bug-gnulib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Problems with directory trees "confdir-14B---" and "confdir3"


From: Peter Dyballa
Subject: Re: Problems with directory trees "confdir-14B---" and "confdir3"
Date: Thu, 21 Mar 2024 13:27:40 +0100

> Am 20.03.2024 um 23:23 schrieb Paul Eggert <eggert@cs.ucla.edu>:
> 
> On 3/20/24 01:51, Peter Dyballa wrote:
>> I was asking this myself: What is this test good for? It's possibly 15 years 
>> after it was introduced?
> 
> If old glibc versions are no longer worth worrying about, a similar argument 
> would apply to old Mac OS X versions, no?

I was thinking something like: This is an old test, maybe checking for a bug 
that's even older…

Mac OS X 10.4, Tiger, is from Easter 2005, almost 20 years old now. Leopard, 
Mac OS X 10.5, is 2½ years or 13% younger. The test 
gl_cv_func_getcwd_succeeds_beyond_4k seems to check whether something's still 
working beyond the official bounds of the system. Would someone use this 
"feature"? So what is the test's use? On these old Mac OS X systems it seems we 
can live without the test… (Are we so unique, so developed?)

A length of 1000 characters for a directory's name could mean a directory depth 
of up to 100 levels. This should be enough. The longest directory name on my 
youngest (and fastest and most energy efficient) Mac is 332 characters and has 
24 levels, it's inside the documentation of legacy objects in Xcode, Apple's 
developer tools.

Deepest directory hierarchy here seems to be 32 (290 characters long) – a 
statistical file system could be nice…

--
Greetings

  Pete

A morning without coffee is like something without something else.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]