Word Aligned - Latest Comments

Re: The Etch-A-Sketch User Interface

Kitty7oz — Mon, 30 Aug 2010 16:10:31 -0000

haha!

Re: A Subversion Pre-Commit Hook

Support — Sat, 28 Aug 2010 07:57:14 -0000

Why would you want to reject files with tabs? Its the only way to indent! We have scripts that take bad code (multiple spaces instead of tab) and correct it. Indent amount is a preference that should be set by each developer, not the developer that necessarily created the file. It annoys me when I need to hit backspace 8 times instead of only 2...move on, spaces are dead!

Re: One svnserve, multiple repositories

Guest — Fri, 27 Aug 2010 17:23:16 -0000

Since each repository was created separately,
> $ svnadmin create red_bear
> $ svnadmin create blue_goat
each repository has its own set of version numbers that are independent of the other as well as being independent of yellow_dog.

A commit to one repository will therefore ONLY affect the version numbers in that repository, and not in any other!

Re: One svnserve, multiple repositories

Cwilde — Fri, 27 Aug 2010 17:10:14 -0000

Helpul post.... thank you!

Re: Converting integer literals in C++ and Python

Punctuation_09 — Tue, 10 Aug 2010 03:26:11 -0000

i already used this codings but it seems that it is hard to understand

Re: One svnserve, multiple repositories

Bokaal — Tue, 03 Aug 2010 10:07:20 -0000

thanks for the info! works perfectly

Re: One svnserve, multiple repositories

Neofish — Thu, 22 Jul 2010 14:38:47 -0000

Nice article, whilst I knew most of this already, you wouldn't believe the trouble caused by lack of up-to-date useful documentation. There are unfortunately too many places (not just config files) that this kind of thing can go wrong, and if you get stuck it can be nigh impossible to fix it. I assumed my config was correct due to the examples, and plowed on further only to find out that it wasn't, and I've had to completely restart XD

Re: One svnserve, multiple repositories

Lviggiani — Wed, 14 Jul 2010 06:24:51 -0000

Hi, I have a doubt (I didn't try so far):
Once having created the two repos, are the two repos versions handled separately?
I mean, at the beginning both blue_goat and red_bear are at version number 1.
Let's imagine I update a file on red_bear. What happen?
Both go to version 2 or blue_goat stais at version 1?

Re: Hiding iterator boilerplate behind a Boost facade

wheaties — Sun, 11 Jul 2010 09:13:46 -0000

Quick comment, on your "missing methods" Python implementation: x < y is not the same as !( y > x ) because you don't check for the possibility of x == y.

Re: Hiding iterator boilerplate behind a Boost facade

Giuseppe — Thu, 08 Jul 2010 05:53:40 -0000

Yes, the tag name misled me too :) The iterator_facade documentation is not the best in Boost...

Re: Hiding iterator boilerplate behind a Boost facade

tag — Thu, 08 Jul 2010 02:47:47 -0000

Thanks Marius. You are quite right: the class passed in to the decorator function is modified directly and is also returned. I have tweaked the wording.

Re: Hiding iterator boilerplate behind a Boost facade

openid-28622 — Wed, 07 Jul 2010 20:42:48 -0000

Nice article. Yes, I often feel that boost documentation could use some work. It seems to have gotten better over time though.

Re: Hiding iterator boilerplate behind a Boost facade

openid-11409 — Wed, 07 Jul 2010 20:12:52 -0000

"total_ordering(X) returns a new version of the class with the missing operators filled-in"

This may give the reader an incorrect impression that the old version of the class is still around, and doesn't have the missing operators. What actually happens is that the class always is modified in-place and then returned.

Re: Hiding iterator boilerplate behind a Boost facade

tag — Wed, 07 Jul 2010 17:37:12 -0000

Many thanks for putting me straight, Giuseppe: you are quite right. The "random access" in "boost::random_access_traversal_tag" misled me. I guess I should have read the reference documentation and indeed the source code a little more carefully. I've corrected the code in SVN, and added a note to the article.

