This is the mail archive of the
mailing list for the Insight project.
Re: [RFC] syntax highlighting
- From: Steven Johnson <sjohnson at sakuraindustries dot com>
- To: Keith Seitz <keiths at redhat dot com>
- Cc: insight at sources dot redhat dot com, Th dot R dot Klein at web dot de
- Date: Wed, 24 Aug 2005 09:11:44 +1100
- Subject: Re: [RFC] syntax highlighting
- References: <42F134B1.email@example.com> <430B8791.firstname.lastname@example.org>
Keith Seitz wrote:
I've ask Vineeth if the developing of the Patch already have started.
Since Vineeth told me that this is not the case.
So I'm sending here a patch, which should carefully reviewed.
The method isn't the fastest but it seems OK for the moment.
I apologize for the long delay, but I'm really conflicted on this patch.
As much as I would like to see syntax highlighting, I'm not entirely
sure that I like the approach that has been taken. I would prefer, of
course, a real parser for these things. Gdb contains parsers for
C/C++, Java, ADA, and whatever else is supported.
Whats the story with those? I thought it was only for expression
passing, so you could enter expressions like the language you are
On the other hand, we have nothing today. Zippo. Nadda. That's a
compelling reason to take what you've kindly offered.
So I guess the best compromise for me is to accept your patch, BUT, I
would like to see it a bit more modular, so that the parsing engine
can be replaced with something more efficient. I think the actual
parsing or regexp'ing should be done in a separate class.
I agree that it being more modular would be good. But what is being
offered is better than nothing, especially as it can be enabled/disabled
as the user feels.
Barring that, there is one more hurdle to overcome: copyright. Unlike
most of the patches that I see here, this patch definitely is not even
remotely trivial. So we need to get you (and your company if involved)
to sign an assignment. The problem is, that assignment is to Red Hat,
not the FSF, since Insight is officially owned (albeit carelessly) by
I dont think there is any rule that says a whole file needs to be
Copyright to a single entity. Parts of a file can have different
Copyright status. How you manage that efficiently and unambiguously in
an open source project, I dont know. What I have seen in the past, in
open source projects that use modified code from different sources, is
headers that say things like (as an example):
# Original file Copyright Redhat (c) 2000,2001,2002,2003,2004,2005
# Portions Copyright FSF (c) 2005
Whether that is appropriate in this case is unknown.
As the whole file is distributed under the GPL, there is no impediment
to modifying the file and having those modifications have any copyright
you like, provided you obey the GPL. Which is what this project does,
so no problem. The only reason I think the FSF insists on copyright
assignment is so it can keep better control of its core projects, and
litigate on infringements if required. Otherwise it isnt required given
the words of the GPL. There are certainly lots of GPL projects that
dont worry about assignment of copyright, and instead just rely on the
words of the GPL to cover the modifications contributed by various others.
It would obviously be cleaner to separate the substantial parts out into
separate files, and then there would be no impediment to assigning those
files to the FSF. Then the "hooks" in the original code would belong to
the "insubstantial" category of change.