Nicolas Boulenguez nico...@debian.org a écrit :
Package: src:gprbuild
Followup-For: Bug #763727
Hello.
If one of you can reach a kfreebsd-i386 machine right now, please
rebuild the package with DEB_BUILD_OPTIONS=nostrip, so that the trace
produced by gdb contains more useful information.
I
Nicolas Boulenguez nico...@debian.org a écrit :
Package: src:gprbuild
Followup-For: Bug #763727
Hello.
If one of you can reach a kfreebsd-i386 machine right now, please
rebuild the package with DEB_BUILD_OPTIONS=nostrip, so that the trace
produced by gdb contains more useful information.
Can
I have a kfreebsd-i386 virtual box on my laptop and I can confirm that
Nicolas' latest patch [https://bugs.debian.org/763727#25] reproduces the
same symptoms as in my message [https://bugs.debian.org/763727#10].
That is, he solved the handling of exceptions in a different way (not
raising the
Here is the result of gprconfig under gdb :
(gdb) where
#0 0x08142483 in __gnat_backtrace ()
#1 0x08137fb0 in system.traceback.call_chain ()
#2 0x080eeb7b in ada.exceptions.call_chain ()
#3 0x080eef6b in ada.exceptions.complete_occurrence ()
#4 0x080eef88 in
Strangely I am unable to reproduce either of the bugs in gprbuild,
namely the mishandling of exceptions and the exception inside
GNAT.Expect. The following program runs successfully on kfreebsd-i386.
This suggests that something really fishy is going on in the way GNAT
compiles gprbuild but not
D'oh.
The problem is caused not by raising or handling exceptions but by
storing the exception traceback in the exception occurrence, as
specified by passing -E to gnatbind. If I bind hello.adb with -E I can
reproduce the storage_error in GNAT.Expect.
As I don't see any reason why gprbuild
The Constraint_Error is triggered when parsing the following node in
gprbuild's compilers.xml:
compiler_description
nameGCC-28/name
executablegcc/executable
version
externalgcc -v/external
grep regexp=^gcc \S+ (\S+) group=1/grep
must_match2\.8\./must_match
Package: src:gprbuild
Followup-For: Bug #763727
Hello.
If one of you can reach a kfreebsd-i386 machine right now, please
rebuild the package with DEB_BUILD_OPTIONS=nostrip, so that the trace
produced by gdb contains more useful information.
Thanks.
--
To UNSUBSCRIBE, email to
Package: src:gprbuild
Followup-For: Bug #763727
Please ignore my previous message, sent while you were answering.
From the gprbuild documentation, I understand that non-number values
for the prefix attribute are meaningful.
If the bug really lies in the way gnat handles Constraint_Error, it
Package: src:gprbuild
Followup-For: Bug #763727
With the diff…
#
# old_revision [950bbf7e9813c77e32c46681053963808b47815c]
#
# patch debian/patches/763727.patch
# from [6023754390a73c950a02592a9e006c3f23cb0a3c]
#to [aa2bbcb0c95f667934874f7d26929d3e2909b5ab]
#
Source: gprbuild
Severity: important
Hello, here are the symptoms visible in the buildd log of libgnatcoll on
kfreebsd-i386, which I could reproduce in a virtual box:
gprbuild -v -j4 -R -v -eS -XADAFLAGS=-g -O2 -fstack-protector-strong
-XCFLAGS=-g -O2 -fstack-protector-strong -Wformat
11 matches
Mail list logo
