This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [rfc] Implementation of qXfer


> Date: Thu, 22 Jun 2006 16:13:11 -0400
> From: Daniel Jacobowitz <drow@false.org>
> 
> On Thu, Jun 22, 2006 at 11:07:50PM +0300, Eli Zaretskii wrote:
> > They are OK, except for this issue with anchors whose names contain a
> > colon: this is against the Texinfo language rules.  You might be lucky
> > with a particular Info reader, but there are others out there, and
> > each one of them uses a different method of searching for
> > cross-reference names (some use fixed strings, others use various
> > regular expressions).  Using a colon makes them fail in different
> > situations and in different interesting ways.
> 
> Do you have a reference for this?  I spent a while with the Texinfo
> documentation when I was working on that, and could not find it.  Just
> curious.

The Texinfo manual doesn't say that explicitly for anchors.  It does
say that for node names:

   * Unfortunately, you cannot use periods, commas, colons or
     parentheses within a node name; these confuse the Texinfo
     processors.

These limitations are due to the syntax of menus and cross-references
where these characters delimit syntactically significant parts of the
reference.

In the section about anchors, there's this text that hints that
anchors and nodes share the same limitations:

      Anchor names and node names may not conflict.  Anchors and nodes are
    given similar treatment in some ways; for example, the `goto-node'
    command in standalone Info takes either an anchor name or a node name as
    an argument.

I will ask the Texinfo maintainer to add an explicit text about anchor
name limitations.

> > So please let's remove the colons and use some other mnemonic methods
> > to make the xref reflect the packet descriptor.
> 
> Blech.  Well, I don't know what else to use.  I tried using
> qXfer-auxv-read, but it's pretty ugly.

How about "qXfer auxv read" or "qXfer read auxiliary vector"?

> > AFAICS, you just moved the text elsewhere and replaced qPart with
> > qXfer, right?  If there are new portions of text, please tell where
> > they are.
> 
> No, much of the text is changed; the packet has a different format
> and different replies.  I definitely rewrote the Reply: table and
> the paragraphs right after qXfer:OBJECT:read and qXfer:OBJECT:write.

Well, at least in this hunk the text was not touched:

> -Fields within the packet should be separated using @samp{,} @samp{;} or
>  @cindex remote protocol, field separator
> +Fields within the packet should be separated using @samp{,} @samp{;} or
>  @samp{:}.  Except where otherwise noted all numbers are represented in
>  @sc{hex} with leading zeros suppressed.

There are several commas missing in these two sentences.

Anyway, I've read all of your text yesterday, so it's okay with me.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]