Closed eternal-sorrow closed 2 years ago
Ty for the heads up, will investigate.
For my own reference, http://www.latofonts.com/2014/02/27/lato-2-0-released/.
Lato 2.0 was tested a while back but rejected due to bad hinting that caused glitchy rendering on Windows. However, the hinting bug has since been fixed upstream: http://www.latofonts.com/2014/09/03/lato-family-updated-version-2-010/
@twardoch has promised to publish the sources on Github eventually :)
Yes, I will. But they're in a VERY chaotic state right now due to the complexity of the project. Mixture of FontLab Studio 5, manual stuff, hardcoded Python snd some unpublishable 3rd party tools. I'm the only person who can build them right now but I definitely plan to upgrade them to a more portable form this summer.
Is there any news on this issue?
Need cyrillic symbols of Lato so much! So, is there any progress?
Happy issue anniversary, any word on getting Lato 2 support?
up
Soon!
The 2.0 version was released in February, 2014 and was asked to update to it. It was 2.5 years ago. There are not so many good cyrillic fonts and Lato is a nice one to have.
What's the status of this, why can't this be done for such a long time?
Hey guys, anything new on this?
What's happening with this one, why it is so hard to do? Early 2016, soon in June 2016. It's almost December, the end of the year.
Do people in Google use another understanding of time? Lato 2.0 was released in February, 2014, almost 3 years ago. It's 1000 days ago. Is it so hard to finally dedicate a day or two and update the font?
@davelab6 Can you finally keep your promise and update this font at least in 2016?
cc @m4rc1e, @alexeiva can someone help him as this promises can take forever. ?
@random-stranger Do any of them owe you anything at all, so that you're entitled to make such exigences? What is stopping you from simply self-hosting the font if you need it so desperately?
@fitojb to be fair, some places require Google Fonts API to be used for their own reasons.
@random-stranger when I pushed this the hinting had a problem, so I need to verify the hinting from this vendor is good in the latest release, and I didn't do this. However, @m4rc1e may be able to check this for us.
I can't commit to a specific date.
I'll look into this today, if they pass our QA, I'll send a pr. However, it will take more time to get them served from the api.
As @fitojb says, self hosting is not a bad option for the time being.
Cheers, Marc
I haven't heard any reports of Lato 2.015 failing on screen rendering, but I did hear of printing problems, even with the most current version, which most likely are related to hinting and ttfautohint
— but it's next to impossible to debug those problems.
Alright,
We have some minor issues and 2 relatively serious ones.
All Weights
Some weights
This will cause clipping on glyphs which ascend past 1974 units, same with the descending glyphs. I Recently found this issue on Overpass. However, the metrics do match the previous release. This means it existed in the previous version which made it into fonts.google.com
Family needs splitting, due to api font name constraints. Thin and Heavy will need their own separate families.
Licensing metadata does not match our spec. OFL and copyright strings... the usual business.
Font names do not match our spec as well. Semibold should be SemiBold.
@davelab6 @twardoch What do you folks think, shall we make the +1 crew happy and push it with these errors? I'm certain many fonts in our collection feature such issues. If not, I can only hot fix them in the binaries, since I don't have sources. I really dislike doing this because we are deviating away from the source.
In the future, if we decide to do an upgrade/language extension for this project, we can fix the issues and convert the sources to .glyphs. For the time being, we could live with the errors, if the authors are happy with this approach and yourself @davelab6.
I will personally do the font splitting so it will work on our api.
ps hinting in Windows looked great. Here's a screenshot of Win 7 Chrome.
Since current version has this issues and users are perfectly fine with them, it would be great to have updated version with them as for 99% of users having more families support is the best wish with this update.
On 21 November 2016 at 07:39, Marc Foley notifications@github.com wrote:
Alright,
We have some minor issues and 2 relatively serious ones.
Thanks for looking into this Marc!
