Re: Make frontend programs relink after libpgfeutils changes

Поиск
Список
Период
Сортировка
Искать

Re: Make frontend programs relink after libpgfeutils changes

От:
Yingying Chen <cyy9255@gmail.com>
Дата:


On Mon, Jun 15, 2026 at 3:29 PM Chao Li <li.evan.chao@gmail.com> wrote:
Hi,

I ran into some trouble while working on patch [1], where I changed print.c and needed to rebuild psql. However, make silently didn't rebuild psql after the print.c change. I wasn’t aware of that at first, and it confused me and wasted quite a bit of time before I figured out what was happening.

I tried a Meson build. Ninja automatically rebuilds 33 client tools after print.c changes, so I changed make to automatically rebuild client tools in the same way as Meson.

But then I found that there are actually 27 client tools that depend on libpgfeutils.a. There are 6 client tools that don’t really require libpgfeutils.a, but Meson still links the library to them. They are:

* pg_archivecleanup
* pg_config
* pg_controldata
* pg_ctl
* pg_test_fsync
* pg_test_timing

I want this patch to focus on make, and I am not as familiar with Meson as I am with make, so I didn’t touch the Meson part.

With this patch, when something like print.c in libpgfeutils.a is changed, all dependent client tools are automatically rebuilt.

See the attached patch for details. BTW, besides the client tools, test_escape also depends on libpgfeutils.a, so I included it in the patch as well.

[1] https://www.postgresql.org/message-id/A44110E7-6A03-4C67-95AD-527192A6C768%40gmail.com

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/

 
Thanks for the patch. I even ran into the same trouble and got frustrated, and ended up I switched to meson build. So I'd give a big +1 to fix the problem for make users.

I tried the patch, and it works properly. But I see meson automatically rebuilds some more tools than src/bin/ tools, such as contrib/oid2name/oid2name, src/tools/pg_bsd_indent/pg_bsd_indent, etc., do you plan to handle them also?

Regards,
Yingying Chen

Re: Make frontend programs relink after libpgfeutils changes

От:
Álvaro Herrera <alvherre@kurilemu.de>
Дата:
On 2026-Jun-16, Yingying Chen wrote:

> I tried the patch, and it works properly. But I see meson automatically
> rebuilds some more tools than src/bin/ tools, such as
> contrib/oid2name/oid2name, src/tools/pg_bsd_indent/pg_bsd_indent, etc., do
> you plan to handle them also?

Hmm, isn't that bogus?  oid2name at least doesn't depend on fe_utils as
far as I can see.  Maybe meson is overeagerly inheriting that dependency
from somewhere else.  If that's the case then we don't want Make to do
the same (though I don't dare suggest to clean up Meson not to.)

-- 
Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/


Re: Make frontend programs relink after libpgfeutils changes

От:
Chao Li <li.evan.chao@gmail.com>
Дата:
Вложения:


> On Jun 16, 2026, at 20:44, Álvaro Herrera  wrote:
> 
> On 2026-Jun-16, Yingying Chen wrote:
> 
>> I tried the patch, and it works properly. But I see meson automatically
>> rebuilds some more tools than src/bin/ tools, such as
>> contrib/oid2name/oid2name, src/tools/pg_bsd_indent/pg_bsd_indent, etc., do
>> you plan to handle them also?
> 
> Hmm, isn't that bogus?  oid2name at least doesn't depend on fe_utils as
> far as I can see.  Maybe meson is overeagerly inheriting that dependency
> from somewhere else.  If that's the case then we don't want Make to do
> the same (though I don't dare suggest to clean up Meson not to.)
> 
> -- 
> Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/

Thanks for your comments. Actually, I checked that while working on v1.

There are 15 non-src/bin tools that get rebuilt by meson:

* contrib/oid2name/oid2name
* contrib/vacuumlo/vacuumlo
* src/interfaces/ecpg/preproc/ecpg
* src/interfaces/ecpg/test/pg_regress_ecpg
* src/test/isolation/isolationtester
* src/test/isolation/pg_isolation_regress
* src/test/modules/libpq_pipeline/libpq_pipeline
* src/test/modules/oauth_validator/oauth_hook_client
* src/test/modules/test_escape/test_escape
* src/test/modules/test_int128/test_int128
* src/test/modules/test_json_parser/test_json_parser_incremental
* src/test/modules/test_json_parser/test_json_parser_perf
* src/test/regress/pg_regress
* src/timezone/zic
* src/tools/pg_bsd_indent/pg_bsd_indent

Only src/test/modules/test_escape/test_escape really depends on libpgfeutils.a, and I included it in v1. I think Meson may involve some broader dependency logic that causes the others to be relinked, but I don’t want to touch that part in this patch.

PFA v2, only rebased, and with the reviewer and discussion added to the commit message.

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/




pgsql: Fix out-of-bound reads with ascii() for invalid multibyte charac

От:
Noah Misch <noah@leadboat.com>
Дата:
Вложения:
Fix out-of-bound reads with ascii() for invalid multibyte characters

This commit addresses two defects in this SQL function, the code
assuming that:
- The user-supplied string was long enough to contain a character of the
length implied by the first byte.  It is possible to provide in input
data that was able to disclose a few bytes of server memory, allowing
out-of-bound reads.
- Specific bytes had values within the expected range, using a set of
assertions to validate them.  The assertions could be triggered on
invalid input.  These are replaced by tests and error reports.

Reported-by: Hcamael 
Author: Michael Paquier 
Reviewed-by: Robert Haas 
Backpatch-through: 14
Security: CVE-2026-18024

Branch
------
REL_19_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/80ce920a5a42a43d98ac40d174fea8ae518abf63
Author: Michael Paquier 

Modified Files
--------------
src/backend/utils/adt/oracle_compat.c  | 25 ++++++++++++++++++++-----
src/test/regress/expected/encoding.out | 19 +++++++++++++++++++
src/test/regress/sql/encoding.sql      | 12 ++++++++++++
3 files changed, 51 insertions(+), 5 deletions(-)

FAQ