« first day (1402 days earlier)   

12:39 AM
I remember the last time I asked about annotating a LaTeX file, per tex.stackexchange.com/q/169808/3406 for example, someone made the distinction about writing on top of the TeX (which means it might overwrite stuff), or putting it in a "provided" slot, say at the top of the page. I don't remember the details. Does anyone have a link to something about this?
@PauloCereda Exploding compilers? What do you people talk about here.
 
 
6 hours later…
6:12 AM
@PauloCereda No idea. We should ask Roberto for insights..
 
6:31 AM
In my experience using TeX Live Manager (any version), I think the background process blocks the GUI thread so the GUI is not responsive. Hopefully my conjecture will be proven soon or later.
 
 
2 hours later…
8:57 AM
@cyanide-basedfood does texlive manager have a gui? :-) I've only ever used tlmgr on the commandline
 
@DavidCarlisle obviously, but I get:
[tohecz@toheshiba ~]$ tlmgr --gui
...
Goodbye.
:D
 
9:15 AM
Damn, I can't decide between a bad notation and another bad notation.
 
@tohecz Make up your own notation ;-)
 
@Johannes_B that's what I did, but now it coincides with the derivative (not that I use the derivative, but still it's rather confusing)
 
@tohecz Using the prime with other meanings is common.
 
@egreg yep, but not with $P'(w)$, where P is moreover a polynomial
 
@tohecz I agree, in that case it's not good. Why not sharp or flat?
 
9:24 AM
@egreg but the problem is that prime is the way how you denote the non-id automorphism of a quadratic field
So when $x=P(w)$, I want to have that $x'=P'(w)$
 
@tohecz It's a conjugation: $x^*=P^*(w)$.
@tohecz Or, as I suggested, \sharp or \flat
 
@egreg yep I know, and the problem is that all people around numeration systems use \prime
 
@tohecz So use it as well! ;-) Just make sure that the symbol is explained.
 
@egreg but then my superior says that it's confusing, and he's right
it is well-explained, and it is all around the article (on more than half of the pages)
 
I don't feel happy with this answers. tex.stackexchange.com/a/199062/37907 It is smurf.
 
9:28 AM
@Johannes_B It is crap low quality. Do you plan to comment on it?
 
@tohecz I don't know what to say. Extendeing the answers would still not add anything useful. I don't think that the OP is looking for geometry.
@tohecz @egreg just left a comment.
 
@Johannes_B I see
well, I was tempted to say something like: There's no good answers to unclear questions.
 
@egreg youtube.com/watch?v=zY7UK-6aaNA Look at 3:41 please
 
9:44 AM
Another LaTeX3 doubt that has been around my mind for a while, but today I thought about it a bit more. All primitives have the form of \(lua|e|pdf)tex_primitive:D, which is nice, because one knows all the primitives are there, always. And they have an “interface” (not all yet, but in the future): \tl_to_lowercase:n <= \tex_lowercase:D, \cs_set:Npn <= \tex_def:D, \int_eval:n <= \etex_dimexpr:D, etc.
 
@Johannes_B They even turn the page! It's so silly! Just like something I saw at the Tate Modern gallery: a canvas painted in a uniform color. The explanation said it was the first in a series of works (with different colors, of course).
 
However… many of them are inside a macro so the \…_primitive:D ends up being used at the very end of expansion, but others are just \cs_set_eq:NN \cs_set:Npn \tex_def:D which means that \tex_def:D is never, ever, being used.
Shouldn't it be that all the primitives are being used at the very end? That way one would absolutely know what are the primitives. In my example, it happens that \cs_set:Npn is a primitive, but no one knows and may think that the one that ends up being used is \tex_def:D.
 
@Manuel Historically the plan was essentially never to use the primitives 'directly' but always to rename (so we had e.g. \pref_long:D = \tex_long:D)
@Manuel That was to be frank a bit of a pain in the neck
 
@Manuel With \let\x\def, a \write\stream{\x} command would write \x and not \def.
 
@Manuel Currently, the idea is that if the team provide a documented interface that achieves the task in hand it can be used if used 'correctly'
 
9:51 AM
@Manuel With \def\x{\def}, \def would be written.
 
@Manuel The :D level is really meant for the team, in the sense that if we are doing our job correctly then only we and people with a real interest need to know the detail
 
@egreg That's something I still don't know about :( But I see that's a reason not to use a \def instead of a \def.
 
@Manuel Currently things are not quite 'perfect' in the sense that over time the conventions have changed a bit
 
@JosephWright Your last message is absolutely true, I was just guessing.
\cs_new_protected:Npn \cs_set_eq:NN #1 { \tex_let:D #1 =~ }
\tex_let:D \cs_set_nopar:Npn \tex_def:D
 
@Manuel Note that not all of the primitives have a direct interface, and as I say some of them have evolved a bit over time. For example, I'd probably not have gone for \__int_eval:w, which could be covered by \etex_numexpr:D
 
9:55 AM
I mean, \cs_set_eq:NN is inside a \def just because one needed the =~, and I just see that it's a different thing than the definition of \cs_set_nopar:Npn (which is just \let).
 