Lato is like Ubuntu in that its a 'special' family - ultra popular but without real sources or a reliably automated build process. Adam is the magician here :)
Serious FB https://github.com/googlefonts/fontbakery report for v2 binaries http://www.latofonts.com/lato-free-fonts/#download
All Weights
- "ERROR: Glyph "uni200C" http://www.fileformat.info/info/unicode/char/200c/index.htm has ink. It needs to be replaced by an empty glyph", "ERROR: Glyph "uni200D" http://unicode.org/cldr/utility/character.jsp?a=200D has ink. It needs to be replaced by an empty glyph" Lato V2 has: [image: screen shot 2016-11-21 at 08 11 11] https://cloud.githubusercontent.com/assets/7525512/20473606/1e912fd0-afc2-11e6-9c38-a8a625b31d9b.png
That seems bad, but you can just drop the glyphs using pyftsubset (or even just drop the link from the GLYF table with ttx) to hotfix this :)
- https://cloud.githubusercontent.com/assets/7525512/20473606/1e912fd0-afc2-11e6-9c38-a8a625b31d9b.png
Some weights
- "ERROR: OS/2 usWinAscent & usWinDescent Changes that must be applied to this font: OS/2 usWinAscent from 1974 to 2109 | OS/2 usWinDescent from 426 to 527".
This will cause clipping on glyphs which ascend past 1974 units, same with the descending glyphs. I Recently found this issue on Overpass https://github.com/RedHatBrand/Overpass/issues/33. However, the metrics do match the previous release. This means it existed in the previous version which made it into fonts.google.com
If we move this to the "khaled" v metrics schema, where will we see reflow? Everywhere or only on a subset of platforms (GDI?)
Minor
-
Family needs splitting, due to api font name constraints. Thin and Heavy will need their own separate families.
I am not sure about this. In http://www.latofonts.com/download/Lato2OFL.zip there are only 18 font files, so they should be fitting within a single GF API family.
Where the style names are out of compliance, they can be hotfixed
- Licensing metadata does not match our spec https://github.com/googlefonts/gf-docs/blob/master/ProjectChecklist.md#ofltxt. OFL and copyright strings... the usual business.
Yeah perhaps this is a good moment to address that since it can be hotfixed
- Font names do not match our spec as well. Semibold should be SemiBold.
Ditto
@davelab6 https://github.com/davelab6 @twardoch https://github.com/twardoch What do you folks think, shall we make the +1 crew happy and push it with these errors? I'm certain many fonts in our collection feature such issues. If not, I can only hot fix them in the binaries, since I don't have sources. I really dislike doing this because we are deviating away from the source.
Knowing your proclivity for automated builds, you would dislike dealing with the sources even more ;)
In the future, if we decide to do an upgrade/language extension for this project. We can fix the issues and convert the sources to .glyphs. For the time being, we could live with the errors, if the authors are happy with this approach and yourself @davelab6 https://github.com/davelab6.
Actually the glyph set of Lato 2.0 is super extensive, so this could be considered more or less a 'final' release
I will personally do the font splitting so it will work on our api.
ps hinting in Windows looked great. Here's a screenshot of Win 7 Chrome.
[image: screen shot 2016-11-21 at 08 34 56] https://cloud.githubusercontent.com/assets/7525512/20474235/ef284e46-afc5-11e6-9cb1-d3e0bf07982f.png
I suggest you prep hotfixes for Adam to approve and we ship this 'soon' :)
Cheers Dave
@davelab6
Here's my approach to kick this out the door:
Since Heavy and Hairline weights exist, they will become their own separate families. Our api only support fonts from Thin–Black
The win asc and desc doesn't bother me for the time being. The fonts already use Khaled's approach, only the win asc and desc are bad. I don't want to tinker with these too much in FontTools. This already existed in the previous versions as well
I'll change the name Semibold to SemiBold
I think this is a nice temporary solution. It also means I won't spend longer than a few hours getting this done. If we're all happy with this approach, I'll submit the pr tomorrow :-)
Cheers, Marc
I don't think the fonts need to become separate families, since there are 9 weights in Lato 2.015 (specifically done so at Dave's request, the old Lato had a slightly different weight progression), so they map well to font-weight 100 to 900. The internal style names have no meaning in webfonts. The webfont kit for self-hosting has no problem with them being in one family.
The old Lato had 5 weights, and Lato 2.0 has 9 weights, but the design space positions of the previously released 5 weights has been kept (which is why the new Lato Medium is not much bolder than the regular, because both Medium and Semibold needed to fit between the old Regular and the old Bold which were not very far apart in weight).
