Comments by Dustin Sallings

I think enough people do that on accident, and it's a hard enough thing to keep up with that it's a good default at least. It'd be trivial to make a setting that disables that warning for you.

Reply | Original | Permalink | Tweet

> In git, you have to name every branch you make (if you want to get to it later). What if I don't want to name a branch I'm going to merge in ten minutes later? Why do you think you have to name your branches? One really common thing I've seen with new git users is that they create unnamed branches and don't understand why git doesn't know what they mean when they tell it to push them. > You can just commit when/where you want and use hg heads to find the open branches. That's a trade-off. I can find all heads in git, but I have a *lot* of dead ancestry in practice.

Reply | Original | Permalink | Tweet

I'd recommend reading reddit programming some. Preferably before you post since this question is asked weekly.

Reply | Original | Permalink | Tweet

Because it is more than just better performance, and it's fewer than 10 bytes on the output. Little known fact: Most sites aren't google. I've found having valid xhtml makes it far easier to get predictable results. Quite frequently, when someone is having trouble with rendering on one browser or another, I suggest first making the document valid and that makes it far easier to understand, thus express your intention I understand this is about html4, but the same thing still applies. Suggesting that people should break the symmetry of their documents -- making them harder to understand and parse -- for the sake of *optimization* is one of the first things we learned in computer science. Premature optimization is evil. Admittedly, I don't know all of the html 4 rules for what's optional, what's nestable, etc... xhtml is far more simple -- there are block level elements and inline elements. The rules as to which elements can fall within others is pretty intuitive, but easy enough to reference. Here's what I do know, though: It's really easy to tell everyone to always close what they open. Always use quotes for attribute values. Have templates that do nesting in natural ways. Once you get really simple rules in place, **I can easily write an output filter that performs the optimization.** This is the same thing that happens with gzip (unless you're suggesting people also write their HTML snippets pre-gzip encoded). Here's a bit of advice for everyone other than google writing code out there: **You are not google.**

Reply | Original | Permalink | Tweet

Does it really save that much once you compress stuff? I realize once you get a billion hits a day, a couple bytes is in the GB range, but you pretty much have to have perfected everything else before you get to that point.

Reply | Original | Permalink | Tweet

I wash my dishes incrementally while my tea water heats up. More dishes are done for the first cup than for the second. I didn't drink enough tea this weekend, so I had to block for a spoon before breakfast this morning. :(

Reply | Original | Permalink | Tweet

You certainly can commit from gitx. Look at the little split button on the top left. It changes you from view mode to commit mode.

Reply | Original | Permalink | Tweet

Note that this should also work: cat file | cat | cat | cat | cat | sort -R | head -1

Reply | Original | Permalink | Tweet

That is kind of an awful changeset, though. If you find yourself making a list of things that you've done as wide as that one is, you're definitely doing it wrong. It's really hard to follow projects that do giant code drops.

Reply | Original | Permalink | Tweet

Sure, it's been supported for years. First clue, though, should be that svn "changeset" IDs don't look like this: 77e51a304d1b4034614d75c5bf4c07b216400a42 They also don't have a `git-committer`.

Reply | Original | Permalink | Tweet

appjet used to be a javascript-based web host, but it seems they're in the process of shutting it down. :(

Reply | Original | Permalink | Tweet

How does one get "svn diff" out of this?

Reply | Original | Permalink | Tweet

C is not out of date, it's just less of a general purpose tool. We have higher level options now.

Reply | Original | Permalink | Tweet

I agree with a lot of that, but comments are for either documentation (javadoc, doxygen, etc...) or to make excuses for bad code. I find reading and understanding code easier than English. The code doesn't lie and has no ambiguity. Pieces of it can be replaced or modified as I need. I rarely read any comments when I'm working in code because they've mislead me too often.

Reply | Original | Permalink | Tweet

I never close emacs. `emacsclient -n` is about the same thing as `mate` and will just shove stuff into my existing frame. I've got some customizations over midnight mode to make sure my list of open files remains small enough.

Reply | Original | Permalink | Tweet

1. « Previous
2. Page:
3. 1
4. 2
5. 3
6. 4
7. ...
8. 11
9. Next »