svn commit: r303626 - head/www/xpi-firebug

Chris Rees utisoft at
Wed Sep 5 21:21:54 UTC 2012

On 5 September 2012 22:09, Doug Barton <dougb at> wrote:
> On 09/05/2012 14:01, Chris Rees wrote:
>> On 5 Sep 2012 21:54, "Doug Barton" <dougb at
>> <mailto:dougb at>> wrote:
>>> Hash: SHA256
>>> On 09/05/2012 13:35, Florian Smeets wrote:
>>> > On 04.09.12 18:35, Doug Barton wrote:
>>> >> Portmaster, and other tools, use this method extensively to read
>>> >> those files, and this issue causes those tools to break while
>>> >> processing those files.
>>> >>
>>> >
>>> > You keep repeating this, without proof.
>>> Um, what kind of proof do you need? I wrote the code, I know how it
>>> works, and I know that the parts of it that use "while read" to
>>> process distinfo files breaks because of this issue. There have also
>>> been other people who have raised this issue, as relates to their tools.
>>> I also showed you how to reproduce it yourself.
>>> How much more proof do you need?
>> I think Florian is asking for a real example, perhaps a portmaster command;
>> % portmaster reproduce-this-error
> If Florian isn't interested in taking my word for it, he can look
> through the code and find ways to reproduce it himself. :)

Well, since you provoked me into it, I've had a look, the main
distinfo part I've distilled into something simpler:

[crees at pegasus]~% svn co $ports/head/www/xpi-firebug at r303627
A    xpi-firebug/distinfo
A    xpi-firebug/pkg-descr
A    xpi-firebug/pkg-plist
A    xpi-firebug/Makefile
Checked out revision 303627.
[crees at pegasus]~% cd xpi-firebug
[crees at pegasus]~/xpi-firebug% <distinfo while read s f discard ; do
while>  echo $s
while> done
[crees at pegasus]~/xpi-firebug%

Granted, that leaves off the SIZE line, but I can't see any problems here...

The only part of portmaster that I can see that cares about SIZE
depends on ports/106483, which is still in waiting (is this going to
be needed with pkgng anyway?)

Given this... specifics could be nice for people who can't distil the
actual problem.

I understand that this is correct, but you can't claim it affects
portmaster when it actually doesn't.


More information about the svn-ports-all mailing list