I also like your suggestions on how best to use ifstreams.

Thanks again.

Re: Hiding iterator boilerplate behind a Boost facade

Giuseppe — Wed, 07 Jul 2010 17:02:44 -0000

A quick check reveals that distance_to is never called, which means that binary_search is using the (slow) linear std::distance. This is because boost::random_access_traversal_tag is not a random access iterator, but a "new style iterator" which has no O(1) distance.
Changing the category in iterator_facade parameters to std::random_access_iter_tag gives the same performance as the iterator implementation.

About the "mutable", iterators are value types, so they are often copied. They should be as lightweight as possible. Having an ifstream as a member, requiring to reopen the file on every copy of the iterator (because the ifstream is noncopyable), is IMHO not a very good idea. Would be probably better to have the ifstream object instantiated outside of the iterator, and keep a pointer to it in the iterator object, so the state of the iterator would be (std::ifstream* in, std::streamoff pos), and you would have to do the seek at every read operation. ifstreams don't have a very nice interface for random access...

Re: Hiding iterator boilerplate behind a Boost facade

Andrew Fray — Wed, 07 Jul 2010 14:29:08 -0000

/const/ methods are there for the caller, not the method body. They demonstrate that no matter how many times you call them, their behaviour won't change. This is logical constness. Often your methods are bit-const, so no memory changes, as well as logical. Hoewever this is just a coincidence.

/mutable/ allows you to break bit constness to keep logical constness, and I don't think this makes it a hack or a cheat, or that you should be uncomfortable using it.

Re: Keyword Substitution - Just say No!

Sean Healey — Wed, 07 Jul 2010 12:01:21 -0000

I know I am commenting on a really old thread, but I only just found it and the point I am about to make might be useful to others.

We have a large developer base who maintain SQL scripts which contain definitions of among other things, database stored procedure code. It is immensely valuable for a production support bod to be able to examine source code on the dataserver and immediately know from the header block who modified the current version and which change request the modification relates to.

We have traditionally manually maintained header blocks in these scripts, but frustratingly people are not discplined about this. It is very powerful to have the source control system enforce the basics of the audit trail using keyword replacement in this manner.

I grant you, although ideally suited to this scenario, it probably should not be used in others.

Re: A Subversion Pre-Commit Hook

Waseem Bokhari — Mon, 05 Jul 2010 01:26:57 -0000

Good :)

Re: A Subversion Pre-Commit Hook

Liuzengyou — Mon, 05 Jul 2010 00:52:47 -0000

你这个 prevents commit of certain file extensions such as .exe, .jar, .tar, etc.
脚本在哪呢能否用一下 thanks

Re: Bike charts by Google

floater81 — Fri, 02 Jul 2010 12:29:33 -0000

The available google charts (line, bar etc.) didn't offer what I needed so I wrote my own. Very satisfying!
http://code.google.com/p/gwtstationmodel/

Re: Bike charts by Google

tag — Wed, 30 Jun 2010 11:49:09 -0000

Yes, I submitted those charts. Thanks for the timely reminder. This year's tour is shaping up nicely. Roll on the prologue!

Re: Bike charts by Google

floater81 — Wed, 30 Jun 2010 11:17:33 -0000

I read this post back when you wrote it but today I was looking at the chart API and recognised these at the top of the user submitted gallery...
http://code.google.com/apis/chart/docs/user_submitted_gallery.html

Re: Essential Python Reading List

jimmy choo — Thu, 24 Jun 2010 02:12:52 -0000

http://blog.objectmentor.com/articles/2009/11/21/

Re: Ignoring .svn directories

Anonymouse — Sat, 19 Jun 2010 22:29:32 -0000

I get too lazy to type print, so I use:
find . -not $ -name .svn -prune $ -type f

But I like James' suggestion which uses -print0 and -0 to support filenames with spaces.

Re: A Subversion Pre-Commit Hook

Justin — Sat, 19 Jun 2010 06:05:36 -0000

Excellent. Thank you. Just what I was looking for.