@JosephWright I disagree. With \__int_eval:w you are consistent with the general approach.
 
En route to SP. You guys behave while I am away. :)
 
@egreg Like I say, I'd have gone for something different :-) A few years ago I did a 'small scale' rewrite of expl3 where I did stick to the :D versions as far as possible, i.e. only using defined interfaces when they 'added' something. Those experiments gave us some useful ideas but not everyone agreed with everything I tried :-)
@egreg That's broadly Frank's argument
@egreg The issue comes where you don't use the primitive more than once or twice or only in a very constrained situation
@Manuel Yes
@Manuel Like I said, the idea is that really only the team should need to read all of the code part
 
@JosephWright I didn't mean to stick to the :D versions, but to use a general \def\x{\tex_primitive:D} instead of mixing with \let\x\tex_primitive:D.
 
@Manuel If we are getting it right then interface3 should describe correctly how to use expl3
@Manuel Not as simple as that
 
10:00 AM
@JosephWright Okey, okey. That's true. Thanks for the comments :)
 
@Manuel For example, \exp_not:N or \exp_after:wN have to work 'as is' with no expansion
 
@JosephWright Ahá, I did not think about it. By the way, just another thing I thought, why \exp_after:wN and no \cs_set_eq:Nw?
 
@egreg The issue comes less with the 'core' stuff but more with specialist primitives. For example, the galley needs quite a number of primitives that don't appear in any other module. For those, after a bit of discussion we decided to stick with the :D names.
 
@Manuel That's \token_new:Nn
 
The line tends to be that if it's used only in one module :D is OK, but the moment it pops up 'here and there' we should give it some internal name
 
10:04 AM
@JosephWright Reasonable.
 
@Manuel That is a slightly tricky one
 
@egreg :) Thanks for the fast question solving here!
 
@Manuel Historically, several of the 'low level' commands had the wrong name (\exp_after:NN, for example)
@Manuel I've wondered about that myself, I have to say
@Manuel You are likely right, but selling that change might be tricky :-)
@Manuel As we provide \c_group_begin_token and \c_group_end_token I'd imagine that in use #2 should always be N-type
@egreg That one is definitely wrong: better fix it!
 
@Manuel \token_new:Nn is actually an alias to \cs_new_eq:NN; but the advantage is being able to change the implementation at any time.
@JosephWright Of course it should be \token_new:NN
@JosephWright Or somebody could try \token_new:Nn \my_space_token {~}
 
@egreg And also include the =~ inside the definition. Then \token_new:NN \c_space_token {~}.
 
10:10 AM
@egreg Yes
 
@JosephWright It would work, actually. But it's not consistent with the naming scheme.
 
@egreg I know: I am one of the people who is expected to know the detail here :-)
 
@JosephWright On the other hand, one can do \cs_set_eq:NN \my_space_token \c_space_token; a \token_set_eq:NN function is missing.
This seems a duplicate (see my comment)
0
Q: Apply macro to .tex file and get a new .tex file

Dennis YurichevAre there a preprocessor for TeX, like for C/C++? Are there a way to apply a macro to .tex file in order to generate new file? For example, I have macro: \renewcommand{asd}{some phrase here} I have a .tex file: bla-bla \asd bla-bla I would like to get a new .tex file where all macros are r...

 
10:58 AM
@egreg I've raised this with the team
@egreg I'm not quite sure \token_... things make all that much sense, as the first token has to be either a control sequence or active char
 
Good MAEN...
 
11:14 AM
@ChristianHupfer Hey. The latest issue of »Me question is smurf the smurf« is out there. golatex.de/problem-mit-twoside-t13765.html
 
@Johannes_B: Und nun wieder eine Folge "Schweine im Weltall" ;-)
@Johannes_B: Lol ... ich kann meine Socke nicht finden :D :D
 
@ChristianHupfer Exactly.
@ChristianHupfer I had a friend sleep over a few days ego, she said the exact same thing in the morning. ;-)
 
11:29 AM
@JosephWright It all depends on what the token data type is meant to be. Possibly \chartoken_... would be more meaningful, as we don't really want that these functions are used for symbolic tokens.
 
@Johannes_B: Sorry, I have to be offline for some hours. See you later
 
11:54 AM
@DavidCarlisle Beat me by a minute. I had to upvote you. Damn!
 
12:09 PM
@egreg excellent news.
@egreg oh but I see you still posted to try to deprive me of the 15
 
12:27 PM
Do we have a canonical question concerning starred chapters and headers?
 
@egreg Damn, $x^*$ doesn't work for me because * is the Kleene star :-/
 
1:09 PM
@egreg oh shame, you didn't get the tick :(
 
@DavidCarlisle I had to do it!
 

« first day (1402 days earlier)