What is with the insanely huge temptation to make millions of typedefs of "int"???

[Trishia Rose]

God dang it all, i am so sick of programmers who feel compelled that
every single time they use an int variable, they give it some brand
new name! for example they would take an innocent function like
int multiply_by_2( int argument )

and change it to

multiply_by_2_return_t multiply_by_2( multiply_by_2_argument_t
argument )

What is the bloody point? Is it a deliberate attempt to render your
code unreadable?

[Artie Gold]

Trishia Rose wrote:

God dang it all, i am so sick of programmers who feel compelled that
every single time they use an int variable, they give it some brand
new name! for example they would take an innocent function like
int multiply_by_2( int argument )

and change it to

multiply_by_2_return_t multiply_by_2( multiply_by_2_argument_t
argument )

What is the bloody point? Is it a deliberate attempt to render your
code unreadable?

Actually, it's often an attempt to make things _more_ readable.

In the example you've quoted, of course, that attempt has failed.

<serious>
In some cases, there's a good reason to do this; sometimes you want to
decouple the concrete type of an argument (particularly of an integral
type) from the signatures of families of functions in order to account
for differences in platform. That way, the particular concrete type can
be dealt with through a typedef.
</serious>

HTH,
--ag

--
Artie Gold -- Austin, Texas

[Sven Gohlke,,,]

Trishia Rose wrote:

God dang it all, i am so sick of programmers who feel compelled that
every single time they use an int variable, they give it some brand
new name! for example they would take an innocent function like
int multiply_by_2( int argument )

and change it to

multiply_by_2_return_t multiply_by_2( multiply_by_2_argument_t
argument )

What is the bloody point? Is it a deliberate attempt to render your
code unreadable?

Depends on the environment. Sometimes the style is used to make code more
readable and sometimes it is used to establish something like a signature.
Often YOu want functions with the same name and different arguments. While
C++ supports this (and internally does something similar) You have to work
around in C.
A different application of this style is writing portable code. This way You
can typedef the neccesary machine dependency (usual at the start of the
file or in an extra header) and don't have to rewrite tons of code (while
You have to read tons of code to understand it -- nothing comes
costless:-(.

While it isn't so bad in principle, an overusing is of cause unreadable.
--
Best Regards
Sven

[James Kew]

"Trishia Rose" <ad********@yahoo.com> wrote in message
news:b2**************************@posting.google.c om...

God dang it all, i am so sick of programmers who feel compelled that
every single time they use an int variable, they give it some brand
new name!

IMHO, to add additional information to the interface and the user's code.

An int is just an int -- the type says nothing about what it represents. A
frequency and a time delayare both (unsigned) integers, but they represent
very different and distinct concepts.

Typedeffing distinct types gives the user of an interface additional
information:

* it makes function prototypes more descriptive
* it discourages the user from assigning between variables of those types

It may also provide additional information for tools: for example, PC-Lint
can be configured to treat typedefs of integral types as strong types and
warn on inappropriate assignment.

It can, as you noted, be taken too far: I tend to typedef for quanties of
distinct physical units (FrequencyHzT; DelayMsT; and TimeoutMsT is distinct
as it has special "immediate" and "infinite" values) but not for quantities
of things (size-of-buffer, read-count, number-of-sectors).

--
James Kew
ja*******@btinternet.com