I'd very much prefer if the 9 weights remain part of one family. But I realize that the way old Lato has been deployed on GF is that the Thin is 100, while in Lato 2, Hairline is 100 and Thin is 200.
Also, Lato 2.0 has a bit tighter spacing than Lato 1.0, because we used iKern for the spacing and kerning. But tighter is better than looser, because this won't cause overset text.
The design changes regarding the two glyphs will need to be tabled for our side. We do have a few issues on our todo list already but they keep sitting there, especially since in September, Łukasz has become a father to twins, and he's currently on paternal leave from work (mostly).
And yes, the build process is currently complicated and semi-manual. I've been hoping to upgrade it to a more automated process but it has been unobvious which one. fontmake
is the most promising so I started experimenting with it. But due to the glyphset size, we're currently using some non-opensource tools in the middle of the process and I still haven't found good replacements. :/
Ps. In short, Lato 2 is supposed to be a good replacement for Lato 1, vertical metrics and weights have been kept but the spacing (sidebearings and kerning) is tighter throughout the family, so there is no guarantee for full backwards compatibility.
On Nov 21, 2016 8:12 AM, "Marc Foley" notifications@github.com wrote:
Since Heavy and Hairline weights exist, they will become their own separate families. Our api only support fonts from Thin–Black
Heavy and Hairline must be renamed to Black and Thin.
The win asc and desc doesn't bother me for the time being. The fonts already use Khaled's approach, only the win asc and desc are bad. I don't want to tinker with these too much in FontTools. This already existed in the previous versions as well
Just fix them
I'll change the name Semibold to SemiBold
Yep, careful of osx though, you need to rename from x to x1 to X, as renaming from x to X actually nothing and git goes millennial, err, mental. Autocorrect 🙈
I think this is a nice temporary solution. It also means I won't spend longer than a few hours getting this done. If we're all happy with this approach, I'll submit the pr tomorrow :-)
I think the above 3 actions don't take very long either 🙊
Cheers, Marc
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.
On Nov 21, 2016 12:56 PM, "Adam Twardoch" notifications@github.com wrote:
I don't think the fonts need to become separate families, since there are 9 weights in Lato 2.015 (specifically done so at Dave's request, the old Lato had a slightly different weight progression), so they map well to font-weight 100 to 900. The internal style names have no meaning in webfonts. The webfont kit for self-hosting has no problem with them being in one family.
Agree
The old Lato had 5 weights, and Lato 2.0 has 9 weights, but the design space positions of the previously released 5 weights has been kept (which is why the new Lato Medium is not much bolder than the regular, because both Medium and Semibold needed to fit between the old Regular and the old Bold which were not very far apart in weight).
Agree, and variable fonts will sort this out very soon so I'm not worried at all about it
I'd very much prefer if the 9 weights remain part of one family. But I realize that the way old Lato has been deployed on GF is that the Thin is 100, while in Lato 2, Hairline is 100 and Thin is 200.
I'm ok with this
Also, Lato 2.0 has a bit tighter spacing than Lato 1.0, because we used iKern for the spacing and kerning. But tighter is better than looser, because this won't cause overset text.
I'm also ok with this, and indeed there will be quite a bit of tighter and looser spacing changes coming across the collection as quality improves.
The design changes regarding the two glyphs will need to be tabled for our side.
Why? 😄
We do have a few issues on our todo list already but they keep sitting there, especially since in September, Łukasz has become a father to twins, and he's currently on paternal leave from work (mostly).
🐣🐣😄📣
And yes, the build process is currently complicated and semi-manual. I've been hoping to upgrade it to a more automated process but it has been unobvious which one. fontmake is the most promising so I started experimenting with it. But due to the glyphset size, we're currently using some non-opensource tools in the middle of the process and I still haven't found good replacements. :/
I hope you can speak with the Noto team about this as they also have some large glyph sets and if Python based fonttools can't cut it, that's important feedback :)
@davelav6 Ok cool.
I may as well write a script so we can do this for other families as well. I'll make sure it's all done for FB.
