    John Hudson said:
    It would be nice if the 'About Font Validator' pop-up in the Help menu presented the current version number instead of 1.0.00

    You found an obscure bug in mono :). Just filed:

    https://bugzilla.xamarin.com/show_bug.cgi?id=38515

    It is the sort of info on windows which is derived from multiple sources and from X and fallback on y and if y hasn't been filled, use z. So I 'll let the mono people figure where they missed something.

    The 'about ...' info is correct if you run it on windows :( . So I guess you are using the OS X binary? Please consider making a donation.

    I have a work-around, but I'd prefer the mono people to fix the bug, instead of me working around it.
  • John Hudson
    John Hudson Posts: 3,186
    The 'about ...' info is correct if you run it on windows  . So I guess you are using the OS X binary? 

    No, I'm using the Windows version.

    The version string in the fontval.exe file properties reads 1.1.5865.23048, but in the app, the about info still reads 1.0.0.0
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    John Hudson said:
    No, I'm using the Windows version.

    The version string in the fontval.exe file properties reads 1.1.5865.23048, but in the app, the about info still reads 1.0.0.0
    Okay, there is a bug in mono but not where I thought it was. I'll go and add a note in the xamarin bugzilla soon.

    Anyway, how soon/keen do you want this? Because the CFF update will be out possibly Monday or thereabouts. I could have something that works in, but on the whole it is basically stabilized snapshots until rasterisation test & 2.0, I think; and I'd rather give the mono people a chance to do it correctly.
  • John Hudson
    John Hudson Posts: 3,186
    Not critical. I just found it unnerving because I was replacing my old version and seeing an in-app version string that was older. Thanks.
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    John Hudson said:
    Not critical. I just found it unnerving because I was replacing my old version and seeing an in-app version string that was older. Thanks.
    It seem that you found not just one bug in mono but two :( .
    Filed a 2nd one:

    https://bugzilla.xamarin.com/show_bug.cgi?id=38553

    I changed my mind - the work around is:

    https://github.com/HinTak/Font-Validator/commit/ed14e76df6b1699cb39bcd59702e3b4bc4cec9aa
  • Hin-Tak Leung
    Hin-Tak Leung Posts: 361
    edited April 2016
    I mentioned earlier in this thread, that I registered a proposed talk about the Font Validator in 2016 Libregraphics in London, 15-18 April (i.e. Weekend two weeks from now). The proposal was accepted, and so people who want to catch my talk, or etc, please do in two weeks' time.

    I probably will only be there for one or maybe two days. Since I live within drivable distance from the venue, I do not need to decide now. LG is well worth going to for many of the interesting topics/projects.

Categories