[Top][All Lists]

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

Re: [lwip-users] Lots of RAM

From: Andrew Lentvorski
Subject: Re: [lwip-users] Lots of RAM
Date: Sun, 09 Sep 2007 11:38:16 -0700
User-agent: Thunderbird (Macintosh/20070728)

Frédéric BERNON wrote:
I'm agree with Jonathan & Simon (but I think we should add $date$ and $revision$ CVS keyword, is there any objections to add them ?).

I object.  Strenuously.

It's already tough enough to mesh the fairly unusual organization of the lwip code with other projects as well as track upstream changes. Adding CVS keywords is going to make diffing unnecessarily annoying. If you must, at least put them at the *bottom* of the file.

The source code is your gold standard. Your repository is where all the metadata for you source code goes. The two should not meet.

The whole "CVS keyword" problem is why you need weird "Oh, leave this file alone" flags so that CVS doesn't do keyword interpolation (want to check in a tar file? oops. Have a script that looks for CVS keywords? Oops.)

There are lots of CVS annotation tools that exist. Most of them are *far* more informative than any meager amount of information a CVS keyword could convey.

The correct answer to this is either A) use/learn how to use a better editor/IDE/whatever for computer work or B) use a better prettyprinter/CVS browser/etc. for hardcopy/web display.

The wrong answer is modifying the checked in code.

If the problem, however, is: "I can't do CVS operations when offline." then the solution is to upgrade to a better source control system.

This objection comes from the fact that I actually check out lwip from CVS and then check it right back into my mercurial repository (mercurial is much nicer when trying to merge upstream changes than CVS). I can think of several things that will likely break if keywords get added to the mix.


reply via email to

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