Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Created attachment 33185
-- https://gcc.gnu.org/bugzilla/attachment.cgi?id=33185action=edit
librdf.gpr
gnatmake -d -P/home/porton/Projects/redland-bindings/ada/test.gpr
-XRUNTIME=full -XMODE=Install -s --subdirs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61911
Victor Porton porton at narod dot ru changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
The below program, compiled with GNAT 4.9, calls Adjust two times whenÂ
copying a T1 object.
But it does the same operations with a T2 object without calling Adjust.
So calling Adjust on a T1
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Created attachment 33370
-- https://gcc.gnu.org/bugzilla/attachment.cgi?id=33370action=edit
An Ada file whose semantics does not verify
I attach an Ada package file which does not verify
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Created attachment 33381
-- https://gcc.gnu.org/bugzilla/attachment.cgi?id=33381action=edit
Apply gnatchop to this file
Apply gnatchop to all.chop (attached).
$ gnatgcc -v
Using built-in specs.
COLLECT_GCC=gnatgcc
Priority: P3
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Created attachment 33382
-- https://gcc.gnu.org/bugzilla/attachment.cgi?id=33382action=edit
A program which doesn't compile
The below program (its copy is also
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62236
--- Comment #1 from Victor Porton porton at narod dot ru ---
Confirmed in GNAT GPL 2014.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #2 from Victor Porton porton at narod dot ru ---
Confirmed in GNAT GPL 2014.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62205
--- Comment #1 from Victor Porton porton at narod dot ru ---
It does not work with GNAT GPL 2014.
$ gnat compile -gnatc -gnat12 test.ads
gcc -c -gnatc -gnat12 test.ads
test.ads:19:28: Default Iterator must be a primitive of Descriptions_List
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
--- Comment #1 from Victor Porton porton at narod dot ru ---
Confirmed in GNAT GPL 2014.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62205
Victor Porton porton at narod dot ru changed:
What|Removed |Added
Version|4.9.1 |4.9.2
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62236
--- Comment #3 from Victor Porton porton at narod dot ru ---
It does not work also with GANT 4.9.2.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #4 from Victor Porton porton at narod dot ru ---
Also does not work with GNAT 4.9.2.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65863
--- Comment #1 from Victor Porton porton at narod dot ru ---
Linux victor.local 3.10-2-amd64 #1 SMP Debian 3.10.7-1 (2013-08-17) x86_64
GNU/Linux
with 32 bit userland
: bootstrap
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
I can't compile gcc-5.1.0 with gcc-4.9.2.
./configure --prefix=/usr/local/stow/gcc-5.1.0 --enable-libada
--enable-languages=c,c++,java,lto,objc,ada
make[3]: Entering directory
'/home/porton
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65863
Victor Porton porton at narod dot ru changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Compiling GCC 5.1.0 with GCC 4.9.2.
Debian jessie
Linux victor.local 3.10-2-amd64 #1 SMP Debian 3.10.7-1 (2013-08-17) x86_64
GNU/Linux
with 32 bit userland
./configure --prefix=/usr/local/stow/gcc-5.1.0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65868
--- Comment #2 from Victor Porton porton at narod dot ru ---
Your suggestion to build in a separate directory helped. It compiled
completely.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
Victor Porton porton at narod dot ru changed:
What|Removed |Added
Version|4.9.0 |5.1.0
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
Victor Porton porton at narod dot ru changed:
What|Removed |Added
Version|4.9.2 |5.1.0
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62236
Victor Porton porton at narod dot ru changed:
What|Removed |Added
Version|4.9.2 |5.1.0
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
Victor Porton changed:
What|Removed |Added
Version|5.1.0 |5.2.1
Severity|normal
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #6 from Victor Porton ---
Not fixed in GNAT 5.2.1 20150903 (i586-linux-gnu) (Debian gnat-5 package).
However the exact error message is somehow different.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
--- Comment #8 from Victor Porton ---
> Only under certain circumstances though.
Yes, such circumstances as my toy (but representing a real case) program.
Such programs SHOULD be optimized, because the standards allow optimization
under such
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
--- Comment #4 from Victor Porton ---
I utterly disagree.
Avoiding calling unnecessary routines when copying an object, is a VERY useful
optimization, just because copying operation is VERY often in real software.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
--- Comment #6 from Victor Porton ---
> But you generally cannot do this optimization with controlled objects
> in real software.
Why we cannot?!
If you are afraid that this may introduce errors, you are not right, as this is
allowed by Ada
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62236
--- Comment #5 from Victor Porton ---
Not fixed in 5.3.1.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
--- Comment #9 from Victor Porton ---
Not fixed in 5.3.1.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #7 from Victor Porton ---
Not fixed in 5.3.1.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #9 from Victor Porton ---
A similar bug (but with another error message) appears in GNAT GPL 2016 for
Windows x32 32bit.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #10 from Victor Porton ---
(In reply to Victor Porton from comment #8)
> Note fixed in GCC 6.1.1.
Sorry, should read "Not fixed in GCC 6.1.1."
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62042
--- Comment #10 from Victor Porton ---
Not fixed in GCC 6.1.1.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #8 from Victor Porton ---
Note fixed in GCC 6.1.1.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62236
--- Comment #7 from Victor Porton ---
Not fixed in GCC 6.1.1.
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
The following valid code throws an exception:
-- test_count.adb start
with Ada.Strings.Fixed, Ada.Integer_Text_IO;
use Ada.Strings.Fixed, Ada.Integer_Text_IO;
procedure Test_Count
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62236
Victor Porton changed:
What|Removed |Added
Status|NEW |RESOLVED
Known to work|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81328
Victor Porton changed:
What|Removed |Added
Status|RESOLVED|REOPENED
Resolution|WONTFIX
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 41688
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41688=edit
A source with a deliberate error
GNAT provides an error list for my erroneous input in wrong order.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62235
--- Comment #11 from Victor Porton ---
The bug was NOT fixed in GCC 7.1.0.
This is a long standing bug :-(
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42345
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42345=edit
The compilation units which should produce a warning
The command &quo
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42023
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42023=edit
A file which does not compile
The compiler refuses to verify the code in attached f
Priority: P3
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42734
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42734=edit
a valid program which does not comp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #8 from Victor Porton ---
Created attachment 42634
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42634=edit
Sources in gnatchop format
Run
gnatchop all.chop to extract the sources.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #6 from Victor Porton ---
I found that the presence of the bug depends on presense `-lahven` in the
linker command.
Adding -lahven leads to the bug, removing it removes the bug.
Weird?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #7 from Victor Porton ---
There is an easier way to reprise the bug:
Copy all source files into a directory and run:
gnatmake simple_test -largs -lahven
./simple_test
To reprise the bug, you need Ahven installed.
I will attach
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #5 from Victor Porton ---
Created attachment 42633
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42633=edit
The source reprising the bug without "ainet"
Needs "ahven" to reprise the bug.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #10 from Victor Porton ---
I notified the Ahven author.
However, the bug is quite weird: Isn't linking with a library just no-op if no
functions from the library are called? I wonder how the bug may depend on
linking with it. So I
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #11 from Victor Porton ---
Debian libahven27 package version 2.6-1.2.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #17 from Victor Porton ---
Created attachment 42666
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42666=edit
Minimal example reprising the bug
I've created the minimal example reprising the bug.
The bug is actually awful:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #16 from Victor Porton ---
I've discovered that Ahven source uses tasking.
So it is most likely some tasking problem with Ahven.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #21 from Victor Porton ---
@Eric Botcazou:
Adding binder flag -F solves the problem in the attached examples
https://gcc.gnu.org/bugzilla/attachment.cgi?id=42630 and
https://gcc.gnu.org/bugzilla/attachment.cgi?id=42666 but not in my
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42630
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42630=edit
The source reprising the bug
Extract attached bug.zip into an empty directory.
Run
make comp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #3 from Victor Porton ---
GCC 7.2.0, Debian Linux 4.13.0.
Eric Botcazou: I can't make a standalone reproducer without project file,
because the bug appears only when I compile in two stages: creating an .a
library and then linking
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #4 from Victor Porton ---
If I remove `with "ahven";` from spawn_tests.gpr the bug disappears.
Seems to do something with elaboration order.
I also deleted `with "anet";` from spawn_common.gpr without any effect on the
bug.
I will
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83027
--- Comment #13 from Victor Porton ---
It is a GCC bug rather than an Ahven bug, because the bug is triggered by `with
Ada.Text_IO;` (and disappears if we remove this line from spawn-signals.adb)
which should not influence semantics of the
: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42805
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42805=edit
The source causing the bug
Use the attached all.chop. Chop it with gnatchop.
I have not succeeded to create a mini
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83310
--- Comment #2 from Victor Porton ---
Created attachment 42807
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42807=edit
project file
note that sources should be put into src/ subfolder
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83310
--- Comment #1 from Victor Porton ---
Created attachment 42806
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42806=edit
forgotten Makefile
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82639
Victor Porton changed:
What|Removed |Added
Status|RESOLVED|REOPENED
Resolution|INVALID
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82639
Victor Porton changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Dynamic predicate for a non-private record should give a warning, because Ada
RM does not warrant consistency when a field is assigned (or changed in other
way).
So
: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42413
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42413=edit
Compiler input triggering bug
The files are attached. Use gnatchop to extract them.
$ gnatgcc -c uri.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82637
--- Comment #1 from Victor Porton ---
Possibly related bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82638
: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42414
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42414=edit
Compiler input triggering bug
GNAT 7.2.0 for Debian.
This bug may be rela
: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
Created attachment 42421
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42421=edit
Compiler input triggering bug
We get the following nonsense error message with GNAT 7.2.0 on Debian Li
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82639
--- Comment #1 from Victor Porton ---
This was an example of a legal Ada program which does not compile with GNAT.
Workaround: Don't use a private type here.
Priority: P3
Component: c
Assignee: unassigned at gcc dot gnu.org
Reporter: porton at narod dot ru
Target Milestone: ---
GCC reports unused result even when explicitly converted to void to show the
developer's intent not to report a warning.
lib.c: In function
67 matches
Mail list logo
