The method is to run "make check" in the fixincludes directory.
If you've made changes, it is supposed to fail.
The diff between expected and actual results get displayed.
The author of the changes is expected to verify that the new results are
correct and then adjust the appropriate file in the "res" directory.
I guess I didn't make that clear enough. Sorry.
At this point, I haven't rebuilt GCC in more than a decade &
the repo says my credentials are invalid, so I'd have to start over
to be able to check up on the current state. :(
Regards, Bruce
On 9/10/25 15:54, Jonathan Wakely wrote:
On Wed, 10 Sept 2025 at 22:55, Bruce Korb <[email protected]> wrote:
Looks reasonable to me. I don't see a patch to the testing results tho.
The "SVN-MERGE" warns against using subversion merging the file instead
of rebuilding and replacing it. The text for "SVN-MERGE" lives in the
template used to create the "fixincl.x" file.
I pushed the patch as r16-3446-g59db4ce2df1db3, including a fix for
the fixincl.x file.
Should there have been a change to the tests? The README talks about a
$(builddir)/tests/res directory but I don't see that.
I'll need some guidance if a change to the tests is needed. Should I
verify that it does fix pthread.h without the bypass string matching,
and also verify that it doesn't fix it when the bypass matches? How
would I do that?
On 8/28/25 06:56, Jonathan Wakely wrote:
Ping
Original: https://gcc.gnu.org/pipermail/gcc-patches/2025-August/692933.html
On 20/08/25 17:30 +0100, Jonathan Wakely wrote:
The pthread_incomplete_struct_argument fix was intended for ancient
versions of Glibc (only 2.3.3 and 2.3.4, I believe). From Glibc 2.3.5
the pthread.h header already included the change to use a pointer
instead of an array, so the fixinclude was no longer used.
However, the https://sourceware.org/bugzilla/show_bug.cgi?id=26647 fix
changed the __setjmpbuf declaration to use struct __jmp_buf_tag __env[1]
again, which caused this fixinclude to start matching again. This means
that GCC now installs a "fixed" pthread.h with a change to a declaration
that guarded by #if ! __GNUC_PREREQ (11, 0), i.e. it's not even relevant
for modern versions of GCC. The "fixed" pthread.h causes problems for
users because of changes to internal implementation details of the
pthread_cond_t type, which require the "fixed" pthread.h to be updated
with mkheaders if Glibc is updated.
This change adds a bypass to the fixinclude, so that it no longer
matches modern Glibc versions, and only applies to glibc versions 2.3.3
and 2.3.4 as originally intended.
fixincludes/ChangeLog:
PR bootstrap/118009
PR bootstrap/119089
* inclhack.def (pthread_incomplete_struct_argument): Add bypass.
* fixincl.x: Regenerate.
---
Tested x86_64-linux.
OK for trunk?
And release branches?
Aside: should we change the comment talking about SVN-MERGE?
fixincludes/fixincl.x | 15 +++++++++++----
fixincludes/inclhack.def | 1 +
2 files changed, 12 insertions(+), 4 deletions(-)
diff --git a/fixincludes/fixincl.x b/fixincludes/fixincl.x
index 819c02c483c7..d28262ef9eeb 100644
--- a/fixincludes/fixincl.x
+++ b/fixincludes/fixincl.x
@@ -2,11 +2,11 @@
*
* DO NOT EDIT THIS FILE (fixincl.x)
*
- * It has been AutoGen-ed August 15, 2025 at 05:30:32 PM by AutoGen
5.18.16
+ * It has been AutoGen-ed August 19, 2025 at 04:54:32 PM by AutoGen
5.18.16
* From the definitions inclhack.def
* and the template file fixincl
*/
-/* DO NOT SVN-MERGE THIS FILE, EITHER Fri Aug 15 17:30:32 CEST 2025
+/* DO NOT SVN-MERGE THIS FILE, EITHER Tue Aug 19 16:54:32 BST 2025
*
* You must regenerate it. Use the ./genfixes script.
*
@@ -7547,8 +7547,15 @@ tSCC zPthread_Incomplete_Struct_ArgumentList[] =
tSCC zPthread_Incomplete_Struct_ArgumentSelect0[] =
"struct __jmp_buf_tag";
-#define PTHREAD_INCOMPLETE_STRUCT_ARGUMENT_TEST_CT 1
+/*
+ * content bypass pattern - skip fix if pattern found
+ */
+tSCC zPthread_Incomplete_Struct_ArgumentBypass0[] =
+ "bits/types/struct___jmp_buf_tag.h";
+
+#define PTHREAD_INCOMPLETE_STRUCT_ARGUMENT_TEST_CT 2
static tTestDesc aPthread_Incomplete_Struct_ArgumentTests[] = {
+ { TT_NEGREP, zPthread_Incomplete_Struct_ArgumentBypass0,
(regex_t*)NULL },
{ TT_EGREP, zPthread_Incomplete_Struct_ArgumentSelect0,
(regex_t*)NULL }, };
/*
@@ -11202,7 +11209,7 @@ static const char* apzX11_SprintfPatch[] = {
*
* List of all fixes
*/
-#define REGEX_COUNT 318
+#define REGEX_COUNT 319
#define MACH_LIST_SIZE_LIMIT 187
#define FIX_COUNT 274
diff --git a/fixincludes/inclhack.def b/fixincludes/inclhack.def
index 35ccaf03c4fa..71777cb10e73 100644
--- a/fixincludes/inclhack.def
+++ b/fixincludes/inclhack.def
@@ -3803,6 +3803,7 @@ fix = {
hackname = pthread_incomplete_struct_argument;
files = pthread.h;
select = "struct __jmp_buf_tag";
+ bypass = "bits/types/struct___jmp_buf_tag.h";
c_fix = format;
c_fix_arg = "%1 *%2%3";
c_fix_arg = "^(extern int __sigsetjmp \\(struct __jmp_buf_tag) "
