gnu-arch-users
[Top][All Lists]
Advanced

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

Re: [Gnu-arch-users] File naming conventions


From: Dustin Sallings
Subject: Re: [Gnu-arch-users] File naming conventions
Date: Sun, 17 Oct 2004 23:20:51 -0700


On Oct 17, 2004, at 21:43, Zenaan Harkness wrote:

0 = neutral, 1 = annoyance, 2 = potential trouble, 3 = serious

FILETYPE  bash  tcsh  csh  cygwin  mingw  w32_native  OSX
{files}    1     1     1     ?       ?        ?        ?
+files     1     1     1     ?       ?        ?        ?
=files     1     0     0     ?       ?        ?        ?
,files     0     3     2     ?       ?        ?        ?

Recent OS X without preferences use bash. Longer-term OS X users without preferences use tcsh (I don't remember exactly when they changed the default shell). Users with preferences obviously use their preferences. The only issue with OS X specifically is that it has case insensitive filenames by default.

I don't think comma files are all that serious for {t,}csh unless you're trying to do {} expansions (which should be the same for bash).

Now, the one place I've had a nuisance with case insensitive filesystems is in bringing in patches from more than one contributor to a project when two different users used the same name for a branch in unrelated archives.

        I.e. if I have the following:

        address@hidden

        And I merge in something that has a patch from the following archive:

        address@hidden

And then commit, I have created a broken archive (actually, that recipe may not be *exactly* correct, but I did something similar to this and broke an archive completely). When I was done, the revision could not be checked out anywhere.

There was some discussion about this on the list. Basically, the argument was that other stuff can break when you're dealing with case insensitive filesystems, and it should be a policy issue. I argued that the tree seems inverted ({arch}/c/c--b/c--b--v/address@hidden instead of {arch}/address@hidden/c/c--b/c--b--v) which both makes more sense and is more consistent to me, and would also avoid this class of problem.

I personally have come to like {arch} and ,files and think they're good.

I still don't get the appeal of {arch}, but I haven't had too much of a problem with the comma files. In particular, they only seem to occur when your tree is in a ``weird'' state (you've got undos, or some operation broke). Of course, looking in an {arch} tree gives one the Angry Fruit Salad feel:

-rw-------  1 dustin  staff    50 14 Jan  2004 ++default-version
drwx------  7 dustin  staff   238 19 Sep 11:22 ,,inode-sigs/
-rw-------  1 dustin  staff    52 14 Jan  2004 .arch-project-tree
-rw-------  1 dustin  staff  6795 14 Jan  2004 =tagging-method
drwx------  3 dustin  staff   102 14 Jan  2004 misc/


--
SPY                      My girlfriend asked me which one I like better.
pub  1024/3CAE01D5 1994/11/03 Dustin Sallings <address@hidden>
|    Key fingerprint =  87 02 57 08 02 D0 DA D6  C8 0F 3E 65 51 98 D8 BE
L_______________________ I hope the answer won't upset her. ____________





reply via email to

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