Rob Earhart's WebLog - All Comments

Rob Earhart s WebLog SEH stack walking | Quick Diets

Rob Earhart s WebLog SEH stack walking | Quick Diets — Sat, 13 Jun 2009 16:19:18 GMT

PingBack from http://quickdietsite.info/story.php?id=14022

re: Threadpool

Ramanujam — Tue, 15 Feb 2005 14:11:00 GMT

wow that's great..........to hear...
even we are working on kernel 32....in our IBM labs

re: Threadpool

Pavel Lebedinsky — Mon, 24 Jan 2005 12:12:00 GMT

Are the docs for the new threadpool publicly available?

Good luck fixing the loader lock mess. The current design is (in my opinion) just asking for problems:

1. Create a set of rules that forbid doing most non-trivial things in code that runs inside the loader lock.

2. Expose a very easy way for people to write such code without even realizing it (static C++ constructors/destructors).

3. Provide no mechanism to enforce the rules.

4. Make sure that even when the rules are violated, things still seem to work correctly 99% of the time.

5. Watch people fall into the same trap over and over again.

Really, how are people supposed to find out that (for example) declaring a global CComBSTR object is a big no-no, when it works just fine in the vast majority of cases?

It's Hungarian-esque

Sorting It All Out — Thu, 20 Jan 2005 11:31:00 GMT

It's Hungarian-esque

Sorting It All Out — Mon, 10 Jan 2005 00:53:00 GMT

re: Hungarian

Michael Grier [MSFT] — Thu, 06 Jan 2005 11:04:00 GMT

I think there is a useful subset of Hungarian.

rgElements - an array of elements
iElement - the index (0-based) of an element
cElements - the count of elements in a collection
rgpElements - an array of pointers to elements
prgElements - a pointer to the head of an array of elements
pElement - a pointer to a single element

specifically, the "i" and "c" prefixes are very useful IMO as a shorthand for "ElementCount" or "ElementIndex".

pElement vs. prgElements is useful to differentiate a pointer to a singleton vs. a pointer to a vector/array/"range" of elements (range is where rg comes from; I think it's obtuse but whatever.)

Further

pszName
rgchName

is a good distinction since the intent is clear that pszName is terminated with a null character where rgchName is just an array of characters which may or may not have a null character inside the array.

In Office, we had to assign a Hungarian prefix to every type which was rediculous. That's where you get some of the craziness.

re: Hungarian

Rob Earhart — Wed, 05 Jan 2005 20:51:00 GMT

It wouldn't surprise me if study's a hoax... the underlying idea remains, though--people seem to be able to handle a certain degree of scrambling, but that mechanism's probably not as useful with Hungarian. At least, that matches my intuition, having used Hungarian a bit.

I'd love to see some sort of "reading comprehension" study done on source code written in various languages and in various styles... it'd be hard to factor in people's existing language knowledge, but it might still be useful.

re: Hungarian

Uwe Keim — Wed, 05 Jan 2005 20:40:00 GMT

That study that your wife stated is a hoax.

This is a german magazine's article about that stuff:
http://www.heise.de/tp/r4/artikel/15/15701/1.html

This is the google-translation:
http://translate.google.com/translate?u=http%3A%2F%2Fwww.heise.de%2Ftp%2Fr4%2Fartikel%2F15%2F15701%2F1.html&langpair=de%7Cen&hl=de&ie=UTF-8&oe=UTF-8&prev=%2Flanguage_tools

re: Hungarian

Rick Schaut — Wed, 05 Jan 2005 20:34:00 GMT

Take a look at <a href=http://blogs.msdn.com/rick_schaut/archive/2004/02/14/73108.aspx this>. Wikipedia references the article <a href=http://en.wikipedia.org/wiki/Talk:Hungarian_notation here>.

Rick

re: Yet another reason not to handle all SEH exceptions

Seth McCarus — Mon, 03 Jan 2005 23:40:00 GMT

Rob, nice to have found your blog. SEH is simple on the surface but there are so many hidden caveats - I wish that MSDN or some such reference had a real user's guide to SEH, not just the functions and keywords, but how it should be used.