On 04/12/2018 17:04, Fritz Reese wrote:
On Tue, Dec 4, 2018 at 10:12 AM Jakub Jelinek <ja...@redhat.com> wrote:
Just a couple of random comments.
-fdec-pad-with-spaces option name doesn't look right, because it doesn't say
what the option affects. So perhaps have transfer in the option name?
[...]
Wouldn't it be better to allow specifying whatever character you want to pad
with, so that users can choose something even different?
I concur with this. In that case some option like -ftransfer-pad-char=
may be a more appriopriate name, where -fdec may enable a default
transfer-pad-char of \x20 rather than \x00.
How would the value be specified?
-ftransfer-pad-char=0x20
-ftransfer-pad-char=32
-ftransfer-pad-char=' '
If -fdec and -ftransfer-pad-char are both specified which should have
precedence?
For -fdec and -fno-dec both specified would I be correct in assuming 0x0
should be used?
--- a/gcc/fortran/simplify.c
+++ b/gcc/fortran/simplify.c
@@ -7830,7 +7830,7 @@ gfc_simplify_transfer (gfc_expr *source, gfc_expr *mold,
gfc_expr *size)
[...]
This affects just the simplification when the argument is a known constant.
Shouldn't we handle it also when it is not a constant?
Yes. Mark, you'll need to also patch iresolve.c (gfc_resolve_transfer)
to affect non-constant resolution.
Thanks for the hint.
The tests look too much big-endian vs. little-endian dependent and
ascii dependent. We have pdp-endian and doesn't s390x TPF use EBCDIC?
Wouldn't it be better to compare transfer("ABCE", 1_8) with transfer("ABCE
", 1_8)
Agreed.
---
Fritz
Thanks for the feedback.
Mark
--
https://www.codethink.co.uk/privacy.html