@PauloCereda -- why, yes -- i am pleased to announce that the amsmath users guide has been significantly overhauled, and i've received notice that it has been uploaded to ctan and will be in the next edition of tex live. (how's that for "camouflage"?)
@DavidCarlisle yesterday someone mentioned that the Oxford Dictionary (?) changed recently a lot of hyphenations. Do you know something about it? (There is a question on c.t.t. about it and I wonder if it is related)
@DavidCarlisle just found a discussion on the hyphenation mailing list about it. Something did change ..
@JosephWright yesterday at Passau there was a talk of Oliver Kopp who uses CircleCi and docker images of texlive to run checks on tex code from github (he is a very fast clicker and speaker so don't ask me about the details ...).
@UlrikeFischer Depends partly if you want a fixed image or not
@UlrikeFischer I'm not 100%, but my impression is that this approach means you have 'provided' images rather than something you select yourself. With caching, it makes little odds how you get TL: the tests are what takes the time.
@UlrikeFischer I'm not tied to any particular approach ...
@JosephWright he seems to have setup a system to store docker images of different states of his texlive and then can decide which one to use for a test. (The docker images are store somewhere else I have forgotten how the site is named).
@JosephWright I'm not trying to convince you to do something differently, I know nothing about this stuff but at least understood that is it similar to the travisCI setup you are using.
@UlrikeFischer no, travis provides a separate mechanism to cache stuff so we are using a cached texlive install, so it's a different technology but comes to the same thing
@DavidCarlisle It was quite interesting to see it (even if the talk was a bit full as he not only shows the circleCI stuff but also pull requests on github and creation of docker images and setting up of ssh keys and ...).
@PauloCereda ask @Maïeul to post it on the french groups.
@samcarter Ditto on the apologies for the long delay ... too busy eating roasted birds (not duck, of course). On "official" stuff then I tend to use my real name - this identity is pseudonymous not anonymous. texdoc tikzmark gives the game away if you're not sure.
@ChristianHupfer @Skillmon It's easy: Blame @JosephWright :) Though I believe that it's not difficult to code yourself. I think if you ask on the site, @egreg will think something out :)
@marmot your scroll looks OK, but the text on a Tex scroll should be in the other orientation so you can write longer stuff and some unseen hand appears at random times and rips off one sheet of text.
@DavidCarlisle I got it this afternoon in tl (I was in the train until then and didn't want to strain the wifi) and my first document didn't explode ;-). (miktex hasn't is yet).
@UlrikeFischer I have no reason to doubt it;-) (In the neighboring park there are really lots of ducks and turtles, and they really live happily together. The marmots are a bit shy, though ;-)
@UlrikeFischer feel a bit bad about PL1 when ctan team are under pressure to get things out in time for texlive cutoff, but \UseRawInputEncoding is definitely better now, so....
@marmot Yes. I was just listening today to Radiolab and they had a whole show on the ability to make very plausible sounding made up audio attached to video: futureoffakenews.com/videos.html