help-octave
[Top][All Lists]
Advanced

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

Re: inadequate/inconsisten 'help pkg' message; problems with downloading


From: Ben Abbott
Subject: Re: inadequate/inconsisten 'help pkg' message; problems with downloading
Date: Sun, 07 Aug 2011 12:37:53 -0400

On Aug 7, 2011, at 7:36 AM, Sergei Steshenko wrote:

> --- On Tue, 8/2/11, Sergei Steshenko <address@hidden> wrote:

> In order to avoid the non-informative message which sparked this whole
> investigation I modified the following piece of 'uninstall' function:
> 
>    924
>    925   num_packages = length (installed_pkgs_lst);
>    926   delete_idx = [];
>    927   for i = 1:num_packages
>    928     cur_name = installed_pkgs_lst{i}.name;
>    929     if (any (strcmp (cur_name, pkgnames)))
>    930       delete_idx(end+1) = i;
>    931     endif
>    932   endfor

The version of pkg.m you are referencing is different than that in the 
developer's souces (default branch).

 925       warning ("couldn't clean up after my self: %s\n", msg);
 926     endif
 927   endfor
 928 
 929   ## Add the newly installed packages to the path, so the user
 930   ## can begin using them. Only load them if they are marked autoload.
 931   if (length (descriptions) > 0)
 932     idx = [];

If you were to provide a diff instead, it would likely be possible to apply it 
to the developers sources.  As it is, someone (me) will have to look through 
the sources and edit them manually.

> 
> As I wrote in my earlier Emails, I'm going to release the whole 'pkg.m'
> including the above change.
> 


Including a diff is a simple task. Attaching the original pkg.m without the 
line numbers is even easier. With all the line numbers included someone will 
have to either (1) copy-n-paste that into an editor, (2) remove all the line 
numbers, (3) create changeset/diff, and (4) push it to the tracker ... or just 
ignore your email.

I have no intent so elevate or insult, but why are you making this so painful?

In any event, I've read through your analysis, and looked at the developers 
sources. It appears to me that the same problem is still present. I've attached 
a changeset. I've made soem modifications to your suggestion that I think 
improves the warning message and is consistent with Octave's coding 
conventions. Please let me know if I've missed anything.

Ben

Attachment: changeset.patch
Description: Binary data



reply via email to

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