Adam Tauno Williams wrote:
> > How many fixed the problem and never published the fix?
> Not many I bet, at least not for significant bugs. Perhaps trivial one
> liners.
Trivial one liner can be difference between working and not working code, so
simplicity to solve problem doesn't mean that it should not be reported.
The "trivial one liner" is just a *rare* outlier that it barely merits
consideration.
> > They use free OS, but don't give back anything, keeping their knowledge
> > for another opportunity to cash on it.
> Or, alternatively, have no intent to cash in on it at all, just want to get
> their job done.
Agree. That is the same case as those that fixed, but "have no time, will,
interest, knowledge or even permission to dicuss further".
Which is just dumb since if you fix without sending upstream you will
have to fix again, and again, and again.... The concept that not
submitting a fix saves time is crazy; if you really want to save time
then use packages, no way does patching and compiling your own code
constitute efficient use of time.
If you need to fix a bug that the developers don't care about or
mark as WONTFIX (Ulrich Drepper and glibc, anyone?), you often
don't have the choice.
And believe me, sometimes it's *less* effort to fix it again and
again than to cope with the idiosynchrasies of upstream developers.
I'm doing free software development since 1981, when we received
our first TeX tape from Stanford, and have experienced that problem
many times in those 25 years.
In fact, maintaining such patches to Unix source tree was the
original raison-d'etre for CVS. That's why cvs import doesn't
create revisions on the trunk. (I was involved in the 1989 CVS
sh-to-C rewrite effort of Brian Berliner, going from CVS 1.2 to
1.3, so this statement is from personal experience.)
If you think otherwise, try to get a patch into OpenCA, or search
for WONTFIX in the bugzilla entries of glibc.
Joachim
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Joachim Schrod Email: [EMAIL PROTECTED]
Roedermark, Germany
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
