Hi Janus,
I completely missed that this patch is still pending review as you wrote
"committed". However, it only applied to the ICE and not to the
documentation and backporting.
On 10/01/2011 02:00 PM, Janus Weil wrote:
I have just committed as obvious a one-line patch to fix a regression
which is triggered by -fwhole-file:
http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=179413
Should I backport to 4.6? And also to 4.5?
Yes, please backport to 4.6. If you want, you can also backport to 4.5.
However, I do not think that -fwhole-file gets used in 4.5 - especially,
as it has some issues.
Moreover, I noticed that
http://gcc.gnu.org/onlinedocs/gcc-4.6.1/gfortran/Code-Gen-Options.html#Code-Gen-Options
falsely claims that -fwhole-file is not used by default (which it is
since 4.6, see http://gcc.gnu.org/gcc-4.6/changes.html). Attached is a
small documentation patch to fix this. Ok to apply to trunk and 4.6?
OK. By the way, I was thinking of completely removing that option.
-fwhole-file avoids some wrong-code issues, generates faster code, and
improves diagnostic. Thus, there is no real reason* of keeping
-fno-whole-file. Thus, I would like to deprecate the option. (* = one
reason is to see whether a bug is due to this flag, but that does not
really help with fixing the issue.)
I think your wording is fine, I still wonder whether one should somehow
imply more clearly that -fno-whole-file is not recommended and that the
whole flag is deprecated.
Tobias
PS: I think except for procedure calls where the explicit interface is
not known, gfortran generates finally the correct declaration for the
middle end. At least I have not encountered any multiple-declaration
issues for a long time!