Cheers, Marc
Hey folks,
Apologies for the lack of updates on this. I needed to write/update some tools in order to hot fix these fonts. I can confirm the clipping issue and vertical metrics are solved. I adjusted the values but made sure the appearance matches the version which is already hosted.
Win 7 Chrome
OSX Safari
OSX Chrome
Since I have all the tooling in place, issues and releases such as #10 should happen much quicker. Thank you all for your patience. I still need to submit a few prs to tools I have used, before I can upload the hotfix repo.
@twardoch I'll send you a .zip of the fonts.
Cheers, Marc
Right, I've uploaded the hot fix repo. Once the authors are happy, I'll submit a pr.
@m4rc1e Do you know why Lato-Light (left) vs. Lato-300-normal-outline (right) has text reflow?
Or — is the left one the new Lato 2.0 while on the right it's old Lato 1.0? (In that case the reflow is justified because Lato 2.0 runs a bit tighter).
@twardoch yep the right is what is live in the Google Fonts API at the moment
OK, go ahead and publish it then.
Thanks Adam!
Love you guys! Thank you for taking the time to update Lato. It means a lato :) Jokes aside, it seems the changes were merged into google / fonts in this PR. Any idea on when we'll be getting the updated version live on fonts.google.com? Again, thank you!
Any update on this, when we'll be happy to use the updated version? 🎉
cc @davelab6 , @m4rc1e
Sadly I can't say, as its not ready until its ready, but, I hope soon! :)
@davelab6 hey, any news on this?
Is it possible to be released before the end of the year, seems like everything was fixed?
It will be updated in the API soon, but will not be in December
On Dec 24, 2016 9:02 AM, "Vlad Shcherbin" notifications@github.com wrote:
@davelab6 https://github.com/davelab6 hey, any news on this?
Is it possible to be released before the end of the year, seems like everything was fixed?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/google/fonts/issues/6#issuecomment-269074178, or mute the thread https://github.com/notifications/unsubscribe-auth/AAP9y_5YaWuUIPMSBjO1uABMPb6tjXbBks5rLNGbgaJpZM4D9oBR .
Finding information about when google fonts release Lato 2, and visited this. Waiting this great font to be released in the API.
@davelab6 Will it ever be released ?
@m4rc1e fixed everything 2 months ago, what are we waiting now? I'm not sure how bad your deploy process is, but this attitude to the people, asking for this font forever, is really bad.
@davelab6 Hey Men, we need this font on a Google Fonts! Please do it ASAP
@m4rc1e was https://twitter.com/janakatino/status/821513539919679488 resolved in https://github.com/google/fonts/pull/479 ?
Hey guys, any news on this?
Shame about all the problems since the rollout, hopefully this can get rolled out again soon.
Waiting on @twardoch
Hey, @twardoch / @davelab6, I was wondering if there had been any movement on this recently or if there is an ETA? Thanks!
Adam, any news?
On Apr 4, 2017 4:17 AM, "Will Blackmore" notifications@github.com wrote:
Hey, @twardoch https://github.com/twardoch / @davelab6 https://github.com/davelab6, I was wondering if there had been any movement on this recently or if there is an ETA? Thanks!
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/google/fonts/issues/6#issuecomment-291427527, or mute the thread https://github.com/notifications/unsubscribe-auth/AAP9y5eWESLDcJRQqgZfWxA4_MWfyuAsks5rsfyCgaJpZM4D9oBR .
Not yet. But things look promising. Łukasz got back to work after being exclusively committed dad to his newborn twins for some 8 months. There are a handful of glyphs that need tweaking and I think he'll be able to do it soon, then I can make the update with the newest ttfautohint et al. Maybe I can even scrap a big part of my awkward production workflow in favor of fontmake
And another 4 months have passed ...
@Renkas this is normal for google fonts. When you have to wait for an important fix like this one for half a year (and it's still not fixed), it's no surprise Lato 2.0 is not added for years.
Expect this update in 2020 if you are lucky.
Oh man, I'm waiting to this font to be fixed like 3 years. Google ain't an agile company.
Since version 2.0 Lato has started providing cyrillic symbols along with ones of some other languages. So, please add a new version of this nice font.
(Note accents not working per http://jsbin.com/momova/edit?html,css,output)