[conspire] [Gimp-user] PNG troubles

[Ruben Safir]

> Your claim that PNG is wholly undependable and not a usable format is
> _also_ grossly inaccurate.  

Its not a claim but a hypothesis based on experience.  twice in 20 years
I found problems with previously made PNG files that later failed to be
renderable by the (at the time) current PNG.

I was further confused by an email from one of the develoeprs whewre he
claimed rthat even according to the PNG specifications it is possible to
create PNG files with one application that are not renderable in another
application although they both, by the letter, complly with PNG
standards.

When I went back and reread his email after our conversation, however, I
think his example would require an intepretation of the standard that is
broader than its intent or any reasonable interpretation of the
specification.  Essentailly, he was saying that there is no defined
behavior when an error is found in the file, so if the library in use
just discovers the fault and ignores it, that would be "correct" by the
standard, and would allow it to render images that otherwise might not
render.

eh, he is ignoring that the library would also need to construct broken
files within the specification, so his logic is not convincing to me.

The other fear I raised was that perhaps the PNG format was so complex
that the coders, using a kitchen sink mentality, simply can't produce a
consistantly working libray, just because of human error, which is
always a factor in coding projects.  I'm not convinced at this time that
this is true, but I am warey of this.  I am always concerned about this,
regardless of the project, but more so in what I consider a key free
Software ecosystem component like PNG.

They seem to have a small application which fixes the files I showed
them, so that is good.  One time for having this library bug in a decade
is plenty.  My expection for failure to render is zero, that is they
render forever by any application that conforms with the specification.

That this bug was only for a short period of time which produced images
that can be viewed on some implementations, but not others, is a
consideration, but it still makes me concerned about the stability of
PNG files on the scale of say, 500 or a 1000 years?


> That is not supported by the facts of your
> problem, either, and the problem doesn't even lie in the PNG chunk check 
> values at all, but rather in the zlib check value included as a
> cross-check attempt to verify that the eflate and inflate algorithms are
> implemented correctly.  There is thus an infuriating legacy problem with
> the treatment of (many) older PNG files, but that is not the same thing
> as what you said, at all.
> 
> It appears that Cosmin Truta will be soon offering you a comprehensive 
> fix in the form of the 'special build of OptiPNG' that you can use to 
> mass-fix your PNGs stored with erroneous zlib check values using that
> buggy libpng version available briefly during part of 2004.
> 
> So, happy ending for you.
> 
> Does that mean 'PNG files are wholely undependable and not a usable
> format'?  Hell no.  And it's really bloody annoying to see you cry wolf
> like that.
> 
> It means that software sometimes has bugs, and you got bit by one.
> Sorry about that, but it's the way of the world.
> 
> 
> _______________________________________________
> conspire mailing list
> conspire at linuxmafia.com
> http://linuxmafia.com/mailman/listinfo/conspire

-- 
So many immigrant groups have swept through our town
that Brooklyn, like Atlantis, reaches mythological
proportions in the mind of the world - RI Safir 1998
http://www.mrbrklyn.com 

DRM is THEFT - We are the STAKEHOLDERS - RI Safir 2002
http://www.nylxs.com - Leadership Development in Free Software
http://www2.mrbrklyn.com/resources - Unpublished Archive 
http://www.coinhangout.com - coins!
http://www.brooklyn-living.com 

Being so tracked is for FARM ANIMALS and and extermination camps, 
but incompatible with living as a free human being. -RI Safir